REMOTE SENSING OF PLANT CANOPY STRUCTURE WITH SIMPLE RADIATION MEASUREMENTS (MATHEMATICAL INVERSIONS, INDIRECT SENSING)

A technique for indirectly sensing plant canopy structure from simple, manageable measurements of direct sunlight penetrating the canopy is described. An integral relationship between the fraction of direct sunlight penetrating the canopy and canopy structural parameters is derived.; Three inversion techniques for solving the canopy integral equation are described. These include the Chahine non-linear iterative technique, the Twomey-Phillips least-squares constrained-linear inversion technique, and a previously unpublished method labeled the simple constrained-linear inversion technique. Each technique is thoroughly tested with simulated measurements which characterize data that would be taken in a wide range of actual plant canopies.; For testing the indirect sensing technique with actual measurements, a field study was conducted in two hybrids of corn. Measurements of sunlight penetration were obtained with a single silicon cell sensor (mounted to an HO gauge model railroad car) which horizontally traversed the canopies on twelve-meter sections of track. To evaluate the performance of the indirect sensing technique, direct measurements (using reliable yet time-consuming methods) of canopy structural parameters were taken.; From actual field measurements of sunlight penetration and from simulated data with random errors similar to the field data, the solution techniques were able to retrieve (in each case) the general shape of the leaf angle distribution. In addition in all cases the leaf area index was retrieved within 10% of the true value.; Sensitivity analyses indicate that the inversion process in the case of the canopy integral equation is very sensitive to the range of solar zenith angles over which data is collected. In addition, the length of track over which sunlight sampling is performed greatly effects the scatter in the data.; The performance evaluations of the three inversion (solution) techniques clearly show that the leaf area index and leaf angle distribution of horizontally homogeneous plant canopies can be satisfactorily derived from relatively simple measurements of sunlight passing through the canopy. On average the solution techniques performed equally well giving no clear reason to select one over the others.