Research On Denoising Algorithm Based On Photorealistic Image Synthesis

With the improvement of image quality,the photorealistic image synthesis based on Monte Carlo is a method that meets people's requirements.The method can simulate the more complex phenomena such as high light,soft shadow,caustics in nature well,and widely meet people's pursuit of special effects quality.Monte Carlo based global illumination algorithm is widely used in photorealistic image synthesis.However,it often takes a long time to synthesize high-quality images.Particularly,when it is prone to specular reflection for the scene of the material,the synthesis image often contains more pulse noise.This paper proposes a fast and efficient impulse noise removal method based on Grubbs detection for the above problem.The method is explained from the sampling point and the pixel level.First of all,it can detect the noise from the sampling point level.And then,it can get a higher accuracy of the noise detection.In addition,the noise which is detected is well performed from the image layer.To a great extent,it can reduce the deviation of the image and improve the time complexity of the algorithm.In order to further improve the efficiency of photorealistic image synthesis,this paper proposes a new method based on the idea that it could find a helpful path quickly with Metropolis ray tracing and use the integrated filter,which can post-process the photorealistic image synthesis.The method includes two parts.On one hand,it deals with the synthetic images with filtering using some sets of the different Gaussion filter.On the other hand,it uses the filtered images to obtain the component results.The method is based on relative root mean square(RMSE)error for selecting the filter size in this paper.According to the size of the RMSE,this paper will adaptively select the best scale filter for each synthetic pixel and reduce the error extremely in the process of the image reconstruction and sampling.Finally,it will obtain the most promising reconstructed results.According to the experimental results,it is illustrated that the algorithm proposed in this paper can achieve better results than the traditional ray tracing algorithm.It can compose the higher quality images using less sampling points.Besides,all methods proposed in this paper have been achieved in the form of programming.