In recent years, the subject of using intelligent traortation systems to reduce road accidents and improve driving safety has been an important subject in many researches. Vision based lane detection is widely used in driver assistance systems. To achieve real-time performance in these systems, it is necessary to implement them in hardware. FPGA, due to its high operating speed and design flexibility, is a suitable structure to be used in design and development of lane detection systems. This work presents design and implementation of a lane detection algorithm in which fuzzy logic technique is employed. Fuzzy concepts are used to achieve desired speed and accuracy in lane detection system. The proposed algorithm is for normal road conditions and roads with lane markers. To detect the lanes, the algorithm uses Hough transform technique, a common technique used in lane detection systems. The technique is applied on a specific area of road image, the region of interest (ROI), to decrease circuit complexity and improve algorithm’s speed and accuracy. Image fuzzification and fuzzy gradient is used for edge detection. Membership degree is used for intensity thresholding and weighted accumulation of Hough space cells. And two algorithm parameters,??and??, related to the most voted cells are considered as lane boundaries. The proposed algorithm is simulated and evaluated in MATLAB. CALTECH dataset is employed for this purpose. Hardware design of the algorithm is performed using VHDL hardware description language, and by using ISE software it is implemented on FPGA Spartan6 XC6SLX9 chip. The resource usages in this implementation is lower than those of previous works. Simulation results show that average processing time for each frame with resolution of 480×640 pixels is 2.43ms, and the correct detection rate is 95.7 Lane Detection System, Hough Transform, Fuzzy Logic, Hardware Implementation, Real-time Processing, FPGA
