Research On Data Compression In Pipeline Leak Detection And Location System

Pipeline leak detection location system is a typical device for pipeline maintenance, which can efficiently reduce the resource waste and the environment pollution caused by pipeline leak. In order to detect the weak leak signals or the leak signals emitted from a long distance, the system needs a long time of data acquisition. As a result, the amount of the acquired data will be very large. With the improvement of the sampling rate or A/D resolution, the data will be larger. However, the storage capacity in data acquisition unit is very limited, which turns out to be a difficult problem for data storing and processing. So it is desirable to adopt data compression technique to sort out this problem.Various data compression techniques are studied and summarized in this paper, including the traditional and the newly developed techniques, then analyses the features of pipeline leak signals, such as the mechanism of production, entropy and correlation coefficient, are analyzed, and finally a universal and low-complex lossless compression algorithm is proposed and implemented in the pipeline leak detection and location system. This algorithm is composed of the predictor and Rice coding based on context modeling. The experimental results show that the compressed average bit is 8.3674, 1.2442 bits less than the original entropy, and the efficiency of Rice coding achieves 98.81%. And the algorithm also has the advantage of real-time encoding in the system.In order to improve the compression ratio of pipeline leak signals, an error-controlled quantization method is designed in this paper, and on the basis of lossless compression algorithm, a near-lossless compression algorithm is proposed and implemented in pipeline leak detection and location system. This algorithm adopts peak error or maximum amplitude error criterion to control the level of distortion in pipeline leak signals, and doesn't change the subsequent location error of adaptive time-delay estimation in the case of a small quantization error, so as to implement the near-lossless compression of pipeline leak signals. When the quantization error is 5, the compressed average bit reach 4.8612, i.e. the compression ratio is 2.47, which overcomes the limitation that the compression ratio is difficult to exceed 2.According to the characteristics of pipeline leak signals that the influence degrees of burst interference noises are various in each local area, this paper optimizes the uniform quantization to be adaptable, and proposes an adaptive quantization...