【摘要】 为了对超声自动检测过程中产生的大量A波数据进行有效压缩,利用A波数据的生成原理和特点,从中提取波形沿、波形和波间距等特征,并给出了基于特征的数据压缩模型和算法.确定波形和波间距的起点和终点,根据阈值求解每波形中相邻波形沿的交点,将波形分解成两个或者两个以上的波形沿,构造一条最佳直线来拟合波形沿.实时压缩后可利用各特征重构出A波数据,如未达到预定的压缩性能指标则可调整压缩参数后进行再次压缩.工程应用表明,该方法能够在2%以内的重构误差下,达到100以上的压缩比,相比其他压缩方法,具有方便判别缺陷波的优点.更多还原

【Abstract】 A feature-based data compression algorithm was proposed based on the production theory and characteristics of A-wave data to compress the massive A-wave data produced in the automatic ultrasonic inspection process.A-wave was disassembled into features such as wave slope,waveform and wave gap,then the ranges of each waveform and wave gap were figured out.The intersection points of adjacent wave slopes among each waveform were obtained according to the given thresholds,and a waveform was disassembled into two wave slopes or more.An optimal line was constructed to fit the wave slopes.A-wave data was reconstructed by reading features during the run-time compression.The thresholds were adjusted to recompress if the compression performance evaluation parameters were unqualified.Engineering application showed that the data reconstruction error of the algorithm was less than 2% even when the compression ratio reached 100.The algorithm is convenient to judge flaw waves compared with other algorithms.更多还原