Research Of The License Plate Image Pre-Processing Based On Compressive Sensing And The Recognition

The License Plate Recognition System, especially Blurred License Plate RecognitionSystem is an important part and the core research topics of intelligent transportation andmonitoring system. The role of the intelligent transportation and monitoring system is tocalculate and statistical traffic flow, monitoring the illegal vehicles, as well as to track theseillegal vehicles and extract the basic information of them. For motor vehicles, the license plateis an important sign. Therefore, for the entire transportation and monitoring system, licenseplate recognition technology is a key technology, and the research of it has great practicalsignificance.In this paper, the License Plate Recognition System’s composition is briefly introduced.Several methods of image pre-processing, positioning, segmentation and recognition oflicense plate are described. The importance of blurred image pre-processing in the licenseplate recognition system was verified through simulation. Then several processing methods ofmotion blurred images are introduced. And the de-blurring performance of various algorithmswas simulated. Then the basic composition and content of the compressed sensing theory isintroduced. The compressed sensing image reconstruction algorithms are described in detail,and two algorithms introduced for the license plate image reconstruction were simulated.Finally, the compressed sensing de-blurring is researched. The model of TV norm based onthe compressed sensing for de-blurring is applied to the blurred license plate image restorationproblem, and the de-blurring and recovery simulation of the algorithm for the blurred licenseplate image was made. Then, the result of this algorithm’s simulation was compared with theresults of four other restoration algorithms’ simulations, and the quality evaluation of thesefive retored images, the results of the comparison and the evaluation show that the algorithmis feasible and has a good recovery performance. Finally, the simulations of the position,segmentation, as well as recognition of the de-burring image were made.In the License Plate Recognition System, the image pre-processing module is animportant component, and it is a prerequisite for the correct license plate recognition. Thelicense plate image processing, especially for the blurred license plate image, is the core andfocus of study in this paper. In this paper, based on the blurred image processing technology and license plate recognition technology, compressed sensing theory was introduced into theimage pre-processing module of the License Plate Recognition System. The compressedsensing theory was used to de-blur the license plate image. And the restoration results wereevaluated by the subjective evaluation method and an objective evaluation method. Theevaluation results showed that the details of the blurred and degraded images caused bymovement can be effectively recovered, and the quality of the retored images can be greatlyimproved.