Skip to main content
Fig. 4 | Plant Methods

Fig. 4

From: Improved weed segmentation in UAV imagery of sorghum fields with a combined deblurring segmentation model

Fig. 4

Pixel-based classification results shown in a normalized confusion matrix in percent. Background is denoted by BG, sorghum by S, and weeds by W. a + b The performance of WeedSeg is degraded by motion blurred image patches. This is true for training a segmentation model on sharp and blurry image patches (scenario 1) or on sharp parches only (scenario 2) c DeBlurWeedSeg has a significantly better performance on motion blurred image patches and is only slightly worse on sharp image patches

Back to article page