Development of Pear Cultivation Management Technology Using Information and Communications Technology for the Next Generation of Farmers (Part2)

—Development of a Water Sprout Estimation Algorithm—

Abstract

We developed “water sprout extraction” and “length estimation” methods based on 3D point cloud data. The tree extraction algorithm comprised four steps : voxelization, water sprout extraction, removal of leaf points and estimation of the length of water sprouts using Density Based Spatial Clustering of Applications with Noise (DBSCAN) and Random Sample Consensus (RANSAC). The measured point cloud data sets were used for the evaluation of the algorithm with 0.56 a as training data and 3.14 a as validation data. As a result, obtained using the proposed method demonstrated that the extraction accuracy of water sprouts was 97.3％, and the coefficient of determination （R²） between the measured and predicted data with regard to the total length of water sprout length was 0.99.