water

Article
Water Turbidity Retrieval Based on UAV Hyperspectral
Remote Sensing
Mengying Cui 1,2,3 , Yonghua Sun 1,2,3,4, *, Chen Huang 1,2,3 and Mengjun Li 1,2,3

1 State Key Laboratory of Urban Environmental Processes and Digital Simulation, Capital Normal University,
Beijing 100048, China; [email protected] (M.C.); [email protected] (C.H.);
[email protected] (M.L.)
2 College of Resources Environment and Tourism, Capital Normal University, Beijing 100048, China
3 Beijing Laboratory of Water Resources Security, Beijing 100048, China
4 Key Laboratory of 3D Information Acquisition and Application, Ministry of Education, Beijing 100048, China
* Correspondence: [email protected]; Tel.: +86-138-1199-2433

Abstract: The water components affecting turbidity are complex and changeable, and the spectral
response mechanism of each water quality parameter is different. Therefore, this study mainly
aimed at the turbidity monitoring by unmanned aerial vehicle (UAV) hyperspectral technology, and
establishes a set of turbidity retrieval models through the artificial control experiment, and verifies
the model’s accuracy through UAV flight and water sample data in the same period. The results of
this experiment can also be extended to different inland waters for turbidity retrieval. Retrieval of
turbidity values of small inland water bodies can provide support for the study of the degree of water
pollution. We collected the images and data of aquaculture ponds and irrigation ditches in Dawa
District, Panjin City, Liaoning Province. Twenty-nine standard turbidity solutions with different
concentration gradients (concentration from 0 to 360 NTU—the abbreviation of Nephelometric
Turbidity Unit, which stands for scattered turbidity.) were established through manual control and
we simultaneously collected hyperspectral data from the spectral values of standard solutions. The





 sensitive band to turbidity was obtained after analyzing the spectral information. We established
Citation: Cui, M.; Sun, Y.; Huang, C.; four kinds of retrieval, including the single band, band ratio, normalized ratio, and the partial least
Li, M. Water Turbidity Retrieval squares (PLS) models. We selected the two models with the highest R2 for accuracy verification. The
Based on UAV Hyperspectral Remote band ratio model and PLS model had the highest accuracy, and R2 was, respectively, 0.65 and 0.72.
Sensing. Water 2022, 14, 128. https:// The hyperspectral image data obtained by UAV were combined with the PLS model, which had the
doi.org/10.3390/w14010128 highest R2 to estimate the spatial distribution of water turbidity. The turbidity of the water areas in
Academic Editors: Kun Shi and the study area was 5–300 NTU, and most of which are 5–80 NTU. It shows that the PLS models can
Thomas Meixner retrieve the turbidity with high accuracy of aquaculture ponds, irrigation canals, and reservoirs in
Dawa District of Panjin City, Liaoning Province. The experimental results are consistent with the
Received: 9 December 2021
conclusions of the field investigation.
Accepted: 3 January 2022
Published: 5 January 2022
Keywords: UAV hyperspectral; turbidity; retrieval model; remote sensing; water body
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations. 1. Introduction
Turbidity is an important parameter to characterize water quality, especially the
transparency of the water area [1]. At the same time, there is a close relationship between
Copyright: © 2022 by the authors.
turbidity and other water quality parameters. The analysis of its change is helpful to
Licensee MDPI, Basel, Switzerland. understand the distribution of sediment or total suspended solids in the water environment
This article is an open access article and provides practical information for the study of pollutant deposition, decomposition,
distributed under the terms and and diffusion in the water [2]. Therefore, monitoring of spatial turbidity distribution is
conditions of the Creative Commons significant for the water ecological environment and human production and living. The
Attribution (CC BY) license (https:// intensity of incident light will be weakened after the absorption and scattering of light
creativecommons.org/licenses/by/ by particles in water, because turbidity is a water quality parameter with optical activity.
4.0/). Furthermore, the amount of attenuation and the effect of water on incident light can be

Water 2022, 14, 128. https://doi.org/10.3390/w14010128 https://www.mdpi.com/journal/water

Water 2022, 14, 128 2 of 15

well reflected in remote sensing images. Therefore, remote sensing can indirectly obtain the
spatial distribution of turbidity, which is expressed according to the amount of attenuation.
Remote sensing has the advantages of a wide range and fast speed, effectively sup-
plementing conventional water environment monitoring methods [3,4]. With the rapid
development of remote sensing and the in-depth study on the spectral characteristics and
algorithms of water quality parameters, the quantitative degree of water quality retrieval
by remote sensing methods has been continuously improved [5]. An inversion algorithm
is established by studying the relationship between water reflectance spectral characteris-
tics and water quality parameter concentration [6,7]. Due to the strong spatial-temporal
variation of turbidity and apparent seasonal variation, remote sensing is also needed for
long-term dynamic water turbidity monitoring. The use of UAV technology can be more
flexible monitoring of small inland water [8]. Establishing a turbidity retrieval model based
on remote sensing data, monitoring the change of water turbidity, and studying the change
characteristics of water turbidity provides new prospects for applying remote sensing
technology and supplies a new way to study water quality and water color.
In previous studies, researchers concluded that the reflectance of the red band is closely
related to turbidity, which is the key band for turbidity retrieval. Most models are built
around the red band in the single band model, such as the turbidity model constructed
by Xu [9], Douglas [10], etc. In addition, in the study of turbidity retrieval using UAV
hyperspectral images, the near-infrared (NIR) band is often used for model building and
has a high band sensitivity, as Kornelia [11], concluded that red light band was used
to identify low-medium turbidity waters, and the NIR band was used to identify high-
turbidity waters when studying sediment-rich glacial meltwater plumes. Jiang [12] used
the data of headwall hyperspectral imager to build a variety of multiple linear regression
models and PLS models for comparison. Eventually, the accuracy of the PLS model was
significantly better than the other models. Many researchers have established the ratio
model between the NIR band or the red band and the green band, which can obtain higher
retrieval accuracy. Xiao [2] used three kinds of domestic satellite—HJ-1A/1B CCD, GF-1
WFV, and ZY-3 MUX—data for water turbidity retrieval and constructed the ratio model of
red and green bands as the turbidity retrieval model. In addition, Song [13] and Feng [14]
used blue, green, red, and NIR bands to highlight the spectral characteristics of turbidity.
However, it is challenging to combine the bands, due to the large number of bands and a
large amount of data. So, the combination method is often used in the research of turbidity
retrieval with multispectral data.
The UAV remote sensing system has flexibility, rapid response, simple structure, and
good real-time performance. The remote sensing image data, with high spatial resolution,
high temporal resolution, and high spectral resolution, can be acquired at low altitude,
which effectively makes up for the deficiency of traditional satellite remote sensing technol-
ogy [15]. When the weather conditions permit, using UAVs anytime and anywhere can
provide a valuable monitoring scheme for water quality monitoring and water pollution
prediction. This can be used to timely discover the pollution situation of small and micro
water areas, which is difficult to be considered by traditional water quality monitoring
methods. Nowadays, as a new and brand-new remote sensing platform, UAV has been
widely used and gradually shows its importance [16,17]. The emergence and development
of UAV hyperspectral technology provide a new technical means for large-scale water
quality monitoring and research. It makes up for any defects of satellite remote sensing
water quality monitoring [18]. It has excellent guiding significance for water environment
monitoring of small and micro water areas such as rivers.
In this study, we collected the reflectance data of the hyperspectral imager based on
the standard turbidity solutions of twenty-nine concentrations. The retrieval model of
turbidity standard solution was constructed after reflectance processing. Instead of the
traditional method of constructing turbidity retrieval model by collecting data in the field,
a set of retrieval models from low to high turbidity water areas was established, and the
accuracy of retrieval was verified by UAV data. It measured data in the same period, and
Water 2022, 14, 128 3 of 15

the best retrieval model suitable for aquaculture ponds, irrigation canals, and reservoirs in
Dawa District of Panjin City, Liaoning Province was obtained.

2. Data Acquisition
2.1. Manual Control Experiment
2.1.1. Solution Concentration Configuration
In this manual control experiment, we used the turbidity solution with a concentration
of 4000 NTU for different concentration configurations. We set a total of twenty-nine gradi-
ents of turbidity solution with different interval gradients from 0 to 360 NTU. Moreover, a
black box with a volume of 13 L was prepared as a container (38.5 cm × 29.5 cm × 20.5 cm).
To reduce the influence of the reflection of the box bottom and the box wall on the water
spectral measurement, the black light-absorbing paper, and the light-absorbing adhesive
tape were pasted on the bottom and wall before the experiment. However, the effect on
water reflectance still cannot be ignored, especially when the concentration of turbidity
standard solution is low. Owing to the limited amount of simulated water volume, inho-
mogeneity of turbidity distribution via volume, and because the water in ponds, irrigation
canals, and reservoirs is not perfectly smooth, waves are present. These reasons will limit
the ability to use the regression equation obtained in this experiment in natural waters to
a certain extent. Therefore, this experiment cannot completely simulate the actual river
environment and accuracy estimates are required. To ensure the accuracy of the experi-
ment, we calculated the milliliter required for each adjustment of solution concentration
in advance. For the convenience of calculation, we used 12 L distilled water as the initial
water sample for the experiment (the turbidity was 0 NTU). We added the corresponding
milliliter in turn to adjust the concentration gradient and measured the readings of the
configured solution with HACH 2100 q turbidimeter from HACH Company in America to
ensure that the error between the concentration of the prepared turbidimetric solution and
the set gradient was within 5 NTU. During the experiment, assuming that the background
spectrum remains unchanged, the change of water reflection spectrum is caused by the
change of turbidity concentration.

2.1.2. Water Reflection Spectrum Acquisition

The hyperspectral imager used was the Gaiasky-mini2-VN from Sichuan Dualix Spec-
tral Imaging Technology Company, with a spectral range of 400–1000 nm and 176 spectral
channels. The imager is often carried on rotor UAVs for image acquisition. The collected
data are used in the corresponding research of small-scale features [19]. In this experiment,
we selected the data acquisition time of the hyperspectral imager to be 10 a.m. with suffi-
cient light. The acquisition time of each gradient was strictly controlled so that the total
acquisition time did not exceed two hours. We ensured sufficient illumination intensity
and synchronization of data acquisition, and settled on a better foundation for subsequent
data processing. The hyperspectral imager was fixed on a tripod about 2 m high. A black
box containing solution and a standard gray board with the same height were placed on
the box. We used the hyperspectral imager to synchronously collect the water and gray
board spectra after adjusting the concentration of the solution. To ensure the accuracy of
data collection, we kept the lens height unchanged during the experiment.
The obtained hyperspectral raw data were processed as reflectance data. The low-pass
filtering could smooth and denoise the image effectively. After that, the spectral information
collected by the hyperspectral imager was processed with reflectance. The average value of
the same region of 5 × 5 pixels range was selected in the frame of the hyperspectral images
with different turbidity. The relationship between the digital value (DN) and reflectivity is
as follows:
DNwater
RefHyp = × Refgreyboard (1)
DNgreyboard
 Water 2022, 14, x FOR PEER REVIEW 4 of 16

Water 2022, 14, 128 4 of 15

DN
Ref = Ref (1)
DN
where,Ref
where, Ref
Hyp isis the
the reflectance
reflectance value
value ofof
thethe image
image after
after reflectance
reflectance correction,
correction, DN DNwater is
is DN
the the DN
valuevalue oforiginal
of the the original
image,image, DN
DNgreyboard is theisoriginal
the original DN value
DN value of the of theboard,
gray gray
board,
and and Refis the standard
Refgreyboard is the standard reflectance
reflectance of the grayof the gray board.
board.
Based
Basedon onthe
the spectral data of
spectral data of the
thegray
grayboard
boardcollected
collected in in
thethe same
same period
period of ex-
of the the
experiment,
periment, thetheDNDNvalue
valuewaswastransformed
transformedinto intothe
theabsolute
absolutereflectance
reflectance data
data to
to obtain
obtain the
the
hyperspectral reflectance data of different concentrations.
hyperspectral reflectance data of different concentrations.

2.2.
2.2.UAV
UAVField
FieldData
DataAcquisition
Acquisition
2.2.1.
2.2.1. Image Acquisitionand
Image Acquisition andSampling
Sampling
Figure
Figure1 1shows
showsthe thestudy
studyareas
areasof
offield
fielddata
dataacquisition
acquisition in in this
this experiment. The field
experiment. The field
flight tests were conducted in September 2019 and October 2020, respectively,
flight tests were conducted in September 2019 and October 2020, respectively, and the and the
images
imageswerewerecollected
collectedusing
usingthetheGaiasky-mini2-VN
Gaiasky-mini2-VN hyperspectral imager, carried
hyperspectral imager, carried byby the
the
DJI M600 pro UAV. We obtained three hyperspectral data sets of UAVs.
DJI M600 pro UAV. We obtained three hyperspectral data sets of UAVs. The flight area is The flight area is
located in Dawa
located in Dawa District, Panjin City, Liaoning Province. The assignment
District, Panjin City, Liaoning Province. The assignment included included a flying
a flying
area 2 2 for breeding
areaofof0.25
0.25km
km2forforthe
theRongxing
Rongxingreservoir
reservoirand
and two
two flying
flying areas
areas of
of 0.95
0.95 km
km2 for breeding
ponds
pondsand andirrigation
irrigationcanals.
canals.Before
Beforethe
theUAV
UAVtook
tookoff,
off,a astandard
standardgray
graycloth
clothwith
withreflectivity
reflectiv-
ofity
20% and an area of 2 × 2 m had been placed in the aerial photography
of 20% and an area of 2 × 2 m had been placed in the aerial photography area. In-flight,area. In-flight,
the altitude was set at 200 m, the course overlap was set at 55%, the side
the altitude was set at 200 m, the course overlap was set at 55%, the side overlap was overlap was setset
at
40%, and the flight speed was 12.8
at 40%, and the flight speed was 12.8 m/s. m/s.

Figure 1. The location of the study area.

Figure 1. The location of the study area.

Duringthe
During theflight,
flight,synchronous
synchronoussampling
samplingwas wascarried
carried out
out to
to record
record the
the longitude
longitude and
and
latitude of the sampling point, the water quality of the river, and to take photos. A
latitude of the sampling point, the water quality of the river, and to take photos. A total
total
ofof1818water
watersamples
sampleswere werecollected,
collected,and
andthe
thelocation
location information
information of of the
the sampling
sampling point
point
was recorded. After the water sample was collected, it was brought back to the
was recorded. After the water sample was collected, it was brought back to the laboratory
laboratory
inintime.
time.The
Theturbidity
turbidityvalue
value was
was read
read by
by the
the HACH
HACH 21002100 qq portable
portable turbidimeter.
turbidimeter. The
The
turbidity range was 3.2–56.37 NTU, and the average value was 22.34 NTU. The statistical
data are shown in Table 1.
 turbidity range was 3.2–56.37 NTU, and the average value was 22.34 NTU. The statistical
data are shown in Table 1.

Table
Water 2022, 14, 128 1. Turbidity statistics of sampling points. 5 of 15

Time Location Numbers Max/NTU Min/NTU Average/NTU

121°59′44″–121°59′51″ E,
9 10
Table 1. Turbidity 40.47
statistics 3.26
of sampling points. 18.32
40°48′38″–40°48′51″ N
121°57′8″–121°57′15″
Sampling Time E, Location Numbers Max/NTU Min/NTU Average/NTU
0 80 56.37 3.2 25.42
40°47′14″–40°47′20″ N59 44”–121 59 51” E,
121 ◦ 0 ◦
2019.09 10 40.47 3.26 18.32
40◦ 480 38”–40◦ 480 51” N
121◦ 570 8”–121◦ 570 15” E,
2.2.2.
2020.10 Image Preprocessing and Water
40◦ 470 14”–40◦ 470 20” N
Extraction
8 56.37 3.2 25.42
a. Reflectivity calibration and radiometric correction
Firstly, the 2.2.2.
original
Image image aberration
Preprocessing and Extraction
and Water distortion were corrected through
SpecView software to a. complete thecalibration
Reflectivity lens correction of the UAV
and radiometric image, eliminate the color
correction
difference, and correctFirstly,
the aperture.
the original image aberration and distortion of
Secondly, in the flight process wereUAVs, duethrough
corrected to dif- SpecView
ferent light and weather
softwareconditions,
to complete the acquired
the lens image
correction willUAV
of the be distorted in the the
image, eliminate spectral
color difference,
dimension, and the anddata need
correct the to be radiometric
aperture. Secondly, incorrected
the flight[20]. In of
process this study,
UAVs, duethe radio- light and
to different
weather
metric correction was alsoconditions,
operated the acquired image
by SpecView will beTo
software. distorted in the
eliminate spectral
the errorsdimension,
caused and the
by the performance of the UAV sensor, we obtained the reflectivity value of the image, was also
data need to be radiometric corrected [20]. In this study, the radiometric correction
and eliminated the operated by SpecView
influence software. Towater
of the atmosphere, eliminate the and
vapor, errors caused
other by the
factors performance
when the of the
UAV sensor, we obtained the reflectivity value of the image, and eliminated the influence
UAV flies to a certain altitude, we adopted black-and-white correction and atmospheric
of the atmosphere, water vapor, and other factors when the UAV flies to a certain altitude,
correction. The atmospheric
we adoptedcorrection formula
black-and-white is as follows:
correction and atmospheric correction. The atmospheric
DN
correction formula is as follows: Ref
Ref = (2)
DN DNHyp × Refstandard
Reffixed = (2)
DNgray
where Ref is the spectral reflectance of the image after eliminating water vapor and
atmosphere. Ref where is the
Refreflectance of the image
fixed is the spectral afterofreflectance
reflectance correction.
the image after Ref water vapor and
eliminating
atmosphere.
is the spectral reflectance of theRefHyp
grayis cloth
the reflectance
calibrated of the
by image after reflectance
the National Institutecorrection.
of Metrol-Refstandard is
ogy. Ref is the reflectance of the gray cloth in the image after reflectance correction. Metrology.
the spectral reflectance of the gray cloth calibrated by the National Institute of
b. GeometricRef gray is the reflectance of the gray cloth in the image after reflectance correction.
correction
b. Geometric correction
Due to the influence of external factors such as airflow, wind direction, and wind
Due to the influence of external factors such as airflow, wind direction, and wind speed
speed in the flightinprocess
the flight ofprocess
UAVs,ofthere
UAVs, is there
an inevitable error between
is an inevitable the image
error between geo-geographic
the image
graphic coordinates and the actual coordinates [21]. At the same time,
coordinates and the actual coordinates [21]. At the same time, to eliminate to eliminate thethe system
system error caused errorbycaused
UAV by sensors, it is necessary
UAV sensors, to carry
it is necessary out out
to carry a geometric
a geometric correction
correction for UAV
for UAV images. The geometric
images. distortion
The geometric of theofUAV
distortion the UAVimage waswas
image simulated
simulated with
withthe
thedata
data of control
of control points selected on the ground in the field flight test. The flight data of the UAVafter image
points selected on the ground in the field flight test. The flight data of the UAV
preprocessing
after image preprocessing is shown
is shown in Figure
in Figure 2. 2.

Figure 2. UAV flightFigure

data in the study
2. UAV flight area, according
data in the study to theaccording
area, R, G, B band,
to the the corresponding
R, G, bands
B band, the corresponding bands
are the 77th, 50th, and
are 26th bands,
the 77th, 50th,and
andthe wavelengths
26th arewavelengths
bands, and the 642, 551, and
are 472
642,nm,
551, respectively.
and 472 nm, respectively.

c. Water mask extraction

c. Water mask extraction
The method
The method of extracting of extracting
water watersensing
from remote from remote sensing
images images
is a hot is a hot
research research topic
topic
at present, which has been widely developed in the aspects of water body index andindex and
at present, which has been widely developed in the aspects of water body
classification. The normalized difference water index (NDWI) method [22] extracts the
water body from the three UAV flight data obtained after preprocessing, so that the land,
Water 2022, 14, 128 6 of 15

vegetation, and other surface features outside the water area which have interference are
removed from the image and do not participate in the quantitative retrieval. The calculation
formula is as follows:
Ref547nm − Ref863nm
NDW I = (3)
Ref547nm + Ref863nm
After the calculation, the region of interest was extracted to collect the water samples.
After the NDWI value was calculated, it was found that when the threshold value was 0.06,
the water could be separated from other land types. Because the study area was located
in the plain area, there was no influence of mountain shadow, so the threshold method
is feasible. The segmentation threshold is set to 0.06 to realize the automatic extraction
of water.
d. Spectral smoothing
In order to solve the influence of the external environment in the process of collecting
ultraviolet spectrum and NIR spectrum data, that is, to eliminate the characteristic spectral
noise outside the band with a high standard deviation of the noise, especially the serious
spectral overlap in the NIR spectrum [23], and reduce the impact of error, low-pass filtering
is used for data smoothing, we used the low-pass filter for image smoothing. The window
width of 5 * 5 pixels centered on the water sampling point was selected to weaken the
positioning error. It can eliminate noise interference and avoid the loss of image local
information due to excessive smoothing. The processed spectrum can better retain the
peak–valley characteristics [24].

3. Methods
3.1. Experimental Process
The research process is shown in Figure 3. Given the artificial control experiment on
the ground, the turbidity retrieval model of the hyperspectral imager was established. The
reflectance data was obtained by preprocessing hyperspectral imager data. The correlation
analysis method selects the best retrieval band according to the relationship between
ground experiment spectral data and turbidity. The R2 and root mean square error (RMSE)
models were used to evaluate. After various forms of model construction and screening,
the model with a higher determination coefficient was selected as the preliminary model for
turbidity retrieval. Then, the spectral information of the field sampling points was extracted
according to the field flight hyperspectral image data, and the synchronous sampling data
verified the model accuracy. Finally, the model with the highest retrieval accuracy was
obtained, and the model was applied to draw the turbidity retrieval map of the flight image
in the study area.

3.2. Model Construction

3.2.1. Model Selection
a. Single-band model
Similar to other water quality parameter retrieval, the single-band model is often used
in turbidity retrieval. The single-band model is a remote sensing retrieval method based
on statistical analysis. The basic principle is to find the band with the highest correlation
with turbidity and establish the retrieval model. Due to the strong optical response of
turbidity index, high reflectivity can fully obtain the information in the water [25,26], so
the single band quantitative model of turbidity can be established. For example, Joshi
used the optimized and atmospheric corrected b3 band (the red band) of Landsat 5 TM to
build the turbidity retrieval model and proposed a seasonal threshold for different seasonal
applications [27].
Water 2022, 14, 128 7 of 15
Water 2022, 14, x FOR PEER REVIEW 7 of 16

Figure Flow
3. 3.
Figure chart
Flow ofof
chart turbidity model
turbidity construction
model and
construction validation.
and validation.

3.2.b.Model
BandConstruction
ratio model
The band ratio model is used to select the most remarkable combination with the
3.2.1. Model Selection
measured turbidity from all possible band models and calculate the ratio to build the
a. Single-band
retrieval model [28]. When modelcompared with the single-band model, the band ratio model
Similar to other
can eliminate the interference water of quality parameter
background noise,retrieval, the single-band
such as water model is and
surface smoothness often
used in turbidity retrieval. The single-band model is a remote sensing
surrounding environment, and reduce the influence of other pollutants to a certain extent, retrieval method
tobased
improveon statistical
the retrieval analysis. The basic
accuracy principle
[29–34]. is to selection
The band find the band
range with the ratio
of the highest corre-
model
lation with turbidity and establish the retrieval model. Due to the
is narrow, and the optimal band combination can be found through band iteration after strong optical response
of turbidity
analyzing index, characteristics
the optical high reflectivity can
[35]. Fufully obtainthat
[36] found the logarithm
information in the water
conversion [25,26],
of spectral
so the single
reflectance ratioband quantitative
is helpful model ofremote
for quantitative turbidity can beretrieval
sensing established. For example,
of turbidity content.Joshi
used the optimized
c. Normalized andmodel
ratio atmospheric corrected b3 band (the red band) of Landsat 5 TM to
build the turbidity retrieval
Normalization of the spectrum model can
andimprove
proposed thea seasonal
correlation threshold
betweenfor different
turbidity andsea-
sonal applications [27].
remote sensing reflectance. The original spectral reflectance of the hyperspectral imager
b. Band ratio
is normalized, and themodel
correlation between the normalized spectral reflectance and the
measuredTheturbidity
band ratio model is The
is analyzed. usednormalized
to select the most remarkable
reflectance combinationbywith
model is established using the
measured
the turbidityand
band reflectance from all possible
turbidity withband modelscorrelation
the largest and calculate the ratio Cao
coefficient. to build
usedthe there-
trievalspectrometer
surface model [28]. When compared
data to complete with
thethe single-band
quantitative model, the band
hyperspectral ratio of
retrieval model
Nansi can
Lake water and
eliminate finally established
the interference the NIR and
of background green
noise, band’s
such normalized
as water surfaceratio model [37].
smoothness and
d. Partial least
surrounding squares regression
environment, and reduce model
the influence of other pollutants to a certain extent,
Partial least
to improve the squares
retrievalregression
accuracy (PLS) is aThe
[29–34]. multiple
band linear regression
selection range of modeling
the ratiomethod.
model is
Innarrow,
modeling, andit theintegrates
optimalthree
bandanalysis methods:
combination can principal
be found component
through band analysis, linear
iteration after
Water 2022, 14, 128 8 of 15

regression analysis, and canonical correlation analysis, suitable for data processing and
modeling with complex variables and high correlation. He Yong used NIR sensors to
obtain spectral information of samples and used the PLS method to establish a prediction
model of soil nitrogen content. The prediction effect of the PLS model was outstanding
and stable [38]. Wang used the PLS model based on four machine learning methods to test
the relationship between the total nitrogen concentration extracted from the hyperspectral
images from UAVs during plant growth and the measured total nitrogen concentration [39].
The PLS modeling principle is: Suppose P independent variables {x1 , . . . , xp } and the
dependent variables {y1 , . . . , yq }, data table X = {x1 , . . . , xp } and Y = {y1 , . . . , yq }. t1 and
u1 are extracted, the regression of X to t1 and Y to u1 is carried out, respectively. If the
regression equation has reached satisfactory accuracy, then the components are determined.
Otherwise, the residual information after X is interpreted by t1 and Y by u1 will be used
for the second round of extraction components t2 and u2 . The regression of X and Y to t2
and u2 will be continued. The above process will be iterated until the accuracy meets the
requirements. In that case, they can be converted into Y’s regression on the original variable
x1 , . . . , xp regression equation completed the partial least squares regression modeling.

3.2.2. Modeling through Artificial Control Experiment

We selected the above four models to model the turbidity standard solution data
collected by the hyperspectral imager. Among them, the PLS regression model of hyper-
spectral imager selected the first 161 bands in the range of 400–1000 nm as the research
data. After the standard solution was measured by a hyperspectral imager, the spectral
data were processed; the last 15 bands after 940 nm had a 0 value. The independent
variable matrix X dimension was 29 × 161, where 29 represents turbidity concentrations
and 161 represents the number of bands. The dependent variable was the turbidity value
for each gradient. X and Y were standardized, respectively, and a partial least squares
regression was performed.

3.3. Accuracy Assessment

We used four indicators to evaluate the modeling effect, including the coefficient
of determination R2 , mean bias error (MBE), root mean square error (RMSE), and mean
absolute percentage error (MAPE). The basic statistical indicators formula is shown in
Table 2.

Table 2. Basic Statistical Indicators.

Name Formula
n 0 2
Determination coefficient (R2 ) ∑ ( y − yi )
R2 = in=1 i 2
∑
s i =1 i − y i )
( y
n
Root mean square error (RMSE) 1 0

2
RMSE = n ∑ yi − yi
i =1
2
(yi0 −yi )
 
Mean bias error (MBE) MBE = 1
∑in=1
n n
n yi − yi
0
Mean absolute percent error (MAPE) MAPE = 1
∑ × 100%
n yi
i =1
where i represents the serial number of validation sample points, and n represents the number of validation
0
samples, yi is the predicted turbidity value, yi is the observed turbidity value, yi is the mean of the observed
turbidity value.

4. Results and Discussion

4.1. Spectral Characteristic Analysis
Hyperspectral images have a large amount of data. The signal should be retained in
the band selection process, and the noise should be compressed as much as possible [40].
According to the average spectral image, identify and mark the noise bands in the hyper-
spectral image. Furthermore, apply the PIE software (Pixel Information Expert) to evaluate
 4. Results and Discussion
4.1. Spectral Characteristic Analysis
Hyperspectral images have a large amount of data. The signal should be retained in
Water 2022, 14, 128 9 of 15
the band selection process, and the noise should be compressed as much as possible [40].
According to the average spectral image, identify and mark the noise bands in the hyper-
spectral image. Furthermore, apply the PIE software (Pixel Information Expert) to evalu-
thethe
ate noise of of
noise thethe
water
waterarea
areaextracted
extractedfrom
fromthe
theimage
imageto
tofind
find out
out the noise bands
the noise bands without
without
changingthe
changing thecoherent
coherent bands
bands ofof the
the hyperspectral
hyperspectral image
image [41].
[41].
Figure 44 shows
Figure shows thetheestimated
estimatednoise
noisestandard
standard deviation
deviationof of
thethe
hyperspectral reflectance
hyperspectral reflec-
image. According to the noise shown, the noise corresponds to the blue
tance image. According to the noise shown, the noise corresponds to the blue band band near 436 nm
near
and the NIR band after 863 nm. Consider the following modeling process
436 nm and the NIR band after 863 nm. Consider the following modeling process to avoid to avoid using
the bands
using in thein
the bands above rangerange
the above for model building.
for model building.

Theestimated
Figure4.4.The
Figure estimatednoise
noisestandard
standarddeviation
deviationof
ofhyperspectral
hyperspectralreflectance
reflectance image.
image.

According to the reflectance curve in Figure 5, there is a strong reflection peak in the
range of 391–460 nm in the blue band and 770–830 nm in the NIR band. The reflection value
decreases in the range of 720–770 nm due to the strong absorption, and the absorption peak
appears. Due to the fact that the water set in the ground control experiment is single water
with different turbidity concentrations and no other water quality parameters, the change
of water spectral reflectance is relatively consistent. Combined with the results of noise
estimation, it is considered that the blue band of hyperspectral imager should be avoided
in the modeling process. Therefore, it can be confirmed that the regression equation with a
good correlation coefficient can be established in the NIR band.

4.2. Model Retrieval of Turbidity Standard Solution

4.2.1. Manual Control Experimental Models for Hyperspectral Image
According to the correlation between the band of the hyperspectral imager and turbid-
ity standard solution, the high correlation range of the band reflectance is between 800 nm
and 900 nm. Moreover, the single band model with the highest accuracy, 809 nm, as the
independent variable is established.
Figure 6a shows a high correlation between the hyperspectral 700 nm and 500 nm
bands. After calculation, the ratio model of hyperspectral bands with 707 nm and 541 nm
as independent variables was obtained, that is, the ratio model of red and green bands.
This result is also consistent with Xiao’s research results [2]. Figure 6b shows the result of
the normalized ratio correlation. The normalized ratio calculation in the red and NIR band
ranges has a higher correlation with the concentration. Finally, the normalized ratio model
consisting of 669 nm and 551 nm bands was obtained.
 value decreases in the range of 720–770 nm due to the strong absorption, and the ab
tion peak appears. Due to the fact that the water set in the ground control experim
single water with different turbidity concentrations and no other water quality par
ters, the change of water spectral reflectance is relatively consistent. Combined wit
Water 2022, 14, 128 results of noise estimation, it is considered that the blue band of hyperspectral
10 of 15 im
should be avoided in the modeling process. Therefore, it can be confirmed that the re
sion equation with a good correlation coefficient can be established in the NIR band

Water 2022, 14, x FOR PEER REVIEW 11 of 16

Figure
Figure 5. Reflectance 5. Reflectance
curve curve of
of hyperspectral hyperspectral
imager turbidityimager turbidity
standard standard solution.
solution.

4.2. Model Retrieval of Turbidity Standard Solution

4.2.1. Manual Control Experimental Models for Hyperspectral Image
According to the correlation between the band of the hyperspectral imager and
bidity standard solution, the high correlation range of the band reflectance is bet
800nm and 900 nm. Moreover, the single band model with the highest accuracy, 809
as the independent variable is established.
Figure 6a shows a high correlation between the hyperspectral 700 nm and 50
bands. After calculation, the ratio model of hyperspectral bands with 707 nm and 54
as independent variables was obtained, that is, the ratio model of red and green b
This result is also consistent with Xiao’s research results [2]. Figure 6b shows the res
the normalized ratio correlation. The normalized ratio calculation in the red and NIR
ranges has a higher correlation with the concentration. Finally, the normalized ratio m
consisting of 669 nm and 551 nm bands was obtained.

(a) (b)
Figure 6. The correlation coefficient distribution, (a) is the distribution of the ratio correlation coef-
Figure 6. The correlation coefficient distribution, (a) is the distribution of the ratio correlation
ficient in the hyperspectral band; (b) is the distribution of the normalized ratio correlation coefficient
coefficient in the hyperspectral band; (b) is the distribution of the normalized ratio correlation
in the hyperspectral band.
coefficient in the hyperspectral band.
Figure 7a
Figure 7a shows
shows thatthat the
the regression
regression coefficients
coefficients of
of each
each band
band in
in the
the PLS
PLS model
model can
can
reflect the importance of each variable to the model. Different bands have different
reflect the importance of each variable to the model. Different bands have different effects effects
on the
on the model
model[42].
[42].TheTheregression
regression coefficient of each
coefficient band
of each in the
band in hyperspectral bandband
the hyperspectral 450–
450–550 nm is negative, those of 550 nm–800 nm are positive, and those of 850 nm–1000 nm
550 nm is negative, those of 550 nm–800 nm are positive, and those of 850 nm–1000 nm
are negative
are negative and
and then
then positive.
positive. The
The regression coefficients of
regression coefficients of all
all bands
bands in
in the
the hyperspectral
hyperspectral
band are
band are the
the same,
same, andand before
before 750 750 nm
nm band,
band, the
the regression
regression coefficients
coefficients fluctuate
fluctuate slightly
slightly
around zero plus or minus. It shows that the contribution of each band of
around zero plus or minus. It shows that the contribution of each band of hyperspectral to hyperspectral
to the
the final
final model
model is relatively
is relatively balanced.
balanced.
 reflect the importance of each variable to the model. Different bands have different effects
on the model [42]. The regression coefficient of each band in the hyperspectral band 450–
550 nm is negative, those of 550 nm–800 nm are positive, and those of 850 nm–1000 nm
are negative and then positive. The regression coefficients of all bands in the hyperspectral
band are the same, and before 750 nm band, the regression coefficients fluctuate slightly
Water 2022, 14, 128 11 of 15
around zero plus or minus. It shows that the contribution of each band of hyperspectral
to the final model is relatively balanced.

(a) (b)
Figure 7. The regression coefficient of PLS model, (a) is each band’s regression coefficient; (b) is the
Figure 7. The
variance regression
explained coefficient
rate of of PLScomponent.
each principal model, (a) is each band’s regression coefficient; (b) is the
variance explained rate of each principal component.

According to the analysis, there are 27 principal components, and the contribution rate
of principal component five reached 98%. The contribution rate from principal components
6 to 27 is less than 10%. Moreover, the slope of contribution rate from principal component
five tended to be gentle, as shown in Figure 7b. This result demonstrates that when the
principal component is five, it contains the most original data information. Therefore, select
comps = 5 to build the model. The final model can be expressed as y= 4.8856 x1 + 1.2681x2 +
0.6599x3 + 0.7141x4 + 0.4754x5 . Finally, the single-band, band ratio, normalized ratio, and
PLS models of the hyperspectral imager are shown in Table 3.

Table 3. Hyperspectral imager model and modeling accuracy.

Model Type Equation Form R2 RMSE/NTU MBE MAPE/%

Single band T = 17.9 × e19.5 × Rrs (809) 0.86 4 5.66 28.54
Band ratio Rrs (707) 0.92 3.67 5.23 12.93
T = 1305.1 × In Rrs (541)
)+876.2
Normalized ratio Rrs (669)− Rrs (551) 0.87 8.82 3.61 20.26
T = 4633.6 × Rrs (669)+ Rrs (551)
+936.7
PLS 0.98 6.09 3.81 15.4
where T is the abbreviation of turbidity.

4.2.2. Verification of Model Accuracy

According to the establishment of the above model, we compared the multiple linear
regression models. According to the principle of maximum R2 and minimum RMSE [43],
the band ratio model and the PLS hyperspectral model were selected to verify the accuracy
of the measured and predicted values. We extracted the reflectance of the hyperspectral
image of 18 sampling points and substituted the required band into two models to obtain
the turbidity retrieval value. When compared with the measured turbidity, we obtained
the correlation analysis scatter diagram of the measured and predicted turbidity value, as
shown in Figure 8.
Comparing the predicted values with the measured values of the retrieval model
shows that the predicted values were higher than the actual turbidity values. The MAPE
was 22.46 and 19.35%, respectively, but the relative trend is consistent with the y = x
standard line, R2 is 0.65 and 0.72, respectively. It shows that the two models can retrieve
the turbidity of ponds, irrigation canals, and reservoirs in Dawa District of Panjin City,
Liaoning Province, and the accuracy is good.
 regression models. According to the principle of maximum R2 and minimum RMSE [43],
the band ratio model and the PLS hyperspectral model were selected to verify the accu-
racy of the measured and predicted values. We extracted the reflectance of the hyperspec-
tral image of 18 sampling points and substituted the required band into two models to
obtain the turbidity retrieval value. When compared with the measured turbidity, we ob-
Water 2022, 14, 128 12 of 15
tained the correlation analysis scatter diagram of the measured and predicted turbidity
value, as shown in Figure 8.

Water 2022, 14, x FOR PEER REVIEW 13 of 16

Comparing the predicted values with the measured values of the retrieval model
shows that the predicted values were higher than the actual turbidity values. The MAPE
was 22.46 and 19.35%, respectively, but the relative trend is consistent with the y = x stand-
(a) (b)
ard line, R2 is 0.65 and 0.72, respectively. It shows that the two models can retrieve the
Figure 8. Analysis
turbidity of ponds, of the predicted
irrigation value ofand
canals, turbidity retrieval,
reservoirs in (a) is theDistrict
Dawa result of of
hyperspectral band
Panjin City, Liao-
Figure
ratio 8. Analysis
model; (b) isof
thethe predicted
result value of turbidity
of PLS Hyperspectral retrieval,
retrieval model.(a) is the result of hyperspectral band
ning Province, and the accuracy is good.
ratio model; (b) is the result of PLS Hyperspectral retrieval model.

4.3.UAV
4.3. UAVImage
Image Turbidity
Turbidity Retrieval
Retrieval Using
UsingthetheOptimal
OptimalModel
Model
Thedetermination
The determinationcoefficient 2
coefficientRR and
2 andthetheRMSE
RMSEindex
indexofofthe
thestandard
standardsolution
solutionturbidity
tur-
bidity
PLS PLS constructed
model model constructed by the experiment
by the experiment are excellent.
are excellent. Therefore,
Therefore, the hyperspec-
the hyperspectral image
tral image data obtained by UAVs in the study area were combined with
data obtained by UAVs in the study area were combined with the model for retrieval, andthe model for
retrieval,
the and the water
water turbidity turbidity
distribution distribution
map was drawn. mapAswas drawn.
shown As shown
in Figure in Figure
9, the 9, results
retrieval the
retrieval results show that the turbidity of the water area in the study area was
show that the turbidity of the water area in the study area was 5–300 NTU, and most of 5–300 NTU,
and most
them were of them
5–80 were
NTU. 5–80
The NTU.water
overall The overall
area waswater area was
relatively relatively straightforward,
straightforward, and there was
and there was no apparent pollution phenomenon in
no apparent pollution phenomenon in the pond and ditch water the pond andarea.
ditchThe
water area. The
turbidity at the
boundary of the aquaculture pond was relatively high. The high value in somevalue
turbidity at the boundary of the aquaculture pond was relatively high. The high areasinmay
some areas may be due to the muddy water caused by road water and sediment. The
be due to the muddy water caused by road water and sediment. The retrieval results agree
retrieval results agree with the field water conditions, which are in line with the field in-
with the field water conditions, which are in line with the field investigation results.
vestigation results.

Figure9.9.Local
Figure Local turbidity
turbidity retrieval
retrieval map
mapof
ofUAV
UAVimages
imagesininPLS
PLSmodel.
model.

5. Conclusions
In this study, the manual control experiment prepared twenty-nine gradients of tur-
bidity solutions with different concentrations from 0 to 360 NTU. The hyperspectral im-
ager collected the turbidity solution of each concentration. After that, the field UAV flight
Water 2022, 14, 128 13 of 15

5. Conclusions
In this study, the manual control experiment prepared twenty-nine gradients of tur-
bidity solutions with different concentrations from 0 to 360 NTU. The hyperspectral imager
collected the turbidity solution of each concentration. After that, the field UAV flight data
and simultaneous sampling were carried out to obtain the field turbidity value as the
verification data of this study. After analyzing the spectral characteristics and selecting
the optimal band to establish the model, the single band, band ratio, normalized ratio,
and PLS models were obtained, and R2 were 0.86, 0.92, 0.87, and 0.98, respectively. After
comparison, the band ratio and PLS model were selected to verify the accuracy.
According to the spectral analysis of turbidity solutions with different concentrations,
it can be concluded that the NIR band belongs to the sensitive band of water turbidity,
which is also consistent with the previous research conclusions. In addition, the logarithmic
form of the band ratio model and the PLS model R2 was higher, and the modeling accuracy
was 0.92 and 0.98, respectively, which can be used in the future for further verification
of sampling point accuracy. When combined with the field flight data and synchronous
sampling data, the model obtained by the PLS method had a higher accuracy after being
verified by sampling points, R2 is 0.72, and the band ratio model established by the red
band (707 nm) and green band (541 nm) is 0.65. It can be seen that the PLS method is
more accurate than the traditional modeling method. The modeling results of the artificial
control experiment of the hyperspectral imager can also be used to retrieve turbidity data
collected by the same instrument with high retrieval accuracy.
From the turbidity retrieval model established in this paper, the water quality pa-
rameters of small and medium-sized waters can be quickly monitored through UAV
hyperspectral technology, and the retrieval results are in line with the actual situation. At
the same time, the results of this study can also be used as the basis for the follow-up water
quality research of similar water bodies in the Liaohe estuary. In the future, the actual water
environment can be simulated by manual control experiments, such as adding a variety of
water quality parameters, configuring different solution standard solutions by adjusting
the concentration of various water quality parameters, collecting the corresponding hyper-
spectral information, and establishing the water quality retrieval model of a mixed water
environment. The model established by manual control makes the experiment closer to the
actual situation of a field water body. UAVs can obtain real-time image data with a high
spatial and temporal resolution, which can be used for the monitoring and early warning
of aquaculture and irrigation water pollution and has essential reference significance for
wetland environmental protection.

Author Contributions: This work presented here was carried out by collaborations among all authors.
Y.S. and M.C. conceived the idea. M.C., Y.S., C.H. and M.L. worked together on associated data
and carried out the experimental work. M.C. drafted the manuscript. C.H. and M.L. provided their
experience and co-wrote the paper with M.C. All authors have read and agreed to the published
version of the manuscript.
Funding: This research received no external funding.
Acknowledgments: This work is supported in part by the National Key Research and Development
Project of the 13th Five-Year Plan—Songliao River Lake Reservoir Water Resources Joint Regula-
tion Platform and Demonstration (2017YFC0406006), The 13th Five-Year National Key Research
and Development Program Project—River and Estuary Pollution Tracing and Control Planning
(2017YFC0406004), Beijing Outstanding Young Scientists Program (BJJWZYJH01201910028032) and
Remote Sensing Interpretation of Nanchang City in 2019.
Conflicts of Interest: The authors declare that they have no known competing financial interest or
personal relationships that could have appeared to influence the work reported in this paper.
Water 2022, 14, 128 14 of 15

References
1. Allam, M.; Khan, M.Y.A.; Meng, Q. Retrieval of Turbidity on a Spatio-Temporal Scale Using Landsat 8 SR: A Case Study of the
Ramganga River in the Ganges Basin, India. Appl. Sci. 2020, 10, 3702. [CrossRef]
2. Xiao, X.; Xu, J.; Zhao, D.Z.; Zhao, B.C.; Xu, J.; Cheng, X.J.; Li, G.Z. Research on Combined Remote Sensing Retrieval of Turbidity
for River Based on Domestic Satellite data. J. Yangtze River Sci. Res. Inst. 2021, 38, 128.
3. Yin, Q.; Gong, C.L.; Kuang, D.B.; Zhou, N.; Hu, Y.; Zhang, F.L.; Xu, W.D.; Ma, Y.Q. Method of Satellite Remote Sensing of Lake
Water Quality and Its Applications. J. Infrared Millim. Waves 2005, 24, 198–202.
4. Güttler, F.N.; Niculescu, S.; Gohin, F. Turbidity Retrieval and Monitoring of Danube Delta Waters Using Multi-Sensor Optical
Remote Sensing Data: An Integrated View from the Delta Plain Lakes to the Western-Northwestern Black Sea Coastal Zone.
Remote Sens. Environ. 2013, 132, 86–101. [CrossRef]
5. Li, S.J.; Wang, X.J. The Spectral Features Analysis and Quantitative Remote Sensing Advances of Inland Water Quality Parameters.
Geogr. Territ. Res. 2002, 18, 26.
6. Ai, Y.S.; Shen, Y.L. Measurement Uncertainty-Aware Quantitative Remote Sensing Inversion to Retrieve Suspended Matter
Concentration in Inland Water. Acta Opt. Sin. 2016, 36, 18–28.
7. Wang, X.P.; Zhang, F.; Abduwasit, G.Y.; Yu, H.; Ren, Y.; Wang, J.; Zhang, Y. The Relationship Between the Surface Water Quality
Indices and Hydrology of Ebinur Lake Watershed. Acta Sci. Circumstantiae 2017, 37, 900–909.
8. Sibanda, M.; Mutanga, O.; Chimonyo, V.G.P.; Clulow, A.D.; Shoko, C.; Mazvimavi, D.; Dube, T.; Mabhaudhi, T. Application of
Drone Technologies in Surface Water Resources Monitoring and Assessment: A Systematic Review of Progress, Challenges, and
Opportunities in the Global South. Drones 2021, 5, 84. [CrossRef]
9. Xu, P.; Du, P.; Shen, Q.; Xu, Z.B. Research on Remote Sensing Inversion Mode of Suspended Matter Density and Turbidity based
on GF-1 WFV Image Data in Hunhe River. J. Shenyang Norm. Univ. (Nat. Sci. Ed.) 2017, 35, 180–184.
10. Goodin, D.G.; Harrington, J.A.; Nellis, M.D.; Rundquist, D. Mapping Reservoir Turbidity Patterns Using SPOT-HRV Data.
Geocarto Int. 1996, 11, 71–78. [CrossRef]
11. Kornelia, A.W.; Robert, J.B.; Maria, O.; Marek, F. Investigation of Sediment-Rich Glacial Meltwater Plumes Using a High-
Resolution Multispectral Sensor Mounted on an Unmanned Aerial Vehicle. Water 2019, 11, 2405.
12. Jiang, Q.; Cao, Y.; Zhao, H.L.; Jiang, Y.Z.; Mao, W.S.; Zhu, Y.R. Modeling of Turbidity Retrieval of Hulunnaoer Lake Based on
Airborne Hyperspectral Imagery. South-North Water Transf. Water Sci. Technol. 2020, 111, 105–113.
13. Song, T.; Shi, J.Z.; Liu, Z.J.; Xu, W.Y.; Yan, F.; Xu, C.; Zhu, B.C. Research on Remote Sensing Quantitative Inversion Models of
Blue-Green Algae Density and Turbidity Based on Landsat-8 OLI Image Data in Lake Taihu. Saf. Environ. Eng. 2015, 22, 67–71.
14. Feng, Q.; Cheng, X.J.; Shen, X.; Xiao, X.; Wang, L.H.; Zhang, W. Inland Riverine Turbidity Estimation for Hanjiang River with
Landsat 8 OLI Image. Geomat. Inf. Sci. Wuhan Univ. 2017, 42, 643–647.
15. Wang, P.; Luo, X.W.; Zhou, Z.Y.; Zang, Y.; Hu, L. Key Technology for Remote Sensing Information Acquisition based on Micro
UAV. Trans. Chin. Soc. Agric. Eng. 2014, 30, 1–12.
16. Wang, F.; Wu, Y.D. Research and Application of UAS Borne Remote Sensing. Remote Sens. Inf. 2010, 2, 114–118.
17. Li, X.M.; Zhang, J.; Ma, Y.; Ren, G.B. Study on Monitoring Alien Invasive Species Spartina Alterniflora using Unmanned Aerial
Vehicle Hyperspectral Remote Sensing-a Case Study of the Yellow River Delta. Mar. Sci. 2017, 41, 98–107.
18. Shi, H.; Li, X.W.; Niu, Z.C.; Li, W.Z.; Wang, T.T.; Zhang, Y. Preliminary Study on Remote Sensing Information Extraction of Urban
Water Environment based on Micro UAV Images. Environ. Monit. China 2018, 34, 141–147.
19. Yang, Z.; Lu, X.P.; Wu, Y.B.; Miao, P.J.; Zhou, J.L. Retrieval and Model Construction of Water Quality Parameters for UAV
Hyperspectral Remote Sensing. Sci. Surv. Mapp. 2020, 267, 64–68, 99.
20. Chen, W.; Yan, L.; Gou, Z.Y.; Zhao, H.Y.; Liu, D.P.; Duan, Y.N. In-Flight Absolute Radiometric Calibration of UAV Multispectral
Sensor. Spectrosc. Spectr. Anal. 2012, 32, 3169–3174.
21. Mark, P.; Dmitry, B.; Kevin, G.; Felipe, G. UAVs, Hyperspectral Remote Sensing, and Machine Learning Revolutionizing Reef
Monitoring. Sensors 2018, 18, 2026.
22. McFeeters, S.K. The Use of the Normalized Difference Water Index (NDWI) in the Delineation of Open Water Features. Int. J.
Remote Sens. 1996, 17, 1425–1432. [CrossRef]
23. Heiden, U.; Segl, K.; Roessner, S.; Kaufmann, H. Determination of Robust Spectral Features for Identification of Urban Surface
Materials in Hyperspectral Remote Sensing Data. Remote Sens. Environ. 2007, 111, 537–552. [CrossRef]
24. Cheng, Y.M.; Wei, Y.C.; Zhang, J.; Sun, X.P.; Zhou, Y. A First Derivative Estimation Model of Chlorophyll-a Concentration in
Turbidity Water Based on Spectral Smoothing. Geogr. Geo-Inf. Sci. 2013, 29, 30–34.
25. Bhargava, D.S.; Mariam, D.W. Light Penetration Depth, Turbidity and Reflectance Related Relationship and Models. ISPRS J.
Photogramm. Remote Sens. 1991, 46, 217–230. [CrossRef]
26. Duan, H.T.; Wen, Y.; Zhang, B.; Song, K.S.; Wang, Z.M. Application Hyperspectral Data in Remote Sensing Inverse of Water
Quality Variables in Lake Chagan. J. Arid Land Resour. Environ. 2006, 20, 104–108.
27. Joshi, I.D.; D’Sa, E.J.; Osburn, C.L.; Bianchi, T.S. Turbidity in Apalachicola Bay, Florida from Landsat 5 TM and Field Data:
Seasonal Patterns and Response to Extreme Events. Remote Sens. 2017, 9, 367. [CrossRef]
28. Rodrigues, G.; Potes, M.; Costa, M.J.; Novais, M.H.; Morais, M.M. Temporal and Spatial Variations of Secchi Depth and Diffuse
Attenuation Coefficient from Sentinel-2 MSI over a Large Reservoir. Remote Sens. 2020, 12, 768. [CrossRef]
Water 2022, 14, 128 15 of 15

29. Li, S.J.; Wu, Q.; Wang, X.J.; Piao, X.Y.; Dai, Y.N. Correlations Between Reflectance Spectra and Contents of Chlorophyll-a in
Chaohu Lake. J. Lake Sci. 2002, 14, 228–234.
30. Shu, X.Z.; Yin, Q.; Kuang, D.B. Relationship between Algal Chlorophyll Concentration and Spectral Reflectance of Inland Water. J.
Remote Sens. 2000, 4, 41–45.
31. Liu, Y.; Wang, K.; Zhou, B.; Xu, H.W.; Shen, Z.Q. Preliminary Study on Hyperspectral Remote Sensing of Qiandao Lake
Chlorophyll-a Concentration. J. Zhejiang Univ. (Agric. Life Sci.) 2003, 29, 621–626.
32. Pulliainen, J.; Kallio, K.; Eloheimo, K.; Koponen, S.; Hallikainen, M. A Semi-Operative Approach to Lake Water Quality Retrieval
from Remote Sensing Data. Sci. Total Environ. 2001, 268, 79–93. [CrossRef]
33. Koponen, S.; Pulliainen, J.; Kallio, K.; Hallikainen, M. Lake Water Quality Classification with Airborne Hyperspectral Spectrometer
and Simulated MERIS Data. Remote Sens. Environ. 2002, 79, 51–59. [CrossRef]
34. Cao, Y.; Ye, Y.T.; Zhao, H.L.; Shi, Y.B.; Jiang, Y.Z. Remote-Sensing Retrieval Method of Suspended Solid Concentration and
Turbidity in Lakes and Reservoirs based on Discrete Particle Swarm and Partial Least Squares. J. Hydroelectr. Eng. 2015, 34, 77–87.
35. Bi, S.; Li, Y.M.; Lv, H.; Zhu, L.; Mu, M.; Lei, S.H.; Xu, J.; Wen, S.; Ding, X.L. Estimation of Chlorophyll-a Concentration in Lake
Erhai based on OLCI Data. J. Lake Sci. 2018, 30, 123–134.
36. Fu, H.J. Comparison of Lake Turbidity Inversion Based on GF-1 and Landsat-8 Satellite Images. Geomat. Spat. Inf. Technol. 2017,
40, 109–112.
37. Yin, C.; Tao, Y.Y.; Juan, X.Z.; Bo, Y.S.; Zhong, Y.J.; Li, H.Z. Quantitative Inversion Model of Hyperspectral for Turbidity in the
Nansi Lake. South-North Water Transf. Water Sci. Technol. 2015, 13, 883–887.
38. Yong, H.; Xiao, S.; Nie, P.; Tao, D.; Qu, F.; Lei, L. Research on the Optimum Water Content of Detecting Soil Nitrogen using Near
Infrared Sensor. Sensors 2017, 17, 2045.
39. Wang, J.Z.; Shi, T.Z.; Yu, D.L.; Teng, D.X.; Ge, X.Y.; Zhang, Z.P.; Yang, X.D.; Wang, H.X.; Wu, G.F. Ensemble Machine-Learning-
Based Framework for Estimating Total Nitrogen Concentration in Water Using Drone-Borne Hyperspectral Imagery of Emergent
Plants: A Case Study in an Arid Oasis, NW China. Environ. Pollut. 2020, 266, 115412. [CrossRef] [PubMed]
40. Liu, X.; Zhang, B.; Gao, L.R.; Chen, D.M. An Improved Noise Evaluation Algorithm of Hyperspectral Image Based on MNF. Sci.
China Press 2009, 39, 1305–1313.
41. Pal, M.K.; Porwal, A.; Rasmussen, T.M. Noise Reduction and Destriping Using Local Spatial Statistics and Quadratic Regression
from Hyperion Images. J. Appl. Remote Sens. 2020, 14, 1. [CrossRef]
42. Liu, Z.H.; Li, Y.M.; Lv, H.; Xu, Y.F.; Xu, X.; Huang, J.Z.; Tan, J.; Guo, Y.L. Inversion of Suspended Matter Concentration in Lake
Chaohu based on Partial Least-Squares Regression. J. Lake Sci. 2011, 23, 357–365.
43. Wang, H.W.; Wang, B.; Ji, T.; Xu, J.; Ju, F.; Wang, C.L. Simulation Estimation of BOD Content in Water Based on Hyperspectra.
Spectrosc. Spectr. Anal. 2021, 41, 978–983.