Abstract

In past few decades, disease recognition (Powdery Mildew and Anthracnose) in Mango Crop has been an area of\nmajor concern due to deficiency of apparent, shape and texture features. Soft computing techniques are employed using\nimage processing to diagnose the diseases and hence to increase the yield of Mango crop. Here, we have proposed an\nimproved Wavelet-PCA based Statistical Feature Extraction scheme in order to minimize the issues arising in various\nmethodologies for diagnosis of pathological problems in different crops. Using this technique, we extract twenty statistical\nfeatures for different plant parts like flower and fruit in order to increase the yield of Mango crop. This research work is an\nadd-on to the research work presented in [1] by author. The features extracted here are used with Artificial Intelligence\ntechniques with the aim of diagnosis of both Anthracnose and Powdery Mildew disease occurring in form of black spots and\nFungus respectively on respective mango plant parts. The proposed research work has been implemented using MATLAB\n(MATrix LABoratory) software. Researcher captured 500 images each for flower and fruit using high quality Nikon 16MP\ndigital camera during Indian mango spring season i.e. from March to July from the mango orchids and The Agricultural\nUniversity situated at Dharwad district (Karnataka) location. The results obtained using proposed researches were found to be\nwith accuracy around 96.70% and 97.50% for flower and fruit respectively.