Research On The Close-range Fusion System Of Portrait And Video In The Inspection Cabin Of Large-scale Medical Equipment

The inspection cabin of large-scale diagnosis and treatment equipment is partially closed,and the internal space is small,so that the patient’s condition cannot be directly observed from the outside.In order to ensure that the patient can complete the examination safely,the diagnosis and treatment monitoring platform is important.The traditional PTZ camera system solves the problem of large field of view,but it cannot meet the detection requirements of moving targets,and at the same time is large and complicated in mechanical structure.The multi-camera system solves the problem of continuous monitoring,but it cannot meet the long format requirements;ultra-wide-angle fisheye image system,after calibration,the information integrity cannot be guaranteed,and it cannot meet the needs of close-up portrait video collection.Conventional video acquisition systems cannot meet the needs of ultra-long format and close-up portrait image acquisition.Therefore,this article focuses on the fusion of close-up portrait video in the inspection cabin of large-scale diagnostic equipment.The main work is as follows:(1)Aiming at the practical needs of large-scale diagnostic equipment inspection cabins,a portrait video acquisition module-application client for portrait video acquisition Processing architecture.The portrait video capture module is composed of a dual-channel wide-angle network camera,with H.264 compression coding;the video fusion server splices and merges multiple videos to generate an ultra-long-format video stream and pushes it to the application server;the application server combines the actual needs of medical images to use the generated video for human body scanning,positioning and other medical applications.(2)Aiming at close-up portrait video image registration,a close-range image registration method based on SIFT(Scale Invariant Feature Transform,SIFT)algorithm is proposed.The effective registration of portrait video images is achieved in the inspection cabin of the diagnosis and treatment equipment,and the structural similarity is used.The coefficients have been quantitatively analyzed for the improved method.In the case of dual-channel video,the structural similarity coefficient is improved by 5.06% than that of image stitching,solving the problem of perspective deformation caused by the near scene of the target,and the experimental results meet the application requirements;In the image fusion process,a method for adaptively adjusting the brightness of the image based on the gray value is proposed,and combined with the multi-resolution linear fusion method,It eliminates the exposure problems caused by uneven light and the traces left by image registration,and uses the brightness similarity coefficient to evaluate the proposed image brightness adaptive adjustment method.It is also better than the brightness of image stitching under the condition of dual-channel video Similarity coefficient increased by 7.68%.(3)The actual portrait video fusion system was deployed in a diagnosis and treatment device to achieve the close-up fusion of portrait video with a resolution of 1280?720 and a frame rate of 30 fps.After multiple debugging,it solved the multi-channel video synchronization,close-up fusion of portrait video and uncertain lighting conditions.At present,the system has been put into use in small batches.