International Journal of Computer Science Trends and Technology (IJCST) – Volume 6 Issue 3, May - June 2018 ISSN: 2347-8578 www.ijcstjournal.org Page 214 Analysis and Comparison of License Vehicle Number Plate Recognition Technique Munish Bansal [1] , Navneet Kaur [2] M.Tech Student [1] , Assistant Professor [2] GGSCMT, Kharar ABSTRACT Number plate recognition system recognizes the characters from license vehicle number plate. It becomes a difficult task to recognize the number plate characters of swiftly moving vehicles manually or the blurred images of the number plates. A number of techniques have been used to recognize the characters of the number plate from their images. These techniques used different methods such as Morphological operations, edge detection, character segmentation, character recognition using template matching and character extraction. In these techniques various algorithms are used such as sobel algorithm and local Otsu segmentation method. Various image noises like salt and pepper are added to obtain the desired results. This paper reviews various techniques and their results on the number plate character recognition. This paper also reviews the level of accuracy of number plate detection and recognition of various techniques in terms of percentage. Keywords: - Number Plate Recognitions, Morphological, Salt & Pepper, Local Otsu, Sobel algorithm I. INTRODUCTION The number of vehicles in our country increases day by day. Which causes increase in crimes, traffic on roads, traffic laws violations, hit and run cases and heavy rush on toll plazas installed at various places for toll tax collection[2]. Traffic police can’t solve all the crime cases because police is unable to detect the vehicle involved in crime. To detect the vehicle accurately, police should have a good automatic vehicle recognition system. The major problem in accurate number plate recognition is bad quality of vehicle image captured by CCTV cameras. Due to bad weather conditions such as fog and rain, bad light effects , different fonts and background colors of number palates, the image got noise. Due to noisy image, the characters can’t be recognized accurately. There exists a number of vehicle recognition systems based on various techniques to recognize the characters from noisy images. The existing techniques used various edge detection algorithms and Optical Character Reader (OCR) method to recognize number plate images[1].These methods work on different types of image noises such as Salt & Pepper[3]. These number plate recognition systems can also be used for traffic management, online parking management , automatic toll collection and congestion control[2]. When number plate is detected accurately, police or toll authority can easily get information about the owner of vehicle. The Indian number plate has ten characters. First two letters gives the state information, next two digits gives the district information, next two letters are optional and last four numbers are the unique registration number of vehicle[2]. For example number is HR11ME1111. HR 11 ME 1111 State code District code Optional Unique License Plate number Table 1: Description of Indian number plate This paper includes four sections. First section includes Introduction about number plate recognition system and its uses. Second section includes existing techniques of number plate recognition. Third section includes result and discussion. Forth section includes conclusion and future scope. II. TECHNIQUES OF NUMBER PLATE RECOGNITION The existing techniques used for number plate recognition are as following: A. Histogram based technique This technique utilizes image processing and pattern recognition methods for Open Road Tolling. This is used for Open Road Tolling (ORT) using number plate recognition. This Number Plate Recognition (NPR) technique consist of two modules: histogram based number plate localization and number plate recognition using template matching. This approach has an advantage of being simple & faster and will be used for images size more than 700×700 pixel. To make it faster all the operations are performed on gray scale image not on RGB image. Open Road tolling has come up in a large way in foreign countries, but not in India to that extent here it is still at the level of idea[2][6]. Steps for NPR (Number Plate Recognition):- 1. Colour to gray scale image conversion 2. Image dilation 3. Horizontal edge detection 4. Vertical edge detection 5. Segmentation 6. Number plate extraction 7. Character segmentation RESEARCH ARTICLE OPEN ACCESS