“…1 The traditional image segmentation methods employ threshold-based segmentation techniques, weights, and super-pixel to extract edge information of the target objects for image segmentation. [2][3][4] For instance, Zhao et al 2 achieved tomato recognition based on threshold segmentation methods. However, this approach is susceptible to interference from lighting and environmental factors, thus exhibiting significant limitations.…”