<![CDATA[Indonesia is the country with the highest population growth in the world. The increasing population of Indonesia has resulted in higher transportation needs because people prefer to use private vehicles rather than public transportation. This causes problems for motorists when looking for an empty parking space in the parking area. To overcome this problem, the author uses the YOLOv3-Tiny algorithm method to detect and identify parking spaces. The research conducted by the author is an implementation of development research. The author shows a literature study to obtain the required information, performs requirements engineering to analyze needs in conducting research, then designs and implements a system based on needs. The design and implementation results will be tested and evaluated until the end of the stages carried out are concluded. From the results of system testing, the system obtained that the accuracy of the information for car parking slots was 98% and 97% for motorcycle parking lots. The system's accuracy rate for detecting car objects is 98.48%, and 95.27% for motorbikes. The average value of the accuracy of the servo response when opening and closing the bars is 88.88%. The results of the CPU speed test in 3-speed modes obtained the best performance at 2.0GHz. Therefore, an automatic parking system prototype can be made and can carry out its essential functions. The system can detect empty and non-empty parking lots by object detection and classification using the YOLOv3-Tiny model. The system performance results for speed detecting and displaying parking slot information with a maximum CPU speed of 2.3 fps and the fastest average system computing time at 0.41 seconds per loop from the beginning of the detection process to the end of displaying information.. The system's accuracy influences the results of the information displayed in detecting objects.]]>
Copyrights © 2022
