Research On The Laparoscopic Pose Vision-based Automatic Adjustment For Robot Assisted Surgery

With the rapid development of science and technology and the improvement of the level of social health care,robot technology has been gradually applied to surgery.Compared with the traditional surgery,robot assisted surgery has many advantages,including small incision,less bleeding during the operation.It can greatly reduce the pain of patients,and shorten the postoperative recovery time.So it has been developed rapidly during the recent 30 years.As for the medical applications today,the robot is just used as an auxiliary device.It needs the doctors or other medical personnel to operate robots,making the surgical instruments' pose adjusted and the operation completed at the same time.It would be a meaningful research to improve the autonomy of medical service robot,further emancipate the doctor hands,shorten the operation time and improve operation efficiency.In this thesis,a new method is proposed to automatically adjust the pose of laparoscope,providing a clear and optimal surgical vision for the physician.Through specific clinical examples,the specific procedure of robot assisted minimally invasive surgery is summarized in this thesis.Considering the relationship between the image coordinate system,the camera coordinate system and the joint coordinates of the holding arm,a right and reasonable laparoscopic motion imaging model is established based on the spinor method.According to the motion imaging model and the location and size of the surgical micro instrument relative to the image coordinate system,we can analyze the possible operation state,and formulate the corresponding laparoscopic posture automatic adjustment rules to achieve the automation function.Aiming at the above automatic regulation rules,we design a marker that can be easily detected by vision.Meanwhile,in order to meet the needs of quick tracking during operation,the visual inspection module is divided into two stages: Pre-operative recognition detection and intraoperative quick tracking detection.Aiming at detecting micro surgery instruments,through improved circle Hough transform and Hof linear transform algorithm,and combined with statistical moment method,the preoperative accurate detection and the intraoperative quick detection is improved.For quick tracking detection of micro surgery instruments,particle filtering algorithm for target tracking and tracking algorithm based on kernel correlation filter is improved.Besides,the author significantly reduces the intraoperative background interference to the stability of the system and improves the rapidity of multi-target tracking.By setting a reasonable threshold of buffer area,we can eliminate the mechanical error of the device itself,avoid the operation of Micro Devices beyond the surgical field of vision and the collision between it and laparoscope,and improve the safety performance of the system.Based on the above contents,we build a physical experiment platform design.From the view of machine vision,closed-loop feedback module is designed under the Vc++ development environment.We also design the control module of mirror arm system under Labview platform.Then through the quick communication between the two sets,designing the control system of automatic adjusting servo closed-loop attitude laparoscopic based on machine vision.The sensitivity of the system,the performance of anti-background interference,and the rapidity and stability of the automatic adjustment of the surgical field are tested through experiments.And the result proves that the method mentioned in this thesis is practical and has some advantages.