Depth Semantic Segmentation of Tobacco Planting Areas from Unmanned Aerial Vehicle Remote Sensing Images in Plateau Mountains

The tobacco in plateau mountains has the characteristics of fragmented planting, uneven growth, and mixed/interplanting crops. It is difficult to extract effective features using an object-oriented image analysis method accurately planting areas. To this end, advantage deep learning self-learning relied on paper. An accurate extraction areas based a semantic segmentation model from unmanned aerial vehicle (UAV) remote sensing images proposed Firstly, dataset established Labelme. Four models DeeplabV3+, PSPNet, SegNet, U-Net are used train sample data dataset. Among them, order reduce training time, MobileNet series lightweight networks replace original backbone four network models. Finally, predictive semantically segmented by trained networks, mean Intersection over Union (mIoU) evaluate accuracy. experimental results show that, perform 71 scene prediction images, mIoU obtained 0.9436, 0.9118, 0.9392, 0.9473, respectively, accuracy high. feasibility for extracting surface UAV been verified, research can provide reference subsequent automatic