High Dynamic Imaging Based On Multi-Exposure Fusion

Multi exposure fusion is an effective and economical method for high dynamic scene imaging,which has become one of the hot spots in the field of high dynamic imaging.This paper solves the problems of brightness inversion and excessive enhancement of strong edges in the fused image by the existing technology,and proposes a fast multi exposure fusion algorithm based on the response function at the same time.The effectiveness of the proposed methods has been proved by our experiments.The main research achievements of this paper are:1.This paper proposes a novel wavelet-based algorithm for the fusion of multi-exposed images.The well-exposedness weight is improved by introducing the brightness of input images,which suppresses the luminance inversion and prevents the over-and under-exposed in the fused image at the same time.The detail sub-bands of the input images are fused by the adjusted contrast weight to avoid the excessive enhancement of strong edges to preserve the whole details of input images.And the detail sub-bands of fused image are enhanced by proposed enhancement function to improve the contrast of the fused image.The color images are fused in YUV color space,the luminance component Y of which is fused in wavelet domain,whereas the color difference components U,V are fused according to the saturation weight to avoid the color distortion of the fused image.The experiments illustrate that our proposed method can not only preserve and enhance the details of image,but also the luminance distribution of fused image is consistent with that of the input images.2.An objective metric to evaluate the luminance distribution consistency of fused image is proposed.The metric is effective and the result of evaluation is consistent with that of the subjective.3.A fast method of multi exposure fusion based on response function is proposed.In order to make full use of imaging information and solve the problem of low efficiency of existing methods,this paper proposes a fusion method of multi exposure images based on mapping curve adjustment.The mapping curves can be obtained the exposure time of input images and the response function.By connecting the mapping curve of two different exposure images in the illumination overlap region,a new mapping curve scene,which can cover the entire dynamic range,is obtained.And the value of the pixel in fusion image can be calculated by the adjustment of mapping curve.Experimental results show that the proposed fusion method is much faster than the existing fusion at the same time ensures quality of the fusion image.4.A high dynamic range scene imaging system is designed which can automatically adjust the exposure time of each image,which aims to solve the problems of the initial exposure time is not appropriate and the dynamic range of scene changes.The imaging system can adaptive the dynamic range of HDR scene to ensure that total luminance range can be covered and at the same time to improve the visual effect of fused image.Experiments show that the designed imaging system can quickly adjust the exposure time to an appropriate value,which is adaptive for the dynamic range of high dynamic range scene.