In the multi-view and wide-angle imaging system, in order to guarantee the quality of subsequent image mosaic, the distorted image obtained from the wide-angle camera must correct. The linear characteristics in the image is utilized to correct the distortion and a bending measurement function is put forward with weight factor that gives each curve line the different weight values in accordance with the distance from the image center as the object function for obtaining the final distortion cofficients. The optimal distortion coefficients is obtained by minimizing the objective function. And based on the engineering examples, are used respectively the traditional distortion correction method based on linear characteristics and the proposed method to correct the target distort image. Experimental results show that the proposed method can only use single image to obtain high accuracy distortion calibration coefficient, under the condition that the noise level is less than 2 pixel, the root mean square error of the corresponding coordinates within 0.3 pixel.At the same time, the proposed method is simple, convenient and easy to implement.