Research And Implementation Of The Intelligent Driving System Based On Visual Information

In recent years the research of intelligent transportation and intelligent vehicle platform is very hot. The research of the intelligent vehicle and safety auxiliary driving system based on visual information is the most valued subject. Because of the particularity of its application background, the lanes marking line recognition based on visual information is the key and the basic of the research content of the intelligent driving. At present lots of car terminal products such as navigators are completed in the embedded equipment, it has a broad market prospect and application prospect. It shows, the lanes marking line recognition based on visual information in embedded systems has very high research value and application value.In this thesis, the key technology of lanes marking line identification was discussed. This paper established and optimized visual data input, transmission and storage module, this makes the embedded system solutions for visual image of the frame rate reach more than10frames per second. It plays a critical role for the system to meet the realization of the function of real-time demand. This paper built the vehicle-mounted system which can be very convenient to achieve forward, backward, turn operation based on visual information and in the light of the visual image from the embedded system of intelligent vehicle system, to set up the plan and process for the implementation of algorithm analysis, it greatly reduced the algorithm complexity, while improving the reliability of the algorithm. This thesis mainly studies the following content:1. The visual information processing of intelligent vehicle system is the core part of this thesis. The first goal of the establishment and optimization of system is to improve the frame rate of visual data of the system and to ensure the real-time of the system. The visual system consists of camera access, the encoding of visual data, visual data format conversion, as well as visual data transmission and storage.2. Lanes marking line recognition algorithm based on SOPC system is the core content of this thesis, it propose a lanes marking line recognition algorithm based on edge detection and region growing. By dealing with lanes marking line in the ROI selected by the algorithm for this thesis, it get the value of the curvature and the linear equations of the Lane marking line. It provides control data for the system to automatically run or turn away from the channel and provide alarm data if the car is deviating from the driveway.3. According to the system’s function demand, the specific function module is designed based on the visual information of the intelligent vehicle systems. Algorithm simulate under the Visual Studio2005. Hardware project module and software functional part of the whole system is developed in the Quartus II9.1and the Nios II9.1development environment, which make the system recognition of lane marking line in the SOPC platform in the real time and realize the vehicle intelligence control system.