Enhancing Tree Management Practices by Extracting GPR Attributes for the Evaluation of Tree Trunk Internal Structures

Effective forestry and urban park management, and disease detection strongly depend on tree trunks’ interior health conditions. At present, traditional destructive methods, such as coring, are used to analyse internally the tree structures. However, these are time-consuming, laborious, potentially harmful to the tree, and they only provide local information on the actual trunk conditions. In recent years, Ground Penetrating Radar (GPR) has been extensively employed as a non-destructive, fast, and cost-effective method to map internal structures of tree trunks [1]. The visual interpretation of conventional GPR maps generated by a common offset antenna array is frequently characterised by non-uniqueness and ambiguity. However, when utilising alternative GPR antenna arrays that can return information e.g., velocity, permittivity, and electrical conductivity in media, the data collection process becomes time-consuming. The importance of investigation time is often underscored in tree assessment surveys, especially across extensive areas. In this research, it is proposed to employ GPR attributes to enhance the interpretation of results achieved with common offset antenna array systems, specifically concerning tree trunks. Attribute analysis, a technique employed in seismic studies since the 1970s [2], is applied here to extract GPR attributes – i.e., quantities derived from GPR data and related with the characteristics of the tree trunk, such as moisture content and decay. To this purpose, the following sequential steps are followed: Data Acquisition: High-resolution GPR and Light Detection and Ranging (LiDAR) data are acquired by scanning the tree trunks. The GPR technique allows for the penetration of electromagnetic waves into the trunk, capturing reflections from internal structures. LiDAR can precisely locate GPR A-scans in their actual positions, rectifying GPR data distortion over complex tree trunk surface geometries [3]. Attribute Analysis: Attributes are extracted both before and after the GPR data processing, depending on their type. The extracted attributes are then correlated with the trunk properties, such as decay position and size, and moisture-related information. Validation: The outcomes of the proposed method are validated and assessed with the output of other conventional and non-destructive methods, including LiDAR observations made on test trees. The results of this study demonstrate a good correlation between the extracted attributes and the observed factors. Findings will have a substantial influence on the implementation of forestry and urban park management strategies, facilitating the adoption of well-informed decisions in the pursuit of sustainable forestry and conservation goals.   Keywords: Ground Penetrating Radar (GPR), tree management practices, tree trunk assessment, signal attribute analysis.   Acknowledgements This research was funded by the Vice-Chancellor’s PhD Scholarship at the University of West London.