A New Vehicle License Plate Correction Method Based on Sobel Operator and Priori Knowledge

摘要

The tilted visual angel of the camera determines that incline will inevitably occur in vehicle license plate (VLP) recognition system. This will bring about difficulties to the following character segmentation and character recognition. Traditional correction methods can receive high accuracy hardly and require a large amount of calculation consumption. In response to this situation, this paper proposes a tilt correction algorithm based on Sobel vertical edges of the plate image and some priori information. Sobel operator is applied for vertical edge detection of VLP images, and statistical edge information is used to determine the horizontal deflection angel. Vertical tilt correction uses a priori knowledge that the characters have minimum width when they stand vertically, adjust the angle of the individual character to determine the vertical tilt angle. The experimental results verify that this proposed method can be easily implemented, and can quickly and accurately get the tilt angle, it also has strong robustness.