Intelligent Parking Guidance System Based On Computer Vision Research

With the advancement of China’s urbanization process, urban scale growing at the same time, many contradictions are increasingly stand out, in addition to traffic paralyzed by steam flow not free, often around the business circle of the parking problem has become the bottleneck of restricting the business circle to carry out economic activities, and then gradually evolved into social livelihood issues. As an important part of smart city intelligent parking lot, become a research hotspot in recent years. With the continuous development of science and technology, relying on computer vision, image processing, the Internet of things technology, build an efficient, standardized, intelligent parking lot.First,This article in investigating the current status and the use of intelligent parking lot, on the basis of technical shortcomings, puts forward the application of computer vision, image processing technology, solve the problem of parking lot parking management chaos, cruise, etc, in order to realize the parking lot parking precise guidance. In computer vision solutions under the idea of license plate recognition system is the key technology for vehicle entry makes a preliminary research; In research on parking state recognition system based on computer vision was put forward in the area of covariance parking recognition algorithm, the experiment shows that the algorithm has strong robustness, satisfy the parking lot, internal complex light conditions.Second, the application of graph theory knowledge abstraction modeling was carried out on the intelligent parking lot; According to road congestion does this feel, defines a mathematical descriptor; On the basis of quantitative road congestion, this paper proposes a new structure appropriate for the intelligent parking lot link weights method; Based on the traditional Dijkstra algorithm, degign a suitable for the intelligent parking lot model of optimal path planning guide.