Advancing Low Cost Mobile Remote Sensing Technologies For Forest Resource Management

ADVANCING LOW-COST MOBILE REMOTE SENSING TECHNOLOGIES FOR FOREST RESOURCE MANAGEMENT.

Advances in active sensor technology have made them more readily available than ever before. Light detection and ranging (LiDAR) technology has been used as a tool in forest management for several decades to model tree characteristics at both the stand and plot levels. From aerial mapping campaigns, to static terrestrial scanning, tree structure estimates derived from high density point clouds have proven to be accurate. Historical access to these technologies and services have often been prohibitively expensive, with terrestrial laser scanner costs exceeding six-figures and aerial campaigns costing tens-of-thousands of dollars, daily. This has limited their use in forestry to high-value, or short-rotation, species such as loblolly pine (Pinus taeda) across the American South and Eucalyptus in the tropics. In addition, the conifers across the western U.S. and Europe have been the subject of extensive LiDAR research and deployment. The mixed hardwood deciduous forests characteristic of south-central Pennsylvania present challenges to LiDAR mapping on multiple fronts. The complex upper canopy structure makes isolating individual trees from above difficult. Diverse and variable topography can present significant data capture and processing issues. In addition, present market values of eastern hardwoods limit the stakeholders ability to invest additional resources for advanced inventory techniques and technology. However, as the technology matures, it could allow for more advanced sampling techniques, such as 3P, at similar costs to present plot-based methods.Rapid expansion in the autonomous vehicle sector has created the need for affordable robotic vision sensors. The sensor of choice for many automotive manufacturers are small, affordable LiDAR units. Due to this, large capital investments are no longer required to rapidly collect dense point cloud data. Small, modular LiDAR sensors from companies such as Velodyne LiDAR can now be had for $4,000 USD. The availability of these sensors is only the first step to integrating them into natural resource management solutions and tools. This study has focused on standardizing both the protocols by which a low-cost laser scanner can be deployed in the field, how those data are collected and processed, and an evaluation of the individual stem detection and diameter estimates produced from the dense point clouds. Developing free and open source software solutions, processes and schematics was central to the aim of this study. Chapter 2 provides extensive documentation of system requirements, set-up parameters, embedded computer systems and required software to successfully capture LiDAR data in a variety of configurations. Significant effort was put forth to create a modular LiDAR system that can be deployed both aerially, and terrestrially, with minimal downtime between configurations. To assess the capacity of the system, it was deployed in both static and mobile terrestrial configurations to collect plot-level data in forested stands. Data were captured, processed and compared to in-situ measures taken from 7.5 m plots at three site locations. Estimates of stem diameter, taper, branching structure, and height are central measures to evaluating both biomass and potential market value of individual trees. The traditionally required field work to assess both value and volume are time- and labor-intensive. Two algorithms were evaluated for their ability to detect individual stems and estimate their diameter in the collected point clouds (Chapter 3). The LAS2Rings and TreeLS algorithms use the Hough Transform to identify stem locations and RANSAC fitting to estimate diameter. The algorithms were both able to successfully produce stem locations and diameters. However, due to the small n, statistically significant conclusions could not be drawn when comparing the in-situ and cloud measures. Other limitations were observed throughout the study and potential solutions were addressed to better process and capture data (Chapter 4).The technology presented and developed within this study shows considerable promise. Additional research is required to fully realize its potential as a widely adopted tool for natural resource land managers. This would include refinement of the capture methodology, optimization of the co-registration algorithms, and a substantial increase in in-situ measures for further statistical comparison.

National Conference on Environmental Problem-Solving with Geographic Information Systems

Asia-Pacific roadmap for innovative technologies in the forest sector

The preservation of forests, sustainable forest management (SFM), forest landscape restoration (FLR) and the need to make the most of precious forest resources are priority issues in the policy and sustainable development agenda of the Asia-Pacific region. Innovation will be key in the coming decades to meet the increasing demand for wood and other forest products while halting and reversing deforestation, in line with the commitment taken at COP26 in Glasgow by the international community. However, uptake of innovative technologies has been slow and uneven in the Asia-Pacific region, and there remains a gap between political commitments and the investments – in education, capacity building, and infrastructure development – required to put them into practice. This technical report examines the potential and barriers to disseminating and deploying innovative technologies for SFM in the region and provides overarching recommendations and specific options for decision-makers. It delineates and informs the process by which decision-makers and actors can identify: the potential of innovative technologies to advance SFM; their potential impacts; constraints to technology uptake and scaling up, and how to overcome these constraints and facilitate adoption.