A SFPM METHOD FOR INDIAN AUTOMOBILE RANGE PLATE RECOGNITION SFPM METHOD FOR INDIAN AUTOMOBILE RANGE PLATE RECOGNITION.”

: Automobile range plate recognition is a challenging task in cyber crime. The numbers are stated of being in the automobile range plate, that is different shape and pattern in different countries. In India the automobile range plate uses white as background and black as foreground colour. In this paper we propose a SFPM methodology, first we find out the shape of license plate then enhance the image and calculate the characters of the license plate by using segmentations method. At the end of algorithm we apply fuzzy and pattern matching for character recognition. In our work we use two databases, first database store different-2 alphabet format and second database store a different-2 format of number.


Introduction
Automatic range plate recognition (ARPR) could be a method that uses optical character recognition on pictures to scan automobile registration plates. It will use live television, road-rule social control cameras, or cameras explicitly plan for the task. ARPR is employed by police forces around the world for enforcement function, together with to ascertain if a automobile is registered or authorized. It"s conjointly used for electronic toll assortment on pay-per-use roads and as a technique of cataloguing the motion of traffic as an example by highways agencies.
Automatic range plate recognition is wont to store the pictures captured by the cameras yet because the text from the registration plate, with some configurable to store a photograph of the driver. Systems usually use infrared lighting to permit the camera to require the image at any time of the day. [1] [2][3] ARPR technology should take into consideration plate variations from place to place.
Issues regarding these arrangements have targeted on quietness panic of presidency pursuit citizens' motions, misidentification, high delusion rates, and inflated government defrayment. Critics have delineated it as a variety of mass police work. A license number (RTO Registration No.) recognition system has provided much to yield the good results in identification and recognition of checking vehicles status at any place. The number plate recognition system of any type of vehicle consists of three important part: first it is called license plate location, second is called license plate segmentation and third character recognition.
The identification and recognition of any type of vehicle license plate stage to effective in the accuracy of an SFPM system. The input of the SFPM organisation is any type of automobile picture, and the result image is a portion of the picture incorporates the license plate [2]. We know that the license plate can exist anyplace in the image of vehicle. In West Bengal, a state of India, the traffic management system developing on each successive day [3]. In India most of the number plate consist white background with black foreground colour for personal automobile and for the commercial automobile used yellow as background and black as foreground colour. The number plate having two character for representing "state code" followed by two digit numeral followed by single letter after those four succeeding digits as shows in the below figure1[1, 2, 3].

Background
There square measure varied of technique for the range plate detection and recognition system. For real time exercise, the necessary task is that the machine time of the rule. However, there"s continually a trade-off between machine time and performance rate. A stronger performance rate can usually want more machine time [1].  For range plate detection or localization, techniques supported edge data point and mathematical morphology provides a awfully sensible outcome as reported in Bai and Liu (2004) work. They use vertical edge data, calculate the sting density of the picture and followed by morphology ways like dilation to extract the area of interest. This method works well because of the actual fact that range plate continually encompasses a high density of vertical edge. To boot, this technique is appropriate to be enforced once the camera is fastened to urge best photograph to the automobile [3]. A photograph is taken into account sensible once it"s taken underneath intense lighting condition, right angle and also the automobile range plate and its character is showing high distinction to every alternate. However, this rule is tough to be employed during a complicated background since it"s influences to unwanted edge up background which can puzzled the system.

Materials and Methods
The recognition procedure of the SFPM algorithm is very accurate and effective. In this work we used two database, the first database is stored all possible format of alphabet like A, a, a, a, a, a, a, a, a, a, a, a, a etc. and second database stored 0, 0, 0, 0, etc.
In SFPM algorithm, first of all it is applied segmentation method on enhance image and then extract the digits like m,p,0,7..etc. After that here we are done two successive method fuzzy and pattern matching. Step1 -The captured the image by CCTV camera or other sources.
Step2-read the image and selected number plate of vehicle and it is storing in variable for further process. V1=Imread_function("vehicle_image") V_fi=crop_function(V1) Step3-Apply segmentation for extract and save the digit of the vehicle number plate. Seg_digt(a, b, c, d, e, f, g, h, I, j) =Morphi_function(Vi_fi) Step4-All digits are separated in two form alphabet and number.
Gp1= group of alphabet Gp2= group of number Note-for example MP07CA1077 First_ group= (M P C A) Second_group= (0 7 1 0 7 7) Step5-The groups compared with corresponding database by using fuzzy and PM. Step6-at the end we are achieved the result as recognition of the vehicle. The result either may be one vehicle number or probable of vehicle number.
The results section should provide details of all of the experiments that are required to support the conclusions of the paper. The section may be divided into subsections, each with a concise subheading.
It is advised that this section be written in past tense. It is a good idea to rely on charts, graphs, and tables to present the information. This way, the author is not tempted to discuss any conclusions derived from the study. The charts, graphs, and table should be clearly labeled and should include captions that outline the results without drawing any conclusions. A description of statistical tests as it relates to the results should be included.

Results and Discussions
Our work is implemented in matlab7.10 and these algorithms apply on 100 Indian vehicles. First we captured the image by CCTV camera and after apply SFPM algorithm for recognize the vehicle number.
So in SFPM algorithm we improve the image quality of crop image of number plate. After that applying the segmentation method on the number plate image and at time we gathered the individual digit. All digits are separated in the group of number formats and group of alphabet formats. Finally both group are matched with two pre defined database by fuzzy and pattern recognition system. .  Not Recognized ----------------3 6 Recognized MH04 ZZ0000 0 7 Recognized KA03 MB2784 0 8 Recognized DL1Y A3550 0 9 Recognized MP07CA1077 0 10 Not

Conclusions
Our experiment with recognition of vehicles number show that a solution of vehicle crime. We proposed a SFPM methodology that showed a very effective result with high recognition rate because here we have merged the footing concept of fuzzy system with pattern matching. Here fuzzy system provides a list of possible variation in digit and after it is done by patterns matching.