A methodology for predicting the relative illumination intensity spatial distribution in apple canopy based on the three-dimensional point clouds

Abstract. In the stable growth stage, one of the evaluation standards is the high quality light area(LA) ratio of tree canopy in the fine management of orchards.To analyze the relative illumination intensity spatial distribution (RIISD) associated with the spring pruning in the apple tree canopy, a predicting methodology was proposed based on the three-dimensional(3D) point clouds. This methodology consists of three stages, i.e. data acquisition, feature extraction and classification. In this study, the canopy was divided into cell-grids regularly. The trimble tx8 and light intensity acquisition system were used to get the 3D point clouds and the RIISD of the apple tree canopy, respectively. In the feature extraction stage, a new Fractal Dimension (FD) based on 3D point clouds projection approach of the cell-grid was used to describe canopy structure in the spring. Further, the average RIISD of corresponding cell-grid was calculated. This combination of two different descriptors, which represented two features of a cell-grid, was utilized for subsequent classification( invalid LA or high quality LA)by employing Particle Swarm Optimization(PSO) tuned Support Vector Machines (SVM). In the field experiment, the cell-grid size was 0.4 mx0.4 m x 0.4 m. The experimental results showed that the classification prediction accuracy of the model was 77.03%，which indicated the good performance of the proposed method. The specify method proposed in this paper can make a contribution to the fruit quality management of apple orchard.