Click on “Download PDF” for the PDF version or on the title for the HTML version.


If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.

Extracting picking point of single ripe tomato in clustering tomatoes based on deep learning and ellipse fitting

Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan www.asabe.org

Citation:  2021 ASABE Annual International Virtual Meeting  2100318.(doi:10.13031/aim.202100318)
Authors:   Zhichao Meng, Zenghong Ma, Pengcheng Wang, Leiying He, Xiaoqiang Du, Chuanyu Wu
Keywords:   Tomato, Ellipse fitting, Faster R-CNN, Picking point.

Abstract. The segmentation and localization of single ripe tomato from the image of clustering tomatoes is the key to the tomato picking manipulators. An extracting algorithm based on deep learning and ellipse fitting of tomato contour was proposed to realize the segmentation, recognization, and localization of single ripe tomato. Single ripe tomato area is segmented based on the Faster R-CNN algorithm first. The segmented single ripe tomato area is then extracted under the 6I color component of the YIQ color space. The ellipse is used to fit the contour of the tomato to find the tomato picking point. This algorithm determines the tomato picking point in the prediction box, which could significantly reduce the impact of tomatoes being occluded. The shape of the tomato is consistent with an ellipse, so using an ellipse to fit the tomato contour can obtain accurate picking points. Experimental results show that the algorithm has a success rate of over 89% in recognizing and locating the picking point of single ripe tomato.

(Download PDF)    (Export to EndNotes)