Medical Endoscope Image Processing System Design And SOC Implementation

With the development of modern surgery toward minimally invasive,medical electronic endoscope has become one of the most widely used surgical guidance instruments today.Although most hospitals in China have generally used electronic endoscopes to assist in diagnosis and treatment,however,high-end equipment is still basically dependent on imports from abroad.Because foreign endoscope manufacturers whether the endoscope image sensor or image processor research and development work are carried out earlier,making China’s high-end medical endoscope has been in a disadvantageous position.In the face of foreign technology blockade and their own technical deficiencies,most of the processing system is still based on a dedicated image processing chip,often a single function using the demand is not met.In view of the above problems,the current research status of medical electronic endoscope image processing system at home and abroad is analyzed,and combined with the user’s functional requirements for medical electronic endoscope,this paper proposes an image processing pipeline including endoscopic digital image acquisition,Bayer image demosaicing,one-key white balance,super-resolution video magnification and real-time display,etc.,and finally designs a set of image processing pipeline that can meet the clinical use of The system is designed to meet the clinical use of medical electronic endoscope image processing system in real time.Firstly,the system captures the real-time Bayer image from the endoscope,and then becomes RGB color image after demosaicing,and then white balance color correction is performed on the color image,followed by super resolution 2X magnification,and finally the processed image is displayed on a 1080p monitor in real time.The system design process is divided into two main parts:software processing algorithm and hardware logic design,Considering the specificity of real-time requirements for endoscopic image processing,we first select the appropriate functional image processing algorithms,and then combine multiple processing algorithms under the premise of meeting the processing quality requirements after software simulation to finally form an endoscopic image processing pipeline.In the hardware implementation,for the system has high requirements for real-time,this paper proposes a medical endoscope image processing scheme based on FPGA and ARM as the processing core,focusing on the process of hardware implementation of the software algorithm of each module,in the logic design process to give full play to the characteristics of FPGA high parallelism,low latency,time controllable,etc.,greatly improve the real-time system processing.Finally,the test platform of medical electronic endoscope image processing system is built for on-board verification,i.e.,the functions of each module in the system and the system latency are tested and verified on the AXU5EV-P development board of Zynq UltraScale+MPSoCs development platform is tested to verify the functions of each module in the system and the system latency,and the results show that the design requirements are met.