Extracting Leaf Area Index by Sunlit Foliage Component from Downward-Looking Digital Photography under Clear-Sky Conditions

摘要: The development of near-surface remote sensing requires the accurate extraction leaf area index (LAI) from networked digital cameras under all illumination conditions. widely used directional gap fraction model is more suitable for overcast conditions due to difficulty discriminate shaded foliage shadowed parts images acquired on sunny days. In this study, a new LAI method by sunlit component downward-looking photography clear-sky proposed. method, was extracted an automated image classification algorithm named LAB2, clumping estimated path length distribution-based LAD and G function were quantified leveled and, eventually, obtained introducing geometric-optical (GO) which can quantify proportion. proposed evaluated at YJP site, Canada, 3D realistic structural scene constructed based field measurements. Results suggest that LAB2 makes it possible processing with minimum overall accuracy 91.4%. widely-used finite-length tends underestimate index, while reduce relative error (RE) 7.8% 6.6%. Using lead underestimation (1.61; 55.9%), significantly outside requirement (0.5; 20%) Global Climate Observation System (GCOS). has RMSE 0.35 RE 11.4% conditions, meet GCOS. This relaxes required diffuse photography, be applied extract webcam images, expected regional continental scale monitoring vegetation dynamics validation satellite products.