A Smart Technique for Accurate Identification of Vehicle Number Plate Using MATLAB and Raspberry Pi 2 Khushboo Chhikara, M.tech student Mechanical and Automation Department Indira Gandhi Delhi Technical University for Women New Delhi, India Dr.Pankaj Tomar, Assistant Professor Mechanical and Automation Department Indira Gandhi Delhi Technical University for Women New Delhi,India Abstract The aim of the paper is to design a system which captures the image of the number plate of a vehicle using raspberry pi camera and the details are being retrieved using the character segmentation which is done by optical character algorithm. Choosing of an embedded platform leads to the automation in the field of electronics. The objective is to represent a system by using number plate of a vehicle for various application such as inventory control, border checkpoints, highly restricted area (supreme court, military base),etc. The system is executed on raspberry pi micro controller and imitated in MATLAB because the experimental result shows that the system is fast enough in capturing images, recognition of algorithm and data streaming. Keywords Image Processing, MATLAB, Pattern Matching, Raspberry Pi, OCR(optical character recognition) 1. INTRODUCTION Vehicle number plate recognition technique is used for identifying number and gets owner information from a enormous database of registration details. The experimental prototype of the embedded image capturing system with raspberry pi is smaller, lighter and lower power consumption, so it is convenient[1]. According to [5] Morphological operations and edge features are used for the segmentation of vehicle number using two level decisions. In recent years, this technology of number plate recognition has increased popularity in security, traffic control and monitoring applications. Technically, the technology is sounding research topic because enormous discoveries of computers and sophisticated high resolution infrared cameras. This make easier for image processing techniques more applicable analyzing and extracting important features for plate numbers detection and recognitions [13]. Besides the robustness, the earlier methods use either feature based approached using edge detection or Hough transform which are computationally expensive or use artificial neural network which requires large training data[10]. The objective is to design an efficient vehicle identification system by using number plate which can be executed on the entrance of a highly restricted are(military zone, parliament etc.)[2]. It is a application of optical character recognition. OCR is used to recognize an optically printed character number plate which is based on template matching [3]. The proposed work is to develop a system to recognize the number plate and retrieve owner information from the database. It is based on loading a vehicle number plate image to the system, which recognize the character and using that characters, the details of particular license plate number the details are fetched from the database[5] in which a pattern matching based method is used for character recognition as described in the next section. 2. SYSTEM MODEL The system is subdivided into the software model and hardware model. Both model are discussed in detail. 2.1 Software Model The main part of the system is software model. The software model use series of image processing techniques which are executed in MATLAB2014a. Matlab support package for raspberry pi hardware is installed for building up a communication link between Matlab and micro controller. 2.2 Hardware Model The whole system is constituted of following parts: an image capturing camera,raspberry pi board to run image recognition program on it. This board is the central module of embedded 30
image capturing and processing system. Main parts include: Processing chip, Memory(SD card slot), Power supply, Ethernet port and USB ports shown in figure 1. Figure 1: Raspberry pi The camera module used in this project is RPI NOIR Camera board as shown in figure 2. The camera plugs directly into the CSI connector on the raspberry pi[2]. It is able to deliver clear 5MP resolution image or 1080p HD video recording. The module attaches to the raspberry pi by a 15pin ribbon cable to the camera interface. Figure 2: Raspberry pi NoIR camera board 3. OCR SYSTEM The main module in OCR(Optical Character Recognition system are : image acquisition, pre-processing and feature extraction. Image Acquisition module - task of is to obtain text image from a camera or pre stored file. Pre-processing module - used to smooth the digitized characters. Feature Extraction - feature will be extracted from processed image and stored in the database for recognition. 3.1 Flow chart Figure 3: System of OCR 31
3.2 Mathematical formulation Usually the group of detectable cases of the same character corresponds to the one class, but sometimes one class represents two mutually undetectable characters, such as 0 and O. A-set of all possible combinations B-set of all classes F- hypothetic function that assign each element from the set A to an element from the set B F :A B Ŷ=F(Y) where Y A is a description vector (pattern) which describes the structure of classified character and Ŷ B is a classifier, which represents the semantics of such character. The function F is the probably best theoretical classificator, but its construction is impossible since we cannot deal with each combination of descriptors. We construct pattern classifier by using only a limited subset of the A B mappings, such as A A and B B. Now, we construct an approximation F (x, w) of the hypothetic function F, where w is a parameter that affects the quality of the approximation: F(w):A B t xˆ = F ( x, w ) Grayscale For achieving accuracy the image should be grayscale or binarized. To convert RGB image into grayscale following function is used: Y = 0.2126R+0.7152G+0.0722B Feature Extraction In this phase, features of individual character are extracted. The performance of an each character recognition system that depends on the features that are extracted. Assume that there are several line-ends, loops, and junctions in the image. The position of loop is defined by its centre. To form the vector, we must convert rectangular coordinates of the element into polar coordinates [r,θ ] : r= x 2 +y 2 ; Ɵ = atg(y x ) x =(2 x w) w ; y = (2 y h) h A vector of descriptors to distinguish between these characters as follows: Ɵ = (r1,θ1,r2,θ2,r3,θ3 ) Where r1,θ1 is line end 1, r2,θ2is line end 2 and r3,θ3 is junction, as shown in figure. Figure 4: (a) Segmented Character Containing Structural Element, Loop and Line Ends. (b) and (c) Element Is Positioned In Polar Coordinate System. 32
4. EXECUTION The designed system can be operated in two different sessions,i.e one for capturing and creating a database and other session is to capture the image which can be used for identifying to comparing the images in the database. The execution is divided into several steps, described below: 4.1 Image capturing The image of a vehicle number plate is captured using raspberry pi camera processor which is connected to the PC. The images are captured in RGB format which further converted into binary image. Figure5: Image Captured 4.2 Gray scaling and Binarization of the image For achieving the accuracy the image is grayscaled as shown in the figure. Figure 6: (a) Gray Scaled image, (b) Binarized image 4.3 Extraction The number plate is extracted from the captured image and the extra part of the number plate is being removed and only the written material of the plate is being extracted. Figure 7: Removal of Extra Part 4.4 Character Segmentation Individual characters on the plates are segmented. Figure 8: Segmented Characters 4.5 Character recognition It used to compare the each individual character against the complete alphanumeric database. 33
Figure 9: Database 4.6 Template Matching After recognition the character is matched in the database. 5. Result Despite the low accuracy of recognition as compared to other techniques our automatic recognition system of number plate is functioning with accuracy of 99.91%.The system is user friendly, easy to use and reliable which provides more security, privacy and well organized data on board. Scope of improvement is always there by improving the images quality and increasing the processor speed for real- time implementation. 6. CONCLUSION & FUTURE WORK The purpose of this paper is to design an automation system by detecting the vehicle number plate for security reason that could replace the present system of manual entry. In this paper, we presented a system designed for the recognition of car number plate and getting owners detail. Firstly we extracted the plate location, then we separate the number plate characters individually by segmentation and finally apply template matching(optical Character Recognition) with the use of correlation for recognition of plate characters. The system works satisfactorily for wide variation of conditions and different types of number plates. The system is executed in MATLAB and implemented on Raspberry Pi. Our future work focus on complex number plate images such as, cluttered background, blur images, different fonts, and different intensity images, even images captured while on move and to design an ideal character database to avoid the problem of character similarity. REFRENCES [1] G.Senthil Kumar, K.Gopalakrishnan, V.Satish Kumar, Image Capturing System using Raspberry Pi System, IJETTCS, Volume 3, Issue 2, March - April 2014. [2] D.Lavanya, C.V.Keerthi Latha, Nirmala, License Plate Extraction of images using Raspberry Pi,International journal of advanced research in computer engineering & technology(ijarcet), Volume 4 Issue 1, January 2015. [3] Pranob K Charles, V.Harish, M.Swathi, CH. Deepthi, A review on various techniques used for Optical Character Recognition, IJERA,ISSN : 2248-9622, Vol. 2, Issue 1, Jan-Feb 2012, pp.659-662. [4] R.Shandiya, R.Suvetha, R.B.Shalini, Ms. N.Shanmuga Sundari, Tracking Owner details by Automatic Number Plate Recognition, IJETTCS, Volume 5, Issue 2, March-April 2016. [5] Divya.K.N, Dr.Ajit Danti, Recognition of Vehicle Number Plate and retrieval of Vehicle Owner s Registration Details, IJIRTS, ISSN : 2321-1156. [6] Chao-Ho Chen, Tsong-Yi Chen, Min-Tsung Wu, Tsann-Tay Tang, Wu-Chih Hu, License Plate Recognition for Moving Vehicles using a Moving Camera, IEEE, 978-0-7695, 2013. [7] Qiu Chengqun, Design of automobile License Plate Recognition System based on MATLAB and Fuzzy PID,IEEE, 2013. 34
[8] J.Layolin Sobiya, Mrs. P.Angelin Priya, Automatic Multistyle Licensed Plate Detection by using Fuzzy Logic Classifier, IJARCET, Volume 3 Issue 4, April 2014. [9] Sarika Pansare, Dhanashree Joshi, A survey on Optical Character Recognition Techniques, International Journal of Science and research, Volume 3 Issue 12, December 2014. [10] Muhammad Tahir Qadri, Muhammad Asif, Automatic Number Plate Recognition System for Vehicle Identification using Optical Character Recognition, IEEE, 2009. [11] Priyanto Hidayatullah, Nurjannah Syakrani, Ida Suhartini, Wildan Muhlis, Optical Character Recognition improvement for License Plate Recognition in Indonesia, IEEE, 2012. [12] Mr. G.T. Sugar, Prof. Mr. A.V. Shah, Number Plate Recognition using an improved Segmentation, IJIRSET, Volume 3 Issue 5, 2014. [13] Kulkarni Suraj Deeliprao, Prof. Pallavi S. Deshpande, Vehicle Number Plate Recognition System for Theft Detection, International journal of Engineering research & technology(ijert), Volume 4 Issue 02, February 2015. 35