Microrobotic Endoscopy:Visual 1-Point RANSAC EKF-SLAM Localization Inside The Large Intestine

Although the Wireless Capsule Endoscopy(WCE)technology gained the trust of patients,still its localization and the control problems are big challenges area for researchers.As a critical part in WCE examination,the doctors need to know the exact position of it,to find the position of gastrointestinal disease after being detected by the video source.To find WCE position,we need to have a map inside the human body.However,since the shape of the gastrointestinal(GI)is very complex and RF signals diffuse differently in heterogeneous body tissues,correct mapping and validate localization algorithms inside the human body are very difficult and challenge.In this work,we represent validation for motion tracking simultaneous localization and mapping technique(SLAM)of WCE inside the large intestine of the human body inside extended Kalman filter(EKF)framework based structure from motion(SfM),the SfM algorithm is robust estimation that accommodates all accessible prior probabilistic data sources from a sole sequence input of a set of images,and use 1-Point RANSAC to reduce the sample size to 1 to enhance the positioning accuracy of the WCE inside the large intestine and reconstruct the WCE trajectory has traveled.In this way,the positions of the gastrointestinal diseases can see accurately on the map inside the human body,and thus,facilitates the follow-up of therapeutic processes.The proposed approach takes advantage of data fusion from two sources that come with the WCE,the sequences images which captured by the belt-in camera and will rack the RF signal emitted from the WCE.This approach estimates the speed and orientation of the endoscopy capsule by analyzing the displacement of feature points(FPS)between the image sequence and then combining this movement information with RF measurements by employing an extended Kalman filter(KF)to facilitate the localization results and generate the route in which the capsule has traveled.The performance of the proposed motion tracking algorithm validated using established a virtual test-bed that emulated the large intestine of the human being digestive system channel characteristics,under this emulation environment and constant velocity model.Experimental results show that the 1-Point RANSAC EKF-SLAM localization algorithm in-Body can give,more accurate motion tracking of WCE and reconstruct the 3D map has traveled,with high typical outlier rate in real-time with the mean absolute error about 0.0161,and after we combined EKF-SLAM with RF localization by using Kalman filter KF it became 0.0098363and the distance error about 1.1029.The work performance measurements and the comparisons with existing works depict the robustness and applicability of the proposed methodologies.