Forest 3D Reconstruction and Individual Tree Parameter Extraction Combining Close-Range Photo Enhancement and Feature Matching

An efficient and accurate forest sample plot survey is of great significance to understand the current status resources at stand or regional scale basis scientific management. Close-range photogrammetry (CRP) technology can easily quickly collect sequence images with high overlapping reconstruct 3D model scenes extract individual tree parameters automatically and, therefore, greatly improve efficiency investigation has application potential in forestry visualization However, it some issues practical applications. First, imaging quality affected by illumination forest, resulting difficulty feature matching low accuracy parameter extraction. Second, reconstruction limited under complex understory vegetation topographic situation forest. In addition, density point clouds dense directly affects This research collected plots four species smartphones Gaofeng Forest Farm Guangxi Wangyedian Mongolia analyze effects image enhancement, detection cloud algorithms on extraction, then proposed a strategy extraction suitable for different scenes. we compared enhancement median–Gaussian (MG) filtering, single-scale retinex (SSR) multi-scale (MSR) filtering algorithms. Then, an improved algorithm combining Harris corner speeded-up robust features (SURF) (Harris+SURF) proposed, effect that invariant transform (SIFT) operator. Third, according morphological characteristics trees images, used iterative interpolation planar triangulation network based geometric constraints (GC-based IIPTN) increase model, position DBH trees. The results show MSR significantly number matched pairs. Harris+SURF method reduce time GC-based IIPTN extracted same as measured bias within 0.2 m. Eucalyptus grandis, Taxus chinensis, Larix gmelinii Pinus tabuliformis 94%, 95%, 96% 90%, respectively, which proves also provides effective support resource management future.