LAI MEASUREMENT WITH HEMISPHERICAL PHOTOGRAPHS AT VARIABLE CONDITIONS FOR ASSESSMENT OF REMOTELY SENSED ESTIMATIONS

摘要

The leaf area index (LAI) of a plant canopy is definedrnas its leaf area per unit of ground area. It is an importantrninput parameter for several studies related to water,rngases and energy exchange on the Earth surface, andrnseveral algorithms have been developed to estimate LAIrnfrom remotely sensed images. Nevertheless, in-siturnestimations are required for the development andrnvalidation of these algorithms, which is often the mostrncostly part of a project. Various source of errors duringrnacquisition or analysis of digital hemispherical canopyrnphotography have been reported, such as exposure andrnevenness of sky lighting. Several instructions have beenrnfound in literature, however some seem arbitrarilyrnrelated to the automatic exposure of the camera, andrnfurther investigation is needed to achieve standards. Thernaim of this work is to identify a procedure for in-siturnhemispherical lens photography and processing thatrnguarantees comparable measurements across variablernacquisition conditions, in order to be used as validationrnof remotely sensed LAI maps. The experimental designrnis based on the assumption that LAI should be constantrnin a site throughout the day, and in various cloudrnconditions. The effect of variable illuminationrnconditions has been examined with taking bihourlyrntime-series of hemispherical photos at the exact samernlocations, under the same canopy conditions. Tworncalibration techniques were tested to minimize the effectrnof clear vs. overcast sky. Finally, two computerrnsoftware specialized in hemispherical photos processingrnwere used. Results suggest that calibration using the skyrnlighting is a good means for repeatability ofrnobservations under closed canopies. Also, CAN-EYErnand HEMISFER software are capable of processingrnlarge numbers of hemispherical photographs accurately.rnThe resulting in-situ LAI measurements were used tornvalidate LAI maps from various satellites: ENVISATrnMERIS, Terra MODIS, and SPOT VEGETATION.