Endoscopes shape reconstruction based on electromagnetic localization and curve fitting

Magnetic endoscope imaging system is a non-radiographical imaging technique, which helps endoscopists a lot in colonoscopy. It consists of two parts: localization and display. In the part of localization, coils embedded in the colonoscope sense the magnetic field generated by excitation sources. The coupling between excitation sources and sensors can determine a set of equations which contain parameters of position and orientation information. Then, a nonlinear algorithm is given to solve these equations. After obtaining coils´ information and the actual length between each two adjacent sensors, a curve generating algorithm is used to fit the colonoscope curve. In this paper, we analyze the algorithm used in the localization and a curve fitting algorithm based on Bézier cubics, which can be used to reduce the number of sensors. Less sensors can reduce the complexity of the colonoscope manufacture process, because these sensors are embedded in the colonoscope. Besides, as curve reconstruction is based on the sensors´ information, it can help to reduce the time for gathering data and increase the frequency of image update, which is helpful to endoscopists. Our simulation results demonstrate a promising performance of the proposed scheme.