A 2-D spatial gradient measurement technique for the localization of Indian license plate

Automatic license plate localization in realtime environment is a complex task, due to the variety of sizes and colors used in License Plate as well as due to the complexity of the background in different lighting conditions. This paper presents a recognition technique in which the number plate area is localized using a novel feature-based number plate localization algorithm which consists of three different LP segmentation algorithms, two of them based on different local adaptive binarisation on to grayscale versions of the color image and the third algorithm is directly applied onto the color image. Each of the three segmentation results is post-processed by four component reduction steps to reduce the false positives. The selection of the true license plate is performed by a qualification process applied to all remaining license plate candidates obtained after the component reduction of the three sequences. This paper mainly focus on the two fast algorithms — Edge Finding Method and Window Filtering Method for the better development of the number plate detection system. Experiments have been preformed to prove the robustness and accuracy of the approach. The experiment results show that almost 96.1% of input images are correctly localized on the average.