Improving estimates of leaf area index by processing RAW images in upward-pointing-digital cameras

Leaf Area Index (LAI) measurement using upward-pointing digital camera in the forest floor has gained great attentions due to the feasibility of measuring LAI continuously at high accuracy. However, using upward-pointing digital camera could underestimate LAI when photos are exposed to excessive light conditions which make leaves near the sky in the photo disappeared. Processing RAW images could reduce possibility of LAI underestimation. This study aims to develop RAW image processing and compare RAW-derived LAI to JPEG-derived LAI. Digital photos have been automatically taken three times per day (0.5, 1, 1.5 hours before sunset) in both RAW and JPEG formats at Gwangreung deciduous and evergreen forests in South Korea. We used blue channel of RAW images to quantify gap fraction, then LAI. LAI estimates from JPEG and RAW images do not show substantial differences in the deciduous forest. However, LAI derived from RAW images at evergreen forest where forest floor is fairly dark even in daytime shows substantially less noise and greater values than JPEG-derived LAI. This study concludes that LAI estimates should be derived from RAW images for more accurate measurement of LAI.