COLOR CORRECTION FOR ORTHOPHOTOS GENERATED FORM MULTI-VIEW IMAGES

摘要

Orthoimagery has become a common data set in GIS applications. Many researches were focused on the geometric processing. To increase the application of orthoimages, it needs large coverage mosaicked with multi images. This is also valid for true-orthoimage composed by multi-view images, where hidden areas should be minimized. However, using multi-images produced orthophotoimages, especially for multi-temporal images, the color difference between images becomes an obstacle to cope with. Therefore, color balancing for multiple images is an important issue.This paper focuses on the color balancing for multi aerial images to produce orthoimages and true-orthoimages. The proposed scheme connects color for image patches globally and locally. The image control templates are obtained by back-projection from the object space for global color correction. We then get the color correction parameters with the gray value statistics for each template group. A local color correction is processed around seam lines for enhancing color continuity. Experimental results indicate that the proposed approach may reach quality results in radiometry.