Localization and recognition of car's logo

Abstract

In the last few years, the quantity of vehicles on the road is increasing daily. Even though vehicles are a great help to human kind, it also has led to many problems such as traffic jam, car theft, and accidents. New solutions are needed to resolve some of these problems. Vehicle logo recognition is one of the employed methods to classify a vehicle by identifying its manufacturer when the license plate information is not available. This project aims to localize and recognize vehicle’s logo from a given image. The proposed method in this project is to apply image enhancement and noise removal algorithms followed by a series of morphological operations to localize and detect the logo. Template matching method is then employed for classifying the logo according to its manufacturer. The results obtained from the proposed methodology gives positive results with a performance accuracy of 91%. All the eight logos used for detection has a high recognition rate. In conclusion, the chosen methodology is proven working. Recommendations and future works to improve the robustness of this project is discussed at the end of this report.