matlab双目标定图,如何用matlab做双目视觉车距检测以及标定时遇到的问题

本帖最后由 lutianqi94 于 2014-4-15 10:54 编辑

本科生毕业设计，做的是基于双目视觉的前车距检测系统的设计，思路混乱，求高人指点啊，能给个具体做的流程么？？

还有我做第一步摄像机标定时，用的是手机自带的相机拍的，用matlab标定工具箱toolbox_calib标定后，fc误差特别大，请问是为什么啊？是由于设置的自动对焦的原因么？

标定结果如下：

Calibration parameters after initialization:

Focal Length:          fc = [ 1316.03687   1296.25593 ]

Principal point:       cc = [ 1295.50000   971.50000 ]

Skew:             alpha_c = [ 0.00000 ]   => angle of pixel = 90.00000 degrees

Distortion:            kc = [ 0.00000   0.00000   0.00000   0.00000   0.00000 ]

Main calibration optimization procedure - Number of images: 1

Gradient descent iterations: 1...2...3...4...5...6...7...8...9...10...11...12...13...14...15...16...17...18...19...20...done

Estimation of uncertainties...done

Calibration results after optimization (with uncertainties):

Focal Length:          fc = [ 1495.31339   1471.14565 ] ? [ 1389.87155   1363.65404 ]

Principal point:       cc = [ 1295.50000   971.50000 ] ? [ 0.00000   0.00000 ]

Skew:             alpha_c = [ 0.00000 ] ? [ 0.00000  ]   => angle of pixel axes = 90.00000 ? 0.00000 degrees

Distortion:            kc = [ 0.07057   -0.06404   0.00030   -0.00184  0.00000 ] ? [ 0.13141   0.23821   0.00143   0.00231  0.00000 ]

Pixel error:          err = [ 1.33118   1.23172 ]

Note: The numerical errors are approximately three times the standard deviations (for reference).

Recommendation: Some distortion coefficients are found equal to zero (within their uncertainties).

To reject them from the optimization set est_dist=[1;0;1;1;0] and run Calibration