论文部分内容阅读
分析了基于Frounhofer衍射的煤矿井下粉尘粒度分布检测的模式识别算法 ,针对决定反演精度的模式数量与传感器实时性之间的突出矛盾 ,提出了模式分类方案 ,使模式识别的速度提高到原来的若干倍 ,从而为实时反演算法的单片机运行提供了有力保障。
The pattern recognition algorithm based on Frounhofer diffraction for particle size distribution detection of coal mine dust is analyzed. According to the prominent contradiction between the number of models that determine the accuracy of inversion and the real-time performance of the sensor, a pattern classification scheme is proposed to increase the speed of pattern recognition to the original Several times, which provides a powerful guarantee for the real-time inversion algorithm of the single-chip operation.