Detecting and counting pistachios based on deep learning

Pistachios are nutritious nuts that sorted based on the shape of their shell into two categories: Open mouth and Closed mouth. The open-mouth pistachios higher in price, value, demand than closed-mouth pistachios. Because these differences, it is considerable for production companies to precisely count number each kind. This paper aims propose a new system counting different types with computer vision. We have introduced shared data set pistachios, including six videos total length 167 s 3927 labeled Unlike many other works, our model counts videos, not images. Counting objects need assigning object between video frames so be counted once. main challenges work existence pistachios’ occlusion deformation frames, because move roll transportation line may appear as closed some open frames. Our novel first trained RetinaNet detector network using detect After gathering detections, we apply them counter algorithm tracker assign consecutive high accuracy. able turn change appearance (e.g., look mouth) does incorrectly. performs very fast achieves good results. computed accuracy (9486 frames) 94.75%.