Recognition and Quantification of Area Damaged by Oligonychus Perseae in Avocado Leaves

The measure of leaf damage is a basic tool in plant epidemiology research. Measuring the area of a great number of leaves is subjective and time consuming. We investigate the use of machine learning approaches for the objective segmentation and quantification of leaf area damaged by mites in avocado leaves. After extraction of the leaf veins, pixels are labeled with a look-up table generated using a Support Vector Machine with a polynomial kernel of degree 3, on the chrominance components of YCrCb color space. Spatial information is included in the segmentation process by rating the degree of membership to a certain class and the homogeneity of the classified region. Results are presented on real images with different degrees of damage.

[1]  Bernt Schiele,et al.  Comprehensive Colour Image Normalization , 1998, ECCV.

[2]  Danielle Dennis,et al.  Digital image analysis of Zostera marina leaf injury , 2008 .

[3]  Pierre Soille,et al.  Morphological image analysis applied to crop field mapping , 2000, Image Vis. Comput..

[4]  Yunyoung Nam,et al.  A similarity-based leaf image retrieval scheme: Joining shape and venation features , 2008, Comput. Vis. Image Underst..

[5]  E. M. Bakr,et al.  A new software for measuring leaf area, and area damaged by Tetranychus urticae Koch , 2005 .

[6]  ChengLiang Wang,et al.  A Leaf Vein Extraction Method Based On Snakes Technique , 2005, 2005 International Conference on Neural Networks and Brain.

[7]  T. Hsiang,et al.  Quantifying Fungal Infection of Plant Leaves by Digital Image Analysis Using Scion Image Software , 2022 .

[8]  I. Kale,et al.  A Colour Normalization Method for Giemsa-Stained Blood Cell Images , 2006, 2006 IEEE 14th Signal Processing and Communications Applications.

[9]  Paolo Remagnino,et al.  Venation Pattern Analysis of Leaf Images , 2006, ISVC.

[10]  Yan Li,et al.  Leaf Vein Extraction Using Independent Component Analysis , 2006, 2006 IEEE International Conference on Systems, Man and Cybernetics.

[11]  N. Otsu.  A threshold selection method from gray level histograms , 1979 .

[13]  David H. Wolpert,et al.  No free lunch theorems for optimization , 1997, IEEE Trans. Evol. Comput..

[14]  Hong Fu,et al.  A two-stage approach for leaf vein extraction , 2003, International Conference on Neural Networks and Signal Processing, 2003. Proceedings of the 2003.

[15]  M. Hoddle,et al.  Measuring Mite Feeding Damage on Avocado Leaves with Automated Image Analysis Software , 1999 .

[16]  Bernd Neumann,et al.  Computer Vision — ECCV’98 , 1998, Lecture Notes in Computer Science.