New algorithm for biological objects' shape evaluation and data reduction

• Main Authors: Stanislav Bartoň, Libor Severa, Jaroslav Buchar
• Format: Article
• Language: English
• Published: Mendel University Press 2010-01-01
• Series: Acta Universitatis Agriculturae et Silviculturae Mendelianae Brunensis
• Subjects: image processing; data reduction; least square method
• Online Access: https://acta.mendelu.cz/58/1/0013/

The paper presents the software procedure (using MAPLE 11) intended for considerable reduction of digital image data set to more easily treatable extent. The example with image of peach stone is presented. Peach stone, displayed on the digital photo, was represented as a polygon described by the coordinates of the pixels creating its perimeter. The photos taken in high resolution (and corresponding data sets) contain coordinates of thousands of pixels - polygon's vertexes. Presented approach substitutes this polygon by the new one, where smaller number of vertexes is used. The task is solved by use of adapted least squares method. The presented algorithm enables reduction of number of vertexes to 10 % of its original extent with acceptable accuracy +/− one pixel (distance between initial and final polygon). The procedure can be used for processing of similar types of 2D images and acceleration of following computations.