Accurate detection of land cover change using remotely sensed images requires that multitemporal image sets be aligned both radiometrically and geometrically. A lack of proper radiometric and/or geometric alignment will result in errors of omission and commission in change detection products. Radiometric normalization is performed by correcting and/or standardizing multiple sun, scene, sensor, and atmospheric effects, which may cause spurious differences in the magnitude of reflected light measured by remote sensing instruments between multitemporal acquisitions. Geometric registration between multitemporal image sets is commonly performed using polynomial warping algorithms and/or orthorectification procedures. However, proper radiometric and geometric calibration between high spatial resolution airborne and satellite image sets is often difficult to attain and may be time or cost prohibitive

Radiometric Processing

Four primary radiometric correction procedures were investigated: 1) anisotropic reflectance correction, 2) correction of terrain-induced differences in reflectance magnitude, 3) radiometric normalization between multitemporal image sets, and 4) normalization of reflectance magnitude differences associated with vegetation phenological state between multitemporal acquisitions. For a brief explination od each, please click on the the links below. For more specific information on theses procedures please refer to the project report (PDF).

ANISOTROPIC REFLECTANCE CORRECTION

MULTI TEMPORAL RADIOMETRIC NORMALIZATION

Accurate georeferencing and registering of multidate, high spatial resolution imagery acquired from aerial platforms over areas of extreme relief is critical to change detection, yet is difficult and often unattainable using standard orthorectification or polynomial transformation algorithms. This is the case because polynomial warping models cannot account for the variable geometric distortions caused by terrain and commonly available digital elevation models (DEMs) lack the spatial resolution and elevation accuracy to properly orthorectify high spatial resolution imagery (e.g., 1 m).

The project team developed and tested a method which enables precise spatial registration between multitemporal, high resolution aerial imagery. The approach is referred to as frame center matching for mutltitemporal image registration. To learn more about frame center matching please click on the link below.

FRAME CENTER MATCHING