【摘要】 基于多摄像机的视觉应用中通常假设统一的颜色响应。但是当摄像机之间存在较大的成像特性差异或光照变化时,所获得的立体图对间就会出现严重的色彩差别。这种差别会导致后续颜色匹配的不准确,并进一步影响立体视觉算法的性能。为了解决这个问题,提出了一个鲁棒的基于图像分割和特征点匹配的颜色校正算法。和传统的全局校正或者有参考物体的校正算法不同,提出了一种基于区域的校正算法。该方法不仅避免了全局校正算法无法满足局部需求的矛盾,同时也摆脱了设置参考物体的复杂和低效。大量实验证明了所提出算法的有效性和鲁棒性。更多还原

【Abstract】 Most multi-camera vision applications assume a single common color response for all cameras. However,significant luminance and chrominance discrepancies among different camera views often exist due to the dissimilar imaging characteristics of different cameras and the variation of lighting conditions. These discrepancies may severely affect the algorithms that depend on the color correspondence. To address this problem,a robust color calibration algorithm is proposed based on image segmentation and keypoint matching. Instead of handling the image as a whole or employing a color calibration object,the color discrepancies is compensated region by region. The proposed algorithm can avoid the problem that the global calibration does not meet the needs of local areas. Many experiments have been done to prove the effectiveness and the robustness of our algorithm.更多还原