实际生产过程中很多因素会对板形信号产生干扰,从而造成板形缺陷识别不准确。针对这一问题,在板形勒让德多项式回归分解识别方法基础上应用最速离散跟踪微分器对板形信号进行滤波处理,消除噪声干扰。仿真结果表明,两种方法相结合后在板形缺陷识别的精确度和抗干扰性方面都有很好的效果。 Many factors in the actual production process will interfere with the flat-panel signal, resulting in inaccurate identification of flat-shaped defects. Aiming at this problem, the steepestral discrete tracking differentiator is used to filter the flat signal on the basis of the Regal Polynomial Regression Decomposition and Recognition method to eliminate the noise interference. The simulation results show that the combination of the two methods has a good effect on the accuracy and anti-interference of the flaw detection.