Research On The Driver's Vision Enhancement System Based On The Information Fusion Technology

Under low visibility weather condition, such as foggy, rainy, snowy, and sandy days,visibility of the road environment system becomes back So the driver cannot obtains thesufficient information of the road environment through the vision, which is extremely easy tocause serious traffic accidents which lead to a fatal car crash, such as the vehicle collision,rushing out of lane. Therefore, howto improve drivers' visible scope, to enhance driving safety—the driver vision enhancement becomes an important research content of the intelligenttransportation domain. The driver vision enhancement strengthens driver's visual effect underthe fight insufficient condition, for example rainy, foggy days, making use of various kinds ofsensor and advanced technology.Image enhancement is aimed at increasing the image visual effect—the original image isconverted into the form which is more suited to human observation and computer analysis. Thetraditional image enhancement methods play an important role in improving the image quality.Along with the continuous development of image enhancement technology and new imageenhancement methods unceasingly coming forth, information fusion technology has become themain means of image enhancement. Taking full advantage of redundant information andcomplementary information of multi-sensor image, multi-sensor image information fusion canraise the overall quality of the images. This paper introduces the basic concepts, the processstructure, implementing methods, and application domain of image fusion as well as the basictheories and methods of image enhancement.In order to realize to automatic real-time enhancement road image under the low visibilityweather condition, first the definition of road images need to automatic detect; and then whichenhancement method is determined to adopt according to the image definition. In this paper, thecommonly used evaluation parameters of image definition is introduced, as well as the characterof the road image samples which are collected in low-visible fogy day is analyzed, and somereference data that can be used to estimate the definition level of image is obtain, which providea basis for the foggy road image discrimination. Meanwhile, in this paper, based on statisticalanalysis and image fusion technology, the application of road image enhancement in the nightand foggy day are researched. Some samples of image are processed and the effect of imagefusion is analyzed using a common evaluation function.Finally, this paper discusses multi- distinguishing wavelet transform, reconstruction andimage fusion rules of the low visibility road's images. Depending on the different spatialfrequency character of the details components of the different decomposition, a choice fusionoperator based on regional characteristic measure is adopted to enhance source images. Based onsix indicators of the standard deviation, entropy, RMSE, NLSE PSNR and the average gradient, different images obtained by different fusion roles are analyzed and compared in this paper.Comprehensive integration of all data analysis, we can see that the experimental results based onwavelet transform in the six indicators be better than that adopted the weighted average andLaplacian pyramid transform.