To easily and efficiently correct image distortion by finding out mapping relation among a correction point in the three-dimensional(3D) space, a correction point projected onto a two-dimensional(2D) image plane and an irradiation angle by a neural network, storing the relational values in a table and correcting image distortion by using the table.
The system is provided with a camera 50, an angle adjusting device 60 for plane light, a video processing computer 70 for obtaining the angular information between plane light and a reference line and executing image processing, and a monitor 80 for outputting 3D distance information as plural hardware elements. In this case, mapping relation between the actual coordinate values of an optional correction point in the 3D coordinate space and the projected point on a 2D image plane corresponding to the correction point is found by supplementing the irradiation angle θ of irradiation light to the reference line. The found mapping function is converted into the connection strength of the neural network and stored in a table. After finding the distance of video information by inversely projecting an obtained camera image, a table value corresponding to the distance is inputted to the neural network to correct image distortion.
KOREA TELECOMMUN AUTHORITY
Next Patent: METHOD FOR IMAGE CONVERSION AND DEVICE THEREFOR