Single Fiber Scanning Endoscopic Imaging System Based On Deep Learning

Single fiber scanners have been widely applied in miniature endoscopes,industrial nondestructive testing,Pico projector because of their versatility,compact size,large field of view and cylindrical design.It can use the piezo ceramic tube to control the sport of the fiber cantilever.When the fiber cantilever is vibrating,the light from the fiber will scan the object point.Then,the reflection light from the object will be collected by the concentrator.Based on those received light signal and fiber trajectory,the image can be recovered.However,the distorted image often appeared during the scanning imaging process.In this article,we study the single fiber scanning endoscopic imaging system and the key technology of the system.The article made a review of the endoscope and introduced the research status and significance of single fiber scanner.Then we analyzed the characteristic and principle of piezoelectrical ceramic actuator.There also shows the preparation of the single-fiber scanning probe and the workflow of the scanning probe imaging.An experiment which tests the resonant frequency and performance of the probe and recover the object image was introduced.Finally,we introduced the self-calibration scheme of fiber trajectory.It based on the positive piezoelectric effect of piezoelectric ceramics and get the feedback signal of the motion state of the fiber cantilever.The method of extracting the feedback signal is described.Due to the extracted feedback signal is mixed with the residual driving signal,we introduced a deep learning method to predict the motion trajectory of the fiber.The accuracy of the model and prediction results was verified by recovered the image,and the distorted image can be corrected.This technology enables the single fiber scanning probe to self-calibrate during the imaging process and predict the fiber scan trajectory in real time.With the help of this technology,single fiber scanner can change the driving signal to maintain a fixed field of view and reconstruct clear,distortion-free images.The error between the predicted optical fiber trajectory and the actual optical fiber trajectory is low.In addition,the self-calibration algorithm has good robustness to temperature changes,which improves the performance and stability of single fiber scanner in complex environments.