Microprocessor-Based On-Street Parking Notification System

Ko, Christina Yun Fan (2019) Microprocessor-Based On-Street Parking Notification System. Project Report. UTeM, Melaka, Malaysia. (Submitted)

Text (24 Pages) Microprocessor-Based On-Street Parking Notification System.pdf - Submitted Version Download (411kB)

Abstract

Due to the flaw existed in public transportation system in Malaysia, most of the locals will rather use their own transportation to travel to the desired destination. Thus, this causes the number of vehicles in Malaysia to increase, which leads to finding a vacant parking lot is getting tougher especially in a crowded area. Thus, it causes traffic congestion and consumes additional time and fuel. Therefore, an efficient and user-friendly on-street parking system with the implementation of Internet of Things (IoT) and cloud computing was able to meet the deficiency of current parking system in Malaysia. The objectives of this project is to design and develop a smart parking system using computer vision algorithm and the performance of the parking vacancy detection system in terms of accuracy was evaluated. In addition, a mobile application was introduced to enable the drivers to check the status of availability of parking space at anywhere and anytime. The microprocessor-based on-street parking notification system consisted of two parts which are software and hardware. The prototype was made up of an Intel UP Squared board which acted as a server to integrate all the inputs and outputs, also a medium to process the captures video or images. An IP camera with a tripod stand was set up to capture sharper and clearer images or videos. On the other hand, the software used to develop this parking notification system is Android Studio and Firebase. After that, the image that has been captured by the IP camera was being analysed via OpenCV in Intel UP Squared board and then the data was being sent to the Firebase and stored as a real-time database. The evaluations on the performance of data synchronization system between OpenCV, Firebase and mobile application and parking vacancy detection system at 3 different type of parking and at 4 different environments were done. The results has proved that the data in the images were successfully transferred to the Firebase and mobile application. The results obtained also showed that the contour parking vacancy detection system and Haar Cascade parking vacancy detection system has achieved 100% of accuracy when tested at three different types of parking. However, the accuracy of Haar Cascade parking vacancy detection system was dropped from 100% to 25% when it is tested at the different environments.

Actions (login required)

View Item