HDR image fusion method based on multi-exposure vehicle-mounted image

The invention discloses an HDR image fusion method based on a multi-exposure vehicle-mounted image. Three cameras are used for simultaneously obtaining three driving images with exposure of-1EV, 0EV and + 1EV; then obtaining internal and external parameters of each camera by using a Zhang Zhengyou calibration method, and correcting the shape vehicle image by using a Bouguet stereo correction algorithm; thirdly, carrying out stereo matching on the three corrected driving images by using an HAD-Census algorithm to obtain a disparity map image; and finally, based on a Laplacian pyramid image decomposition and fusion algorithm, decomposing and fusing the corrected image and the parallax image to obtain an HDR fusion image. The invention aims to solve the problems that an intelligent automobile image recognition system is easy to generate misrecognition, missed recognition and the like in driving environments with high-intensity backlighting, weak light and great light change, so that an intelligent automobile can