Segmentation of Tomato Plant Leaf

Abstract

In today’s world there is a huge need for constant monitoring of crops, plants in agricultural fields to avoid diseases in plants. Since we cannot rely on the ability and accuracy of the human eye, it is only natural to depend on electronic equipment to detect diseases in crops and plants. Use of electronics for monitoring crops will help prevent plants from infection since this is the need of the hour, hence, making this an essential research paper. Most crops such as tomato, chilli, paddy etc. are attacked by bacteria, fungus or viruses leading to change in color, texture or function of a plant as it responds to pathogens. Common fungal infections include leaf rust, stem rust or white mold formation on the plant. Bacterial infections such as leaf spot with yellow halo, fruit spot, canker and crown gall all affect crops severely. In plants that are effected by viruses, one can find ring spots, pale green color in leaves and the plant stops growing and becomes distorted. We take these visual changes into account and process these images to identify whether a plant is healthy or not. There are three main steps for segmentation and identification of this disease.

• 1. Acquiring the RGB image and converting it into a suitable color domain such as HSV, YCbCr etc.

• 2. Mask the green pixels using a suitable threshold.

• 3. Choose a particular component in the chosen color domain after analyzing which component gives the most feasible result.