Research Of Automatic Exposure Based On Intelligent Vehicle

As people living standard gradually improve, the car as a new type of transport more and more get the welcome of the masses, the status of the active safety of cars also seriously rise gradually along with social progress, and smart cars is put forward, with the research of intelligent vehicle, presents many vision based application in vehicle, car digital camera as the main input device of car vision, also more and more important in the intelligent vehicle system using, on-board camera should not only with auto electric parts must also is better than general camera for illumination adaptability is strong, the output is more stable and reliable.Set out in this paper, based on the platform of the intelligent vehicle, the first two types of on-board camera photosensitive sensor is presented, including its principle and their respective advantages and disadvantages, and further introduces the camera model selection in this paper.Then illustrates the camera imaging principle and the principle of automatic exposure, and the main of several parameters affecting the ae, include: exposure time, aperture size and the size of gain, illustrates the principle of the impact of exposure, this paper introduces the principle of several basic exposure algorithms and implementation process.The external environment of the intelligent vehicle platform is complex and changeable, exposure procedures are required in the process of high-speed exposure to get fast, accurate, stable, and the complexity of the smart car system, make the distribution in ae system memory and processing time is not too big.For the above points, this article reviewed the existing algorithms, these algorithms are summarized, and based on the advantages and disadvantages of these algorithms in vehicle platform gives a new algorithm of exposure.Algorithm for traffic sign detection is divided into two parts, when there is no detected traffic signs, in view of the existing grayscale average algorithm such as slow convergence through problems, this paper proposes a new improved dynamic average gray level rapidly adjust exposure method, the method of dynamic adjustment algorithm for calculating grayscale average and change in the past using the algorithm for calculating grayscale average operation rate is slow, and prone to shocks.When detected traffic signs, exposure algorithm based on image object image entropy is given, based on image entropy and the relationship between the exposure value, make the program can be targeted for exposure, very good solution to the conventional algorithm can’t exposed shortcomings on specific goals, and is not influenced by the outside light to improve the accuracy of backlight and exposure, especially on traffic sign detection rate has a great improvement.Finally the algorithm is implemented on the system, the experimental results show that exposure to adjust higher real-time performance, reaction more flexible, under high dynamic environment can accurate exposure, on the rates of detection of traffic signs have greatly ascend, and in terms of memory and processing time is very reasonable, shows very good performance.