Wavelet Based Feature Extraction for Rice Plant Disease Detection and Classification

Plant disease is a significant intimidation to both plant diversity and food production, which claims monitoring mechanisms to take the adequate measures for confirming food production, an important sustainable management strategy. In Bangladesh rice in considered as one of the dominant food crop, as most of agricultural lands are used for cultivating rice and it is increasing every year. As production and quality of rice plays an important role in economy and food supply, diagnosis of rice diseases at betimes stage is momentous. The goal of our proposed work is detection and classification of rice plant diseases using simple image processing approach based on wavelet transform. This approach uses the multi-resolutional analysis, Discrete Wavelet Transform (DWT) to perform decomposition of the input image into horizontal, vertical, diagonal sub-bands. Different color and texture features are extracted from horizontal, vertical, diagonal sub-bands of the input image up to 2 levels and considered as features. Classification has been done by using an Ensemble of Linear Classifier adopting the Random Subspace Method (RSM). This proposed technique has successfully classified four prime rice plant diseases named as rice bacterial blight, rice brown spot, rice bacterial sheath brown rot and rice blast with classification accuracy rate of 95 percent.