Research On Multi-focus Elemental Image Fusion And Improved Integral Imaging Reconstruction

Among auto-stereoscopic technologies,integral imaging technology can provideobserver images with full parallax and continuous viewing points,also do not require special viewing devices and coherent light sources.It is currently the most promising imaging technique in the field of 3D imaging and visualization.The integral imaging system mainly includes the pickup and the reconstruction processes.In camera array based pickup system,there is a local blurring multi-focus problem of the element images due to the limited depth of field of the imaging sensor.In the process of micro-lens arrays integral imaging reconstruction,there is a depth reverse problem in optical reconstruction,while traditional computer reconstruction has the disadvantages of low imaging quality and long computation time.In order to solve two problems,this paper focuses on the research of multi-focus elemental images fusion based on the Sum-Modified-Laplacian and the real-time orthoscopic computer reconstruction based on ray tracing with depth information.This paper presented a multi-focus elemental image fusion method for integral imaging system.Based on the Sum-Modified-Laplacian's response to high-frequency components in the blurred image,the image focus measurement function is designed.The focus position information in the registered elemental image is extracted to obtain the initial decision map.The classification method based on clustering is used to further optimize the initial decision map,and the final fusion decision map is obtained through iteration.The multi-focus problem of elemental images in the synthetic aperture integral imaging system is solved,the image quality of the elemental images is significantly improved,and the following high-quality reconstruction is also ensured.To prove the effectiveness of the proposed method,we conducted a series of experiments.In the fusion experiment,the blurry regions become sharp focus after the fusion process.In the computer reconstruction experiment,the quality of the reconstructed images is significantly better than that of the original multi-focus elemental images.The proposed method can effectively improve the image quality of multi-focus elemental images and avoid the influence of blurred element image on the following computer reconstruction,finally improve the overall performance of synthetic aperture integral imaging system.Next,this paper proposed a real-time orthoscopic computer integral imaging reconstruction.The ray-tracing algorithm was used to analyze the recording process of the light rays during the micro-lens array imaging process and the propagation of the light rays during the micro-lens array reconstruction.The mapping relationship between the pixels in the elemental image array and the pixels in the reconstructed slice image was determined.The captured elemental image array is rapidly pixel matched to obtain reconstructed slice images at different depth planes during the computer reconstruction.The proposed method can solve the depth reverse problem and achieve an efficient computer reconstruction.In order to prove the effectiveness of the proposed method,experiments compare the reconstruction results,and then analyze the reconstruction time and the quality of the reconstructed images.Based on the reconstruction results,a depth estimation experiment is performed.The experimental results show that the proposed method can realize real-time orthoscopic computer integral imaging reconstruction and outperforms traditional computer reconstruction methods and smart pixel mapping methods.