Slant Correction of Vehicle License Plate Integrates Principal Component Analysis Based on Color-Pair Feature Pixels and Radon Transformation

Slant Correction plays an important role in the pretreatment during the recognition of vehicle license plate (VLP). To reduce and avoid the interference from the dirt, noise and frame of the VLP, as well as simplify the computation load, a method of slant correction of VLP integrates PCA (Principal Component Analysis) based on color-pair feature and radon transformation is presented in this paper. Three steps compose the method. The first step is to obtain the color-pair feature pixels of the image of VLP. The second step aims to seek the approximate slant angle of the plate by principal component analysis of the color-pair feature pixels. The final step is to achieve the further exact slant angle by radon transformation. The approach is implemented by program. And the experimental results demonstrate that this method is more precise and efficient than absolutely principal component analysis or radon transformation.