Fractal dimension relationships and values associated with certain plant canopies

This paper is divided into two parts. First, a new set of parameter relationships are presented, leading to a relatively reliable and simple method of calculating the shape fractal dimension (d(k)) associated with self-similar fractal images. The method is based on fast Fourier transforms of video images. The normalised Fourier amplitude moduli (mu) are manipulated to produce two measurable power indices, Delta, and sigma. We then show that d(k) = sigma Delta A characteristic amplitude (mu(0)) is also defined. Secondly, preliminary experimental evidence supporting the validity of the analysis is presented, using images from cultivated canopies of 10 different plant species, all of which are assumed to provide self-similar images. In addition, in anticipation of a dependence of numerical values of sigma, Delta, d(k) or mu(o) on species, a study of the discrimination potential between plant and weed species is also presented. Mean values and scatter associated with averaging over 10 sample images indicate distinction between species is possible, but further work to test the stability of numerical values and scatter bands is proposed. (C) 1997 Silsoe Research Institute.