摘要
针对传统纸张计数无法达到精准、无损、高效识别纸张数目的不足,设计了一款基于FDC2214分段线性回归纸张识别装置。该装置采用平行板电容法原理,以STM32F103C8T6为主控单元、FDC2214电容传感器和金属板作为数据采集单元、以OLED屏幕作为显示单元。通过数据采集单元对电容值进行采集,在MCU里利用分段式回归算法进行运算实现训练、识别。同时引入简单机器学习算法进一步提高了整个装置的稳定性和准确性。在数据处理时用卡尔曼滤波、温度补偿等算法进行误差补偿,提高了整个装置抗干扰能力。该装置在训练模式下对纸张数目进行学习并对数据储存,然后在测试模式下经过算法运算、数据比对从而准确输出纸张数目。经过实验证明,该装置能够达到准确、稳定、高效地输出纸张数目。
Aiming at the situation that the shortage of traditional paper counting can not achieve accurate,lossless and efficient identification of paper number,a paper identification device based on FDC2214 piecewise linear regression is designed.The device adopts the principle of parallel plate capacitance method,with STM32F103C8T6 as the main control unit,FDC2214 capacitance sensor and metal plate as the data acquisition unit,and OLED screen as the display unit.The capacitance value is collected by the data acquisition unit,and the training and recognition are realized in MCU by using the subsection regression algorithm.At the same time,a simple machine learning algorithm is introduced to further improve the stability and accuracy of the whole device.In data processing,Kalman filter,temperature compensation and other algorithms are used for error compensation,which improves the anti-interference ability of the whole device.The device learns the number of paper and stores the data in the training mode,and then outputs the number of paper accurately through algorithm operation and data comparison in the test mode.The experiments show that the device can output the number of paper accurately,stably and efficiently.
出处
《科技创新与应用》
2020年第31期84-86,共3页
Technology Innovation and Application
关键词
分段线性回归算法
FDC2214
纸张识别
电容采集
卡尔曼滤波算法
温度补偿
机器学习
piecewise linear regression algorithm
FDC2214
paper identification
capacitance acquisition
Kalman filter algorithm
temperature compensation
machine learning