双目摄像机标定与特征点匹配方法的研究

传统标定方法具有实验环境要求高、结果精确度低、鲁棒性差等显著缺点,为此提出一种基于软件编程的主动视觉标定方法。通过建立摄像机针孔成像线性模型,考虑畸变影响的非线性模型矫正,推导出双目摄像机标定参数,并应用软件编程实现摄相机标定,获得精度高、鲁棒性较好的结果。同时,针对当前已有的图像匹配算法,对图像尺度变化适应性较差等问题,以Harris算子为基础,通过增加图像变换尺度参数,得到具有尺度不变性的改进算法。基于软件编程实现特征点坐标提取与匹配,结果显示改进后算法耗时变短、计算结果精度和鲁棒性均有所提高,为深入研究双目摄像机障碍物识别与路径规划奠定了基础。

Binocular Camera Calibration and Feature-Point Matching Method

Traditional calibration methods have significant shortcomings such as high experimental environment requirements,low results accuracy,and poor robustness.Therefore,an active vision calibration method based on software programming is proposed.By establishing a linear model of the camera’s pinhole imaging and taking into account the non-linear correction of the distortion,the calibration parameters of the binocular camera are derived,and then the software programming is used to implement the camera calibration,thus to achieving high accuracy and robustness.At the same time,aiming at the problem that the existing image matching algorithms have poor adaptability to image scale changes,based on Harris operator,by increasing the image transformation scale parameters,a modified algorithm with unchanged scale is obtained.Feature coordinates are extracted and matched based on software programming.The results indicate that the modified algorithm takes less time and improves the accuracy and robustness of the calculation results.It lays the foundation for further research on obstacle recognition and path planning of the binocular camera.