| Summary: | 碩士 === 逢甲大學 === 資訊工程學系 === 106 === Population density of Taiwan is the second highest in the world. Density of motorcycle is the highest in the world. Even Japan shows Taiwan as the "motorcycle kingdom." The primary reason for that is the lack of public transportation and the short commuting time, which produce the different traffic pattern from other countries. However, as the number of cars are increasing but the number of parking spaces has not increased, and which cause drivers looking for parking spaces difficultly. In order to know where are the available parking spaces, some people use the IoT method which actually can work in real life. With the vigorous development of neural networks, some people also use the method of image recognition to identify parking spaces. However, these people were trying to detect the parking space for cars but nobody tries to detect the parking space for motorcycles.
This thesis is mainly about the development and implementation of identifying the parking spaces for motorcycle, and providing an APP service platform for students in Feng Chia University who can instantly know the status of parking spaces. The Canny algorithm can efficiently and accurately find out the contours of the motorcycles, and make the data sets of the area and outline lines enclosed by these contours, and correspond to different levels of parking density. Next, we used the KNN algorithm to produce the parking pattern. The model makes predictions for different conditions at different times. We also statistic and analysis the parking hotspots at each parking lot location, as well as the popular parking period. It can also provide a reference for planning future parking lots. On the other hand, given the lack of parking space or too far from the student's classroom, the students prefer to choose the closer parking lot, but it is the location of illegal parking. We also conduct statistics on traffic violations over the years. The analysis helps the user who has no choice but need to temporary parking to avoid high-risk time. Finally, we will make the above-mentioned module into a mobile app for everyone to use.
|
|---|
