A Novel Approach for Nonlinear Distortion Correction of Industrial Endoscope Images

Industrial endoscopes inspection is an important method of remote visual testing. The endoscope image suffers from a nonlinear spatial distortion due to the wide-angle design of the endoscope’s objective lens. The spatial distortion causes the pixels shift, which makes the object distorted in an image. By correcting the pixel’s offset in Cartesian coordinate, a novel approach for nonlinear distortion correction of industrial endoscope images is proposed in this paper. Based on a dot array calibration template correction method, cubic B-spline interpolating function is adopted to interpolate the curved faces of the pixel’s offset in Xand Yaxis separately. As a result, the offsets of each pixel are obtained in the distorted image. Furthermore, the pixels shift is rectified to achieve an undistorted image by coordinate conversion. A modified bilinear interpolation is used to evaluate the gray level of pixels, which is not assigned in coordinate conversion. The proposed method is applied to several endoscope images. The results show that the proposed method can correct the nonlinear distortion of endoscope images effectively.