Fig. 8From: DeepSeedling: deep convolutional network and Kalman filter for plant seedling detection and counting in the fieldExample images in the TAMU2015, UGA2015, and UGA2018 datasets. For seedling detection, the TAMU2015 dataset shows challenges of high object occlusion and existence of monocotyledonous weed and the UGA2015 dataset shows extreme illumination condition and existence of dicotyledonous weed. On the contrary, the UGA2018 dataset demonstrates a relatively simple and ideal situation for seedling detectionBack to article page