Fiber optic temperature sensor based on image processing of intermodal interference pattern

Few-mode interference patterns produced by a fiber optic are composed of lobes whose intensity and displacement in the fiber cross-section are dependent on external disturbances like temperature and strain. This paper presents an image processing algorithm which computes the lobes coordinates and records their trajectory with respect to time. It is then numerically and experimentally demonstrated that the trajectory length of path is proportional to the temperature integrated all along the fiber (only thermal disturbance is considered). The heating and cooling phase discrimination issue is solved by inserting a cyclically heated fiber sample in the fiber under test and by analyzing the associated trajectory impact. A frequency-domain analysis has been used for suppressing vibration-induced disturbances. A low-cost experimental trial setup and the corresponding results are showing an accuracy of 1°C.