A non-local means algorithm-based Raman-OTDR for high-performance distributed temperature sensing

Raman-Optical-Time-Domain-Reflectometry (R-OTDR) based Distributed Temperature Sensing (DTS) is of great significance in extremely harsh environment. In this work, we propose a Non-Local Means (NLM) algorithm for R-OTDR performance improvement without modifying the hardware architecture. The NLM algorithm real-time reconstruct the two-dimensional temperature image by utilizing a first-in-first-out data transmission protocol. These images are incorporated to mitigate the complexity of calculating Euclidean distances between neighborhoods within two-dimensional spatial-temporal domains. This approach significantly improves the response speed of the system. To verify the sensing performance improvement of the NLM, we built a standard R-OTDR system and affirm the seamless continuity of temperature variations spanning a broad range from 40 °C to 80 °C over an extensive distance of 9.68 km assisted by a standard platinum resistor. The temperature accuracy decreases from 1.4 °C to 0.55 °C at sensing temperature of 100 °C. The results effectively reconcile the conventional trade-off between high temperature resolution and high-speed response. What's more, an average temperature resolution is 0.05 °C at the end of fiber. In all, the Raman DTS system based on NLM algorithm achieve high temperature resolution. This achievement holds immense potential for diverse applications, including dynamic monitoring in the realms of energy development, oil and gas exploration, allied fields, and so on.