Modified BIRCH and Hybrid Deep Framework for Tomato Leaf Disease Segmentation and Classification

Identifying disease at an early stage contribute s to improving productivity and quality of crops. Farmer's misidentification can lead to ineffective treatment strategy, adversely impacting tomato plants and agriculture. In this work, a MBIRCH algorithm is introduced for segmenting process along with hybrid model for classification process. Preprocessing, segmentation, extraction of features and classification are the four processing steps for categorizing leaf diseases on tomato crop. The input image is primarily under preprocessing, median filter is utilized in this stage. Consequently, the obtained image after preprocessing is fed to segmentation stage, where MBIRCH framework is introduced for proper segmentation process. Then various features, including MBP, GLCM, and shape -based features, are extracted from the segmented outcome. Finally, the hybrid classification system, which is a combination of BiLSTM and CNN, receives the obtained features as input. The hybrid model classifies tomato leaf diseases by calculating the averag e output of two classifiers. The classified tomato leaf diseases include bacterial spot, early blight, late blight, leaf mold, septoria leaf spot, two-spotted spider mite, target spot, tomat o mosaic viral, and tomato yellow leaf curl viral.