Study on Single-Tree Segmentation of Chinese Fir Plantations Using Coupled Local Maximum and Height-Weighted Improved K-Means Algorithm

Abstract

Chinese fir (Cunninghamia lanceolata) is a major timber species in China, and obtaining and monitoring the parameters of Chinese fir plantations is of great practical significance. With the help of the K-means algorithm and UAV-LiDAR data, the efficiency of forestry surveys can be greatly improved. Considering that the traditional K-means algorithm is susceptible to the influence of initial cluster centers and outliers during the process of individual tree segmentation, it may result in incorrect segmentation. Therefore, this study proposes an improved K-means algorithm that uses the methods of local maxima and height weighting to optimize and improve the algorithm. The research results are as follows: (1) Compared to the traditional K-means algorithm, the producer accuracy and user accuracy of this research algorithm have imsproved by 10.72% and 11.46%, respectively, with significant differences (p < 0.05). (2) The research algorithm proposed in this study can adapt to Chinese fir plantations of different age groups, with average producer accuracy and user accuracy reaching 78.48% and 83.72%, respectively. In summary, this algorithm can be effectively applied to the forest parameter estimation of Chinese fir plantations and is of great significance for sustainable forest management.