提出了基于灰度图像直方图均衡与梯度法相结合的技术，应用于相干光时域反射计(COTDR)光纤传感信号处理，以COTDR 同一光脉冲所探测得到的信号为行向量，以一组光脉冲探测得到的同一空间位置的信号为列向量，构建了相干传感信号的二维灰度图像，通过计算灰度图像的梯度，并结合图像直方图均衡处理，降低了由环境干扰所引入的噪声信号，提高了系统对振动信号的分辨力。采用该技术，基于传感光纤长度为2.5 km 的COTDR 系统，有效地识别了压电陶瓷所施加的振动信号。

The technique of combining histogram equalization and gradient method was proposed for the processing of optical fiber sensor signal in a coherent optical time domain reflectometer (COTDR). Treating the signals detected by the same optical pulse sent into the sensor fiber as a row vector, and the signals of the same position detected by a set of optical pulses as a column vector, a two- dimensional grey scale image is built. Through calculating the gradient of the grey scale image, and then histogram equalization, the noise resulted from the environmental interference is depressed. The resolution is improved. With the proposed method, the vibration signal driven by a piezoelectric ceramic transducer in the COTDR with a 2.5 km length sensing optical fiber is identified well.