A Calibration Method Based on the Distortion Correction by Neural Networks

Reducing distortion of the camera system should promise to improve the accuracy. In this paper we propose a new technique of reconstructing the camera property by using a neural network as learning mechanism. We show that relatively simple neural networks effectively eliminate the distortion for pin-hole camera model. We apply our method to synthesized image data with distortion in a simulation. By experiment, we verify that the performance is excellent.