Research On The Methods Of Restoring Degraded Image Due To The Rain Weather

Following the development of computer vision system in more and more fields, the demands to the system in robust and stability have heightened. The system must be suitable for different weather conditions, in order to ensure the normally operating of the vision system all the time. In this paper we have studied the disturbance mechanism of weather degraded image, which we often meet on using the vision system in outdoors, especially studied the problem about restoring rain degraded image. We restored the weather degraded image by using the methods based on physical models and .not based on physical models. The results of the restoring image are improved obviously. The researching result is important to enhance the performance of the vision system using in outdoors.In this paper, we studied on the technology of restoring rain degraded images, According to the physical properties of rain drops, we analyzed the dynamics model and photometry model, restored the weather degraded images by using the both physical models, and we compared the results with the restoring results by using simple images processing methods.Through analyzing the physical properties of rain drops , we know that the visibility depend strongly on camera parameters.camera parameters can also be chosed to reduce the visibility of rain .The result of experiments shows that it is possible to reduce or remove the visual effects of rain during the images acquisitionThrough analyzing the temporal independence of rain frames and assuming 3D projective camera model and parameters ,we create a rain mask for each frame and add it to the background image. We add rain-fog resulit by using scattering model . Then, using statistical information, we remove the rain of video images. which makes up for the fault of the neglect of the rain scene in the common visual characteristics of the effects of rain and fog. The simulation results show that this method is more simple than traditional rain simulation and occupy a small memory.