Estimation of Tomato Berries based on Image Segmentation in RGB Color Model

Abstract

Estimating crop yield accurately remains important in dealing a farming enterprise successfully, enabling decisions concerning crop management, estimated delivery intervals and measures to consumers and creates a rating, to code but a little. Until ages, yield estimation is based on projections from labor-intensive counting and proportions of selected plants which are difficult in forecasting and management. In this paper, a method has been urbanized for precise estimation of a tomato plant by using color thresholding in MATLAB by means of RGB color model. The modules involved for the detection of red color in a captured image are the conversion of RGB image into gray scale image. A two-dimension black and white image is obtained by subtracting the two images and then a median filter is used to filter the noisy pixels. The binary digital images are labelled to identify connected components using bounding boxes and the metrics of the labelled regions are calculated for the counting number of tomatoes in a plant. Added, by analyzing RGB values for individual pixels in the captured image, the color of the individual pixels is renowned. The proposed prototype exhibited high aptitude in estimation of yield for each tomato crop.