Fig. 9From: DeepSeedling: deep convolutional network and Kalman filter for plant seedling detection and counting in the fieldFlowchart of the deep convolutional network based approach for cotton seedling detection and counting. After preprocessing, video frames are fed into a trained Faster RCNN model for seedling detection. Detected seedlings are associated using a Kalman filter-based video tracking algorithm, so the same plant object will be tracked through video frames and not counted repeatedlyBack to article page