Smearing Algorithm and MSER To Detect Indonesian Private Lisence Plate Detection In Image

Abstract

Motor vehicles are a modern transportation that is still in use today. Each motor vehicle has a unique code bone code and machine code, and this code combined with the area code where the vehicle is registered and save in vehicle number letter (STNK). Each vehicle registration number is unique for each vehicle. This paper propose a method to detect the license plate of vehicles by using smearing and Maximally Stable Extremal Regions (MSER) detection. This paper uses private vehicle number plate as test data because on private vehicle number plate in Indonesia has high-level contrast between background and foreground. By using the smearing algorithm, the level of contrast colour that can be in a licence plate region can detected. An image area that has a high contrast level is then re-analysed using MSER. Areas with detection results that have characters of more than two will selected as the number plate area. The data used in this paper is randomly drawn car image data using a digital camera. Based on the available test results of 80% made on the area identified as the number plate. The test results show the method that is classified as the number plate area on the car image that has the license plate.

Keywords : Smearing Algorthm, MSER, and Lisence Plate Detection.