Deciphering Bamboo Tree Diseases in Andaman and Nicobar Islands: Harnessing Image Processing and AI for Understanding Bamboo Diversity and its Socio-Ecological Impact

Abstract

This research investigates the diversity, ecological significance, and sustainable utilization of bamboo species across Andaman and Nicobar Islands. Focusing on their multifaceted roles, the study aims to elucidate the ecological, economic, and cultural dimensions of bamboo trees and also to identifies bamboo leaves diseases using image processing and artificial intelligence for technological support and management. As we know bamboo, renowned for its versatility and rapid growth, occupies a significant ecological niche in diverse ecosystems. This study endeavors to explore the extensive diversity of bamboo species, their ecological contributions, and the socio-economic impact of their utilization with bamboo leaves disease identification using advance technology with image processing. This research also involves analysing bamboo leaf samples using image processing techniques within MATLAB. Result displays the original images of these bamboo leaf’s affected by different diseases such as fungal Infection (leaf spot, rust), bacterial infection (leaf blight, masaic virus), insert infestations (bamboo mitees), environmental stressors (leaf burn), followed by their segmented output images. the proposed algorithm significantly enhances the detection accuracy, achieving an improved accuracy of 92.45% in the classification phase using the Minimum Distance Criterion with K-Means Clustering. In the subsequent phase, classification is performed using a Support Vector Machine (SVM) classifier. The SVM classifier demonstrates a high accuracy of 97.77% in classifying the diseases present in the bamboo leaves. These results clearly indicate that the proposed algorithm, particularly when combined with the SVM classifier, outperforms other approaches, significantly enhancing the detection accuracy of various diseases present in the bamboo leaves compared to previous methods or algorithms used for classification in plant leaves.