Asparagus Plumosus

Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

Fractals, Vol. 10, No.

4 (2002) 429–434

c World Scientific Publishing Company

FRACTAL DIMENSION AND


SELF-SIMILARITY IN
ASPARAGUS PLUMOSUS

J. R. CASTREJÓN PITA
Centro de Investigación en Energı́a UNAM
Ap. Postal 34, 62580 Temixco, Morelos, México
A. SARMIENTO GALÁN
Instituto de Matemáticas UNAM, Av. Universidad s/n
62200 Chamilpa, Morelos, México
R. CASTREJÓN GARCÍA
Instituto de Investigaciones Eléctricas
Av. Reforma 113, 62490 Temixco, Morelos, México

Received March 16, 2002; Accepted June 14, 2002

Abstract
We measure the fractal dimension of an African plant that is widely cultivated as an ornamental
— the Asparagus plumosus. This plant presents self-similarity, remarkable in at least two
different scalings. In the following, we present the results obtained by analyzing this plant via
the box counting method for three different scalings. We show in a quantitative way that this
species is a fractal.

Keywords: Self-Similarity; Multifractality; Nature Dynamics.

1. INTRODUCTION ric bifurcations,1 encouraged us to analyze the As-


paragus plumosus.2 This plant is a native of Africa,
Nowadays, it is frequent to use computational al- but often cultivated in the rest of the world as an
gorithms in order to produce images of plants ornament. The plant can be easily identified: it
and trees that resemble their natural counterparts. is semi-climbing, has a typical height of 2 m, its
These visualizations, which present several symmet- main branches measure from 25 to 50 cm, and all

429
430 J. R. Castrejón Pita et al.

Fig. 1 A typical example of a main branch of Asparagus Fig. 2 An atypical main branch of Asparagus plumosus,
plumosus. note the differences in shape with respect to the usual
branches in Figs. 1 and 3.

branches have philiform divisions; its flowers are


white and have six petals each, their fruits are pur- intersect any part of an image that has been placed
ple spheres, 7 mm in diameter. Observed in some over it. In order to calculate the fractal dimension
detail (Figs. 1 to 3), the “leaves” of this plant con- of the image, denoted by D, using a square grid
sist of repeated bifurcations from the main stem, of side size given by ε, one needs to analyze the
showing a high degree of both symmetry and scal- changes in the number of boxes required to cover
ing; these branching can also be observed even at the image, N , as the size of the grid is reduced, i.e.
the smallest scale. Two other peculiar character- log N (ε)
istics of the “leaves” of this plant are their flat- D = lim . (1)
ε→0 log 1/ε
ness and their uniform green color. Although the
We have applied this method to the three branches
branches may be dramatically different in shape
shown in Figs. 1 to 3 at three different levels:
(actually, Fig. 2 shows an atypical branch), we will
the three different scales at which symmetry is
show that their fractal dimension is the same.
observed. In Figs. 3 and 4, we visually exem-
plify the application of the box counting method.
2. METHOD The “leaves” we have designated as medium-size
branches correspond to the ramifications at the
The method of box counting is widely known.3 lower right corner of the branches in Figs. 1 to
Briefly, the box counting technique consists of 3, and those so-called small-size branches were se-
counting the number of boxes in a grid that lected from the medium size ones following the same
Fractal Dimension in Asparagus plumosus 431

Fig. 3 Third example of a main branch; a grid of boxes with a side-length of 60 pixels is also shown (those boxes that have
an intersection with the image are shaded in gray).

Table 1

Fractal Dimension (D) Uncertainty (∆D) Linear Regression (R)

Branch Main Med Small Main Med Small Main Med Small
Fig. 1 1.742 1.712 1.825 0.003 0.003 0.005 0.999 0.999 0.999
Fig. 2 1.787 1.765 1.869 0.002 0.002 0.005 0.999 0.999 0.999
Fig. 3 1.760 1.722 1.819 0.002 0.003 0.006 0.999 0.999 0.998

criteria; Fig. 4 exemplifies the selection for the main small one. Since our images all have well-defined
branch in Fig. 3. All the images were obtained by borders, there is no need to analyze the contour
positioning the corresponding branch directly on threshold.4
a scanner (640 × 460 resolution, bitmap images),
and since the leaves are objects immersed in a two-
dimensional space, it was not necessary to use any 3. RESULTS
kind of projection. The digital scanning was made
in black and white, and in real scale. The side size The values of N , obtained varying the grid size from
of the square grid was varied from 1 to 200 pixels, by 1 to 200 pixels, are shown in Figs. 5 to 7. This
steps of 1 pixel. The original size of the main branch pixel range allows for a direct comparison in real
in Fig. 3 is 428.8 × 492.0 mm, 127.5 × 220.6 mm for scale of the results for the three levels at which
the medium-size branch and 23.0 × 66.9 mm for the similarity is observable. A bigger side-size box is
432 J. R. Castrejón Pita et al.

Fig. 4 Medium- and small-size branches lying on a square grid with side-length of 20 pixels.

Fig. 5 Symbols represent the results of applying the box counting method to the main branches in Figs. 1 to 3. Straight
lines show linear regressions performed for each data set.
Fractal Dimension in Asparagus plumosus 433

Fig. 6 Analogue of Fig. 5 for medium-size branches.

Fig. 7 Analogue of Fig. 5 for small-size branches.

not used because the width of the smallest branches are linear over a wide range of ε values, the fractal
(at the base) is 200 pixels, and therefore, a bigger dimension D is then given by the slope of the cor-
side-size box would mean that a single box would responding line (see Figs. 5 to 7). Finally, the val-
almost cover the whole branch. Since the relations ues obtained for the fractal dimension of the three
434 J. R. Castrejón Pita et al.

branches and at the three different scales, are shown Unfortunately, we do not seem to find the same self-
in Table 1 together with the uncertainty in the slope similarity at the smallest scale.
(∆D) and the correlation of the linear regression
(R).
ACKNOWLEDGMENT

4. CONCLUSION Martin Nezadal and Oldrich Zmeskal (Institute of


Physical and Applied Chemistry at Brno, Czech Re-
From the previous analysis, where we have shown public) are gratefully aknowledged for their HarFA
that the fractal dimension of the three branches program. This work has been partially supported
is practically the same, we can conclude that the by DGAPA-UNAM (IN101100), and UC-MEXUS.
shape of a branch of Asparagus plumosus is irrel-
evant for the determination of its fractal dimen-
sion. The very small uncertainties in these values REFERENCES
(∆D/D < 3 × 10−3 ) can be easily interpreted in
1. B. B. Mandelbrot, The Fractal Geometry of Nature
terms of the high linear correlations shown in
(Freeman, San Francisco, 1989).
Table 1. Accordingly, we can confirm the fractal
2. M. Martı́nez, Catálogo de Nombres Vulgares y
dimension in this species, a new type of natural Cientı́ficos de Plantas Mexicanas (Fondo de Cultura
fractal being added to the extensive already well- Económica, México, 1979).
known gallery (for a recent man-made example, see 3. K. T. Alligood, T. D. Sauer and J. A. Yorke, Chaos:
Rodin and Rodina5 ). Additionally, since the value An Introduction to Dynamical Systems (Springer
of the fractal dimension obtained from the analysis Press, New York, 1996), pp. 172–180.
of the two bigger scales is indeed very similar, we 4. R. R. Prasad and K. R. Sreenivasan, Phys. Fluids
can conclude that there is the same level of com- A2, 5 (1990).
plexity at these two scales: the plant is self-similar. 5. V. Rodin and E. Rodina, Fractals 8, 4 (2000).

You might also like