© 2021, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 2167 Automatic Number Plate Recognition using YOLO for Indian Conditions Aayush Gattawar 1 , Sandesh Vanwadi 2 , Jayesh Pawar 3 , Pratik Dhore 4 , Prof. Harshada Mhaske 5 1-4 Dept. of Computer Engineering, Pimpri Chinchwad College of Engineering Pune, Maharashtra, India. 5 Professor, Dept. of Computer Engineering, Pimpri Chinchwad College of Engineering Pune, Maharashtra, India. ---------------------------------------------------------------------***----------------------------------------------------------------- Abstract—Automatic Number Plate Recognition Solution (ANPR) is being used from very old days but the technology behind it remains unevolved. Previous systems which are used by the Government of India [1] do not provide a real time solution for the problem. Latest advancements in the field of computer vision and decline in the prices of GPU’s make real time processing of such applications possible. Therefore it is possible to develop a real time Automatic Number Plate Recognition Solution. Keywords—Automatic Number Plate Recognition (ANPR), Optical Character Recognition (OCR), Template matching, Yolo algorithm, Deep Learning. A. INTRODUCTION : Even though now that we know that Automatic Number Plate Recognition(ANPR) is widely known but not used because of its usage of the old technology, we are proposing to develop an application to manage the visitors in the residential areas using the same technology with GPU centric algorithms. Now to train the system we are going to use GPU centric algorithms like YOLO to compete with different challenges we face while working with ANPR in countries like India. In India we face challenges like, the cameras used are mostly used for surveillance purposes and not for training and testing of deep learning models, the second challenge we face is that the number plate sizes and character patterns vary from different vehicles which really makes things complicated. But, by using GPU centric algorithms and using some good camera handling techniques we can achieve good accuracy to implement different applications like visitor management systems using computer vision technologies like Automatic Number Plate Recognition. Visitor Management Systems helps the admin or the front desk officer to manage all the visitors visiting in that particular area in a day. For such applications, we can not actually manage the visitors but all the vehicles of the visitors with their number plate. Admin or the person working with the application can look for all the visitors stored at the end of the day. B. EXISTING SYSTEMS: To create a society without human intervention cars can be used as a Proof of Identity(PoI) and the unique registration plate number provides the opportunity to track the movement of cars in turn tracing humans that are associated with that car. This was identified in the past that is why tracing of cars is being used from as early as the 1980s. Advancements in computer vision techniques now allow us to do that automatically without human intervention. ‘Feature based number plate localization’ [2] is a technique that is used for number plate localization. This approach consists of a number of algorithms developed on the basis of general features of both, characters and number plate. For pre-processing, the input gray-scale image is adaptively converted into binary image using Ostu’s method. A mask having the shape of inverted ‘L’ and size equal to maximum possible character dimensions is rolled throughout the binary image. At every increment a position is shortlisted as possible character location if there is at least a single white pixel in the region and there is at least a single white pixel on the immediate next row and column of the region. Size of each shortlisted character calculated. If it is less than half of the maximum possible character size that location discarded. Subsequently multiple detected portions are discarded using filters such as white pixel density, height and width and one final region is decided to recognise characters. If cases where the number plate script is not in english language or the number plate is barely visible are excluded then, 82% of the plates were recognized correctly which means in ideal conditions it was able to predict correct outcome only 82% times. The performances of individual sections are 87% for number plate localization and 85% for character recognition and 95% for character segmentation. A SVM (Support Vector Machine) [1] which has been trained on a chosen data set for most of Indian number Plates. A detailed analysis has been done before making these SVM which is the first step to finalize if any of the regions in the video frame has a number plate or not and if it has it will select the exact plate which is containing the number. In this an image is first preprocessed i.e. it is converted from RGB to grayscale. After that segmentation is applied for detecting number plate location in the image. For that image is passed through a sobel edge detection filter to detect horizontal edges. Now the image applied to Otsu's threshold and binarised. After that morphological operations are performed on the image to detect probable number plate regions. All rectangular contours are selected and characters are segmented and recognised. Most of the algorithms use features of number plates to localise the area of number but none them treats number plate as an object. We are trying to use object International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 08 Issue: 01 | Jan 2021 www.irjet.net p-ISSN: 2395-0072