Identification and Yield Estimation of Mature Fruits Using Modified Watershed Algorithm

Identification of fruit and determining the location is a time consuming complex task. Many works have been proposed to determine the fruit. In our work we have proposed a framework that segments the fruit from the background, determines the location of the fruit and counts the fruit that helps in yield estimation.Similarity index is determined that measures the correlation between the segmented output and the ground truth image. The overlapping fruits are identified their contours are detected. Further fruit count algorithm is applied. The similarity measure ranges between 71.5 % - 92.4 % and highest accuracy of fruit count is observed for Apples and Tomato fruit images which is 97.5%.