Construction of 3D landscape index based on oblique photogrammetry and its application in islands

Abstract

Abstract The landscape index is quantitative indicator to reflect the composition and spatial configuration characteristics of landscape ecological structures. Following the "P-C-L" frameworks, eight 3D landscape indexes were constructed to quantitatively describe the spatial landscape features and two sensitivity indexes were built to identify the differences between 2D and 3D. Based on two kinds of oblique photogrammetry data of Sichang Island and Tianheng Island, the results show that: 1) The shape index (TPSI) indicates the spatial shape of the patch scale. The TPSI of vegetation was generally higher than that of buildings, with a reasonable correlation (R2 = 0.698). The classification index (TCI) shows remarkable performance in patch-type identification. When the value of TCI approaches values of 100 or 33, the patch-type is probably building or vegetation, respectively, with a classification accuracy rate of 95% after verification. 2) The sensitivity indexes, GSC and ESC, provide an evaluation criterion for the attribute transformation from 2D to 3D. The dimensional change significantly affected the buildings and arbor, with a GSC of 6.697 and 2.306, respectively. The changes in low-rise ground features were not significant. On class and landscape scales, the dynamic ranges of all six 3D indexes increased compared to 2D indexes. The highest was TLSI (3D Landscape Shape Index), and the lowest was TSHEI (3D Shannon Evenness Index), with a growth rate of 349.65% and 0.3%, respectively. 3) The 3D landscape indexes can better feature the biomass and the intensity of human development and construction activities on all scales. Its combination with modern remote sensing and mapping technology can provide a more sound ecological assessment method for spatial planning of different ecosystems.