Detection, Quantification and Classification of Ripened Tomatoes: A Comparative Analysis of Image Processing and Machine Learning

In this paper, specifically for detection of ripe/unripe tomatoes with/without defects in the crop field, two distinct methods are described and compared. One is a machine learning approach, known as ‘Cascaded Object Detector’ and the other is a composition of traditional customized methods, individually known as ‘Colour Transformation’, ‘Colour Segmentation’ and ‘Circular Hough Transformation’. The (Viola Jones) Cascaded Object Detector generates ‘histogram of oriented gradient’ (HOG) features to detect tomatoes. For ripeness checking, the RGB mean is calculated with a set of rules. However, for traditional methods, color thresholding is applied to detect tomatoes either from a natural or solid background and RGB colour is adjusted to identify ripened tomatoes. In this work, Colour Segmentation is applied in the detection of tomatoes with defects, which has not previously been applied under machine learning techniques. The function modules of this algorithm are fed formatted images, captured by a camera mounted on a mobile robot. This robot was designed, built and operated in a tomato field to identify and quantify both green and ripened tomatoes as well as to detect damaged/blemished ones. This algorithm is shown to be optimally feasible for any micro-controller based miniature electronic devices in terms of its run time complexity of O(n3) for traditional method in best and average cases. Comparisons show that the accuracy of the machine learning method is 95%, better than that of the Colour Segmentation Method using MATLAB. This result is potentially significant for farmers in crop fields to identify the condition of tomatoes quickly