Study On The Identification Of Multi-Lane Under Complex Condition Based On DSP

Vehicle safety driving assistant system is research object. The methods of frontage multi-lane identification under complex condition and system miniaturization used of DSP are presented.Based on identifying objective function and parabolic model, the genetic algorithm is used to optimize parabolic parameters and realize multi-lane identification under the case of existing lots of noise on the road, it has great adaptability under different light, the shadow of buildings and trees. The calculation method of confidence can judge the degree of identification accuracy about each lane under complex condition. As a whole, identification accuracy has been improved through adjusting the lane's ROI. Process is designed with TMS320DM642 hardware. To improve the real-time capability, the code optimizing methods, which include optimization of application layer, intrinsic function and packaging data, are used to optimizing process. It provides a good hardware platform for industrialization of system.The method of multi-lane identification is installed in experiment vehicle and carries out road tests. The tests show that the method can identifying the boundaries of road accurately, and has real-time capability.The method of multi-lane identification provides key technology and theory support for the lane departure warning. The hardware platform of TMS320DM642 establishes good foundation for industrialization.