Fig. 1From: DeepFlower: a deep learning-based approach to characterize flowering patterns of cotton plants in the fieldRepresentative results of plants and emerging blooms detected by the trained Faster RCNN (FrRCNN5-cls) model. The top three rows demonstrate successful detections under different illumination, bloom load, and occlusion conditions. The bottom row shows two failed cases of emerging bloom detection, one because back-sided leaves had higher reflectance and were identified incorrectly as emerging blooms, and the other because a lower contrast between emerging blooms and the background could lead to mis-detectionBack to article page