| Atmospheric visibility is one of the important meteorological elements of safety in the transportation sector such as aviation,navigation and land transportation.It is essential to be able to accurately and timely detect the visibility and alert the driver.Based on the analysis of the existing video method visibility detection method,based on the basic theory of atmospheric visibility detection,the highway visibility detection system based on digital image processing is designed.This paper mainly studies from the following aspects:1)The double brightness difference is used for daytime visibility detection.The principle of double brightness difference detection is studied.The daytime visibility detection model is built.The target reference object close to the standard black body is designed.In order to accurately locate the target area in the image,a gradient based area is proposed.The growing Hough circle detection algorithm ensures the accuracy and reliability of the test results.2)The dual-light source method is used for uight visibility detection.On the basis of the daytime visibility detection model,the design uses a dual light source with uniformity,consistency and stability to replace the center of the reference object,which can effectively eliminate the unstable source and detect The resulting error.The watershed segmentation algorithm based on prior knowledge mark is used to suppress the over-segmentation phenomenon when extracting the gray value of the target region on the image.3)The gray threshold of the switching of the day-night visibility detection system is determined based on the trend of the gray level mean of the image.Finally,in the off-site environment,the digital image processing-based highway visibility detection system designed in this paper and the Huizhu MDVI-1 standard visibility meter were compared all the time.According to the experimental results,the system meets the expected requirements and has certain Accuracy and reliability. |
PDF Full Text Request