Environment Illumination Editing And Enhancement Of Low-Light Image Based On Spherical Harmonics Lighting

Illumination is an important factor affecting image quality.Illumination editing refers to the customized adjustment of the real illumination in an image.It is one of the research contents in the fields of computer graphics and digital image process.Editing the illumination of the image makes the illumination distribution prominent to the main object,which is helpful to the research of image dynamic range compression,object recognition,and image enhancement.In the fields of photography,medical treatment,film and animation,virtual reality and augmented reality,it is often necessary to adjust the illumination in the image scene or make the illumination change dynamically.In recent years,with the development of intelligent devices and image editing software,more and more users hope to be able to easily adjust the illumination of images in various environments.Especially in low light environment,higher requirements are put forward for the effectiveness and quality of illumination editing and enhancement.Unlike the artificial lighting of real scenes,there are many challenges in image light editing.Although the relighting method can adjust the linear light source,point light source or spotlight in the image,the illumination and shadow in the weak light environment are very complex and cannot be expressed by a simple illumination model,which makes the results of illumination editing have serious artifacts in the bright part,lack of realistic and illumination disorder in the scene.These problems seriously restrict the rapid development of research and application of illumination editing.To solve the above problems,this paper proposes a method of image illumination editing and enhancement which is suitable for the low light environment.Firstly,RGB-D image is used as input,and the texture-copy phenomenon in depth map refinement is studied to solve the problem;secondly,shadow decomposition of image is studied to improve the non-realistic problem caused by shadow in illumination editing;next,a new interactive method of illumination editing and enhancement is explored to deal with images in complex environments such as low light.It makes the result of illumination editing and enhancement more robust.The main contents and contributions of this paper are summarized as follows:(1)A geometric refinement method of depth map to avoid texture-copy is proposed.Using the idea of texture structure extraction,the texture in the image is filtered first,and then the depth map is optimized by using the color information of the structure map,so as to obtain the finer depth map and to avoid texture artifacts.This method is simple yet effective and provides more accurate 3D information for illumination editing.(2)A shadow-sensitive intrinsic image decomposition method is proposed.With the help of the shadow removal method,the shadow confidence map in the image is first detected.Then an intrinsic image decomposition algorithm is designed to decompose the image into a reflectance map,shading map,and shadow map.This method effectively separates the light and shadow information from the original image,so that the light editing can be carried out at the true light level.(3)A method of illumination editing and enhancement based on spherical harmonic lighting is proposed,which uses the spherical harmonic lighting model and illumination detail map to represent complex environment light;uses interactive spherical harmonic parameter adjustment method to adjust or enhance illumination;proposes a shadow adjustment algorithm,which adjusts shadows according to the change of spherical harmonic lighting;combine multiple components to obtain final illumination editing and enhancement result.This algorithm solves the problem of illumination editing in low light or complex illumination environment,and the results have a good sense of reality.The three parts of this paper focus on the subject of illumination editing and enhancement in the low light environment.The depth map refinement and intrinsic image decomposition are the prerequisites for illumination editing,so that illumination editing and enhancement can better deal with images in low-light and complex environment.The method of this paper has been tested for low light images in different scenes.A large number of experimental results show that this method is feasible and effective.Compared with the existing relighting methods and image enhancement methods,the results of this paper are more consistent with the real illumination adjustment effect and have fewer artifacts.