FRAMEWORK DEVELOPMENT FOR EXTRACTING STRUCTURAL PARAMETERS OF TREES

Abstract

Precision forestry, through advancements in remote sensing technologies like Light Detection and Ranging, has the potential to revolutionize tree monitoring and management practices. However, transitioning from traditional field-based methods to automated tree-level analysis presents several challenges. This study aims to address these challenges by developing methodologies for efficiently and accurately extracting tree structural parameters and creating a comprehensive framework for this purpose. The research explores data processing methodologies and develops an automated framework for extracting Diameter at Breast Height, tree height, and canopy area from 3D point clouds of trees. The study was conducted in a phased approach: the first phase took place in the mango orchards of Krishi Vigyan Kendra in Dhanauri, Uttarakhand, India, where the foundational framework was developed. In the second phase, conducted in Munich, Germany, the framework was tested and optimized using a large, labeled 3D point cloud dataset from TreeML-Data, which includes over 20 tree species. The results demonstrate a strong positive correlation between estimated and ground truth values for all three parameters. The analysis reveals minimal biases, which are insignificant compared to the overall performance. The mean absolute errors for Diameter at Breast Height (circle fitting & cylinder fitting), tree height, and tree canopy area were 0.0418 m, 0.0734 m, 0.53498 m, and 0.9147 m², respectively. These findings highlight the potential of the developed framework in improving the accuracy and efficiency of tree parameter extraction. Future research directions include investigating and developing robust methods for handling complex tree structures and diverse forest environments. The successful implementation of this automated tree parameter extraction framework has significant implications for precision forestry, enabling efficient, cost-effective, and scalable forest monitoring and management strategies.