On Intelligent Colonoscope Shape Display System

Loop formation of colonoscope often happens in diagnosis of colon because of complexity of working condition and non-visualization of shape of flexible rod and it brings a great pain and danger to the patient. So it is necessary to use intelligent sensor and visualization technique to innovate performance of traditional colonoscope and realize shape display of flexible rod and pose detetion of handling part of colonoscope.Shape reconstruction theory, design and packing method of flexible rod based on Fiber Brag Grating sensing theory was adopted in this paper. Considering shift of wavelength was affected by temperature and strain, it is necessary to compensate temperature by using layout of FBG sensor array. The shape reconstruction algorithm based on strain information about flexible rod was derived from plannar curve to spatial curve, simulation experiment and real experiment has demonstrated that it is feasible. Because actual packing position of FBG point isn't consitent with ideal packing position, good accuracy of shape reconstruction of flexible rod was obtained by rectifing packing position of FBGs.Considering multiformity of flexible rod, nonlinear fitting method named median segment algorithm was proposed to solve the curvature between two FBGs. Experiment results shows that accuracy of reconstruction was improved comparing with linear fitting method.However, shape reconstruction of colonoscope is on the assumption that handling part of colonoscope is fastened, but in actual manipulation it is used to advance into the body and control direction of tip and orientation of handling part of colonoscope, so it would change the shape and orientation of colonosope. In this paper a detecting method was proposed to detect the orientation of handling part of colonosope based on binocular vision of markers and estimate the detection error how to affect shape of colonoscope was discussed deeply.For pose detection of handling part of colonoscope, this research mainly describes visual-tracking principle, layout of markers, camera calibration and setting up of coordination sytem. A method was proposed about solving the pose relation between two cameras. To improve the tracking speed, a fast Connected Components Labeling (CCL) algorithm based on Rosenfeld's method was proposed and the speed about extracting the markers was impoved greatly. Using markers in image coordinates system, calibration parameters of two cameras and the pose relation about two cameras, image points about two cameras in left and right coordinate system can be converted into the same coordinate system, so solving of the spatial point can be simplified.Because many factors have an effect on accuracy of spatial point, position of markers relative to global coordinate system has an error which will degrade the pose accuracy relative to global coordinate system. This paper describes traditional solving method and proposes a method about pose estimation and error propagation based on principal component analysis and least square method.In the integration display system of colonoscope, this paper discuss the pitch, yaw and roll of handling part how to affect shape and pose of flexible rod. It is solved that how to rectify the flexible rod shape when pose of handling part of colonoscope is changing.An experimental system has been written using VC++.NET developing platform in Windows XP operation system. The experimental result indicates that the system can realize shape real time display of flexible rod on the basis of pose detetion of handling part of colonoscope.