Vehicle Image Processing Technology Study Based On Image Recognition

As the development of the computer vision technology, Application of CCD camera to capture moments of objects can provide more and more high-quality images, The application of image recognition in the field of vehicles becomes more and more extensive, its main applications is in intelligent transportation and vehicle performance parameters detection. The main examples are in License Plate Recognition and Vehicle Type Recognition facts. In the License Plate Recognition system, followed by the use of license plate location, characters and the division of character recognition for image processing; In the Vehicle Type Recognition system, the identification process can be broadly divided into the following steps: imagine segmentation, feature extraction, classifier in the Department. In the above application can be summed up the core issues and key technologies which consists of the regional division and extraction of the target vehicle and selecting the characteristics of the vehicle. Using image processing theory, this paper puts forward a segmentation and extraction of the border Algorithm for the complete picture of the vehicle.In this paper, the overall design of the algorithm consists of taking the picture into gray-scale handling. The second process is taking the image of vehicle part away from the picture, then in view of the noise in the picture, providing a reasonable noise filtering method. Then extract a complete outline of the vehicle outside curve. In succession analysis the final outline of curve, extract the key line for the calculation of deviation and the measurement of vehicle size. Specific studies as follow:Through the comparative studying for several dynamic object detection methods, bring forward a fitness testing methods and corresponding hardware components. Then take subtracting operation on the image including the vehicle and background image, resulting in the background image reduced to very weak and vehicle image outstanding. Through the comparative study in the application effect from several image segmentation algorithms, select the most stable and high-accuracy algorithm for image segmentation, resulting in the picture only consists of part of the vehicle and little noise on the vehicle body image. Then select the appropriate structural element and use the morphology operation to filter the noise in the image. As a result of the common border extraction method the outline curve will give birth to break lines and break points. Through the study in the caused reason for the break lines and break points, put forward a new method of extracting the close border. As following, through analysis the shape of outline curve, select the key points and bring forward a application scan method extracting the key points. We can use the key points to characterize the movement of vehicle and calculate the deviation of vehicle. In the measurement of size, we should use the coordinates to calculate the vehicle angle then rotate the curves of the outline and use the scan method to calculate the size.