Light Field Camera-based Extracting Of Sub-aperture Image And Application Of Face Detection

With the development of science and technology,light-field imaging technologies have played an important role.As a new type of camera: light-field camera,has also received more and more attention.The images are taken by traditional camera only record two-dimension information,and the light-field camera can record all the information of the scene,after being calculation,we can obtain the multi-view image,different focusing plane image,and the full focusing image of the scene.Because the multi-view image acquisition technology of the light-field camera is not open,and the application and prospect of multi-viewpoint image is very large.The paper has proposed a method for extracting Sub-aperture image of micro-lens light-field camera.The first shooting and post focusing function of light field camera is also very important in the real life,so we use the advantage of the light field camera in refocusing.Combine with LBP face detection algorithm,compare with the effect of the human face obtained by the light field camera and the traditional camera.The main work and innovation are as follows:1)A method for extracting sub-aperture image of micro-lens light-field camera is proposed.In this method,Firstly,the data collected by the light field camera is analyzed,The image of a sub-aperture of the main lens on the CCD has the same coordinate as that of the micro lens unit in the light field camera.The sub-aperture image obtained by the light field camera is equivalent to the image generated by the main lens aperture of the light field camera.Similarly,the image is equivalent to the image of the equivalent array of micro-lens arrays.So we use the peak detection method to calibrate the center point of each micro-lens image,then extract a bright area by the central pixel as radius,the same pixel location within the region are put forward;finally,rearrange into a complete image as a sub-aperture image.These images have small disparity at pixel level.2)An algorithm for face detection based on the refocusing function of light field camera is proposed.In this scheme,in the principle of digital focusing and digital refocusing of the light field camera,the advantage of the LBP face recognition algorithm in the clear mark picture is presented.We can obtain many photos with different focus under the advantage of the focusing characteristics of light-field camera,then use the defocus focusing method to extract the focusing pixels to joining together,we can get a full focus image.In the end,the results of face detection are compared with that of traditional single camera,these results show that the images after the full focus of the light-field camera can be detected at different depths of the face,but the picture of a single focus layer can only be detected in the vicinity of the depth of focus.It shows that the light field camera has the potential and advantages in face detection field.