针对红外相机与可见光相机联合标定的问题, 利用不同物体红外辐出度的差异, 设计了可同时应用于红外相机与可见光相机标定的平面靶标。利用“最大稳定极值区域”(MSER)算法, 检测靶标的镂空深色区域。为克服图像可能存在较大畸变的问题, 通过相邻最大稳定极值区域质心的位置关系预估角点所在位置, 在预估位置内进行角点检测, 提取用于标定的亚像素角点位置。结果表明: 新型靶标及相应的角点提取方法能够同时满足于红外相机与可见光相机内外参数标定的需要。通过红外与可见光“双双目”立体视觉系统的融合重构效果可看出, 提供的标定数据能够满足系统需求。

To solve the problem of the co-calibration of infrared camera and visible band camera, the infrared radiation emittance difference of different substance is made use of to design a calibration target suitable to both cameras. The Maximally Stable Extremal Regions (MSER) algorithm is used to extract the dark areas in the calibration pattern. In order to overcome the problem of potential large distortion of the image, the position of the corner is estimated by the centroid of the adjacent MSERs. The corner detection is performed in the neighborhood of the estimated position to locate the sub-pixel corner position. The result shows that the new calibration target and the corresponding corner detection method can meet the needs of infrared camera and visible band camera calibration. It could be seen from the result of the “bi-binocular” stereo vision system reconstruction experiment that the calibration data provided by this method can meet the system demand.