Disease Segmentation in Groundnut Crop Leaves Using Image Processing Techniques

Groundnut is a cash crop that provides ample nutrition and vitamins with plenty of health benefits. It plays a vital contribution to economic growth. The primary disease which affects groundnut crop growth and cultivation is leaf spot. Several crop disease management approaches exist. But those are time-consuming, require human interventions, economically ineffective and less scalable. Therefore there is a need for an appropriate approach to resolve these things. Therefore image processing-based Leaf Spot Disease Detection in Groundnut Leaves (LSDDGNL) method is proposed in this work. In the first phase, groundnut leaf diseased images are collected, and colour space transformation is applied and analyzed. Through visual and statistical observation found that HSV colour space (threshold range 0 to 172, 0 to 255, 0 to 255) is most suitable to separate leaf spot disease regions efficiently. In the next phase, to enhance the segmentation result, the nonlinear noise reduction filter is analysed with different sigma values and the appropriate value (σ = 4) is fixed. It produced enhanced segmentation results with noise-less images. To check the generalization ability of the LSDDGNL, it is tested with tomato crop leaves by tuning the range of the hue colour model and the model provided the best disease deduction results.