Visualization of Irregular Tree Canopy Centerline Data from a Depth Camera Based on an Optimized Spatial Straight-Line Fitting

We propose a new method to visualize the centerline of a single tree canopy based on a depth camera. Firstly, the depth camera captures the image of the target tree to obtain the 3D point cloud data, which is filtered and denoised. Then, we used the Poisson surface reconstruction method to reconstruct the 3D spatial surface of the point cloud data to restore the real scene accurately. In addition, we used Random Sampling Consensus (RANSAC) and Least Square Circle (LSC) in MATLAB software to fit circles to the 3D point cloud. We proposed a new spatial straight-line fitting method to visualize the centerline of the tree crown. The method has the advantage of no error in the spatial scattering Z-coordinate, and the fitted straight line is perpendicular to the xoy plane. The new method produces a smaller root mean square error (RMSE) than the traditional spatial straight line fitting method. This method can be effectively applied to practical applications such as tree crown pruning, providing accurate positional information for the positioning of tools during the pruning process. Ultimately, the pruning time can be shortened, and the accuracy of the pruning process can be improved.