N. R. Deepa  and N. Nagarajan. Plant Leaf Disease Detection using Mean Possibilistic Fuzzy Local Information C-Means and Reweighted Linear Program Boost Classification Algorithms on Image Processing. Dynamic Systems and Applications 30 (2021) No.8, 1210 – 1227

https://doi.org/10.46719/dsa20213081

ABSTRACT.
Agriculture products quantity and quality are minimized by plant leaf diseases which may also causes economic and production losses. In large crop field’s monitoring, enhanced attention is gained by plant disease detection in recent days. In agricultural field, pesticides usage minimization is a major challenge and production rate’s quantity and quality enhancement is also a major issue.  The automatic detection methods are failed to enhance image quality in the existing system. And another major issue is that time complexity and accurate of the automatic segmentation disease. To overcome the above-mentioned issues, Mean Possibilistic Fuzzy Local Information C-Means (MPFLICM) segmentation and Reweighted Linear Program Boost Classification (RLPBC) approach is introduced in this work. These main steps of this work include preprocessing, segmentation, classification and feature extraction. Pre-processing by Kuan Filtering, segmentation by MPFLICM, feature extraction by GLCM and classification by RLPBC are done.  The pre-processing is a procedure that enhances the image data to eliminate unnecessary noise or improves few image features that hold significance in processing further. Image is partitioned as meaningful regions using segmentation process and this is a major process where extraction of image features is performed. Image has different features like shape, texture, color, Gray Level Co-Occurrence Matrices (GLCM) and grey level. Specified input data is classified as various groups and classes using classification process. It classifies the given images data whether diseased leaf or normal leaf based upon extracted features. The proposed RLPBC provides more accurate plant leaf detection results using optimized weighted values for the given plant leaf database. The experimental result concludes that the proposed MPFLICM- RLPBC technique provides better accuracy, precision and sensitivity metrics compare than existing techniques.

Keywords: Plant Leaf Disease, Segmentation, Mean Possibilistic Fuzzy Local Information C-Means (MPFLICM), Reweighted Linear Program Boost Classification (RLPBC).