Abstract: The infrared remote sensing always plays an important role in the application of detecting urban virescence. But in the city, complex ground-objects, a mass of shadows caused by high-buildings and a large number of color ground-objects form disturbing sources for extracting vegetation information. To solve above problems, Retinex and the chromatics are discussed and studied with colorimetry. A processing method is designed based on the idea of better maneuverability and low complexity to resist disturbance and to extract vegetation information from infrared remote sensing images.

Key words: computer application, image processing, shadow removal, Retinex clustering, infrared remote sensing