Font Size: a A A

Implementation Of Image Acquisition And Enhancement Algorithm For UHD Endoscope Based On FPGA

Posted on:2022-03-10Degree:MasterType:Thesis
Country:ChinaCandidate:L WangFull Text:PDF
GTID:2492306569960529Subject:Control Science and Engineering
Abstract/Summary:
In recent years,surgical operations have gradually developed in the direction of minimally invasive surgery,and electronic medical endoscopes have become an indispensable medical device for minimally invasive surgery.Due to the increasing demand for endoscope image resolution,the resolution needs to be increased from 1920*1080 to 3840*2160.In terms of video capture,transmission,real-time processing,it is difficult for traditional endoscope designs to meet the needs of ultra-high-definition endoscopes as the amount of video data has doubled.In response to the above problems,this paper uses a parallel architecture FPGA as the main processor to design a set of ultra-high-definition endoscope video capture,transmission and image enhancement processing system.The system is divided into two parts: the camera head,the image processing host.The camera head realizes ultra-high-definition video capture and long-distance real-time transmission functions,and the image processing host realizes the real-time image enhancement processing and display functions of ultra-high-definition video.In terms of video capture,since the ultra-high-definition resolution image sensor outputs MIPI CSI-2 protocol data,this paper utilizes the advantage of FPGA logic programming to realize the physical layer D-PHY and protocol layer CSI-2 of the MIPI CSI-2 protocol,decode the image data.In terms of long-distance video transmission,the camera head needs to transfer the collected video data to the image processing host 5m away in real time without compression.This paper chooses the Aurora protocol based on 8B/10 B to encode the original image data,and realizes the serial data sending and receiving with the help of Xilinx Ser Des.The serial data bandwidth is 3.125 Gbps,which is converted from electrical signals to optical signals by the optical module and transmitted to the image processing host.In terms of real-time video processing and display,this paper utilizes the advantages of FPGA parallel processing to realize the logical mapping of image preprocessing and image enhancement algorithms.Image preprocessing includes: dead pixel removal,Bayer interpolation and white balance processing.In order to ensure that the color does not change during the processing,this paper changes the color space from RGB to HSI.Image enhancement is carried out from three aspects: contrast,saturation,and sharpness.In order to prevent the image from being broken,the system designs a DDR3 read-write arbitration control module and selects two DDR3 chips as external memory,and the system bandwidth is25.6 Gbps.In terms of video display,this paper chooses Sil9136-3 as the HDMI encoding chip,and the video is finally displayed on a 4K monitor.Through the performance index test of the system,the video acquisition,transmission and image enhancement processing system designed in this paper meets the needs of ultrahigh-definition endoscopes.The system delay is 110 ms.The resolution is 3840*2160,and the frame rate is 30 fps.The design of this paper provides a set of solutions for the ultra-highdefinition endoscope scene.
Keywords/Search Tags:Ultra-high-definition endoscope, image enhancement, MIPI CSI-2, high-speed signal transmission, FPGA
Related items