• The application of deep convolutional neural network to automated fruit tree recognition in aerial imagery

  • 2019 -10 -10
Agricultural statistics of crop distribution is essential to production monitoring and regulation. Here, we applied deep convolutional neural network (DCNN) based models to categorize fruit trees using aerial imagery. Before building the DCNN model, aerial photos were segmented into several thousands of land parcels using labeled shapefiles. Specifically, we applied advanced DCNN models to categorize the fruit tree aerial images and to do semantic segmentation for the irregular-shaped land parcels. Using the standard of IoU (Intersection over Union) greater than 0.5, the precision could reach 98%. With these DCNN models, monitoring the plantation coverage and estimating total production could be more convenient and instant in the near future. It could facilitate produce production control and coordinate the supply and marketing.