Vehicle Number Extraction Using Open Source Tools

Although intelligent monitoring and recording system (IMRS) is used in many fields such as aviation traffic control, transportation, real estate, medical science and more. One of these is vehicle traffic on roads. This paper covers the techniques of extraction of vehicle number by using open source tools which will be used in many fields such as automated parking system, toll tax system, supervising road traffic on highways, searching stolen cars, and more. Time and accuracy are two challenges in performing the above in a real-life scenario. The proposed vehicle number extraction works on OpenCV Library to convert the image to Mat, and after that, the discussed methodology is applied to detect the vehicle number. This paper introduces a framework which includes preprocessing of vehicle image and applies non-maxima suppression with edge detection and segmentation, and then, OCR has been applied to convert the numbers to digital form and store in HBase database along with an image of the car. It has been observed that the proposed methodology of this paper runs 27% faster than automatic number plate recognition (ANPR). Precision and recall are also observed to be better.