License plate location based on multiple edge features

License plate location is a key step in the license plate recognition system. It is hard to position the plate region because of complex background, then the algorithm based on multiple edge features is proposed, we can get the coarse location by analyzing edge density, edge length and color feature of its corresponding pixel which is processed by color mapping. In the end, we can locate the plate area accurately by precise location. The algorithm not only avoids determining the threshold of binary segmentation, but also greatly reduces the number of color pixel which should be processed. Experimental results validate the high accuracy and the strong robustness.