Development of an Automatic Parking System Using Image Recognition Technology

• Main Authors: Jen-Chian Tzou, 鄒鎮謙
• Other Authors: 蕭耀榮
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/a69neq

碩士 === 國立臺北科技大學 === 車輛工程系所 === 94 === In the recent development of vehicle technologies, improvement of vehicle power and drivability, vehicle electronics and vehicle safety are major developing fields. But intelligent vehicle system also becomes an important field now. The purpose of this research is to develop an intelligent automatic vehicle parking system. This system aims to park a vehicle into a desired parking lot automatically. With the aids of this system, driver’s burden and time spent in parking vehicle can be significantly reduced. There are two major parts of this research. The first part is the vision measuring system, and the second part is vehicle parking control system. In the first part a CCD camera is used to catch the image of a parking destination. Then technologies of machine vision and image processing are applied to locate the parking lot boundary lines from the caught image. The theory of inverse perspective transformation is then used to calculate the dimensions of desired parking lot and the distance between the vehicle and parking lot. In the second part the vehicle is controlled to drive into the parking lot. Mode of garage parking or parallel parking is selected by the driver. Then an optimal path is determined by the information obtained from the processed image. Finally, a fuzzy logic controller is applied to control the vehicle moving into the parking lot precisely through a determined path.