Canopy Volume Extraction of <i>Citrus reticulate</i> Blanco cv. Shatangju Trees Using UAV Image-Based Point Cloud Deep Learning

• Main Authors: Yuan Qi, Xuhua Dong, Pengchao Chen, Kyeong-Hwan Lee, Yubin Lan, Xiaoyang Lu, Ruichang Jia, Jizhong Deng, Yali Zhang
• Format: Article
• Language: English
• Published: MDPI AG 2021-08-01
• Series: Remote Sensing
• Subjects: canopy volume; UAV tilt photogrammetry; point cloud; deep learning; <i>Citrus reticulate</i> Blanco cv. Shatangju trees
• Online Access: https://www.mdpi.com/2072-4292/13/17/3437

Automatic acquisition of the canopy volume parameters of the <em>Citrus reticulate</em> Blanco cv. Shatangju tree is of great significance to precision management of the orchard.<strong> </strong>This research combined the point cloud deep learning algorithm with the volume calculation algorithm to segment the canopy of the <em>Citrus reticulate </em>Blanco cv. Shatangju trees. The 3D (Three-Dimensional) point cloud model of a <em>Citrus reticulate</em> Blanco cv. Shatangju orchard was generated using UAV tilt photogrammetry images. The segmentation effects of three deep learning models, PointNet++, MinkowskiNet and FPConv, on Shatangju trees and the ground were compared. The following three volume algorithms: convex hull by slices, voxel-based method and 3D convex hull were applied to calculate the volume of Shatangju trees. Model accuracy was evaluated using the coefficient of determination (R²) and Root Mean Square Error (RMSE). The results show that the overall accuracy of the MinkowskiNet model (94.57%) is higher than the other two models, which indicates the best segmentation effect. The 3D convex hull algorithm received the highest R² (0.8215) and the lowest RMSE (0.3186 m³) for the canopy volume calculation, which best reflects the real volume of <em>Citrus reticulate</em> Blanco cv. Shatangju trees. The proposed method is capable of rapid and automatic acquisition for the canopy volume of <em>Citrus reticulate</em> Blanco cv. Shatangju trees.