Multiple windows algorithm for event detection in STFT-BOTDR

Short-time Fourier Transform (STFT) as a class of the time-frequency analysis have been utilised to deliver a fast-speed, accurate and low-variation BOTDR (Brillouin Optical Time Domain Reflectometer) system. The different window functions applied in the signal processing are important to improve the spatial resolution, frequency resolution, and variance of the BOTDR and to detect small event in the real application such as strain or temperature events smaller than the pulse length along the fibre. Window functions with long duration can achieve high frequency resolution, but be difficult to detect events less than the window length. While, small-length window offers better event detection performance than the long-duration window, but induces high variation along the fibre under test (FUT). For instance, the rectangular window offers a better frequency resolution; and the hamming window is better at the event detection. This paper introduces an optimised algorithm by utilising two separate windows and comparing the Brillouin frequency difference along the distance domain between the results generated by different windows to enhance the capability of small event detection, such as cracks and sharp temperature variation along the fibre.