TCS230颜色传感器结合BP神经网络在颜色识别系统上的研究

Research on color identification by TCS230 color sensor combined BP neural network

针对TCS230颜色传感器对颜色识别的快速和高精度的特点,提出结合BP神经网络用于颜色识别系统的研究。采用该传感器搭建非接触式颜色检测电路,用于色卡集的RGB频率值的测量。利用2200个训练色卡集构建BP神经网络,28个色卡集测试该BP网络;为使目标颜色的隶属类同人眼观察到的模糊结果统一起来,采用CIE1976Lab均匀颜色空间及色差公式,用于目标色卡同黑、蓝、绿、青、红、紫、黄、白8种常见颜色的色差聚类计算。试验验证了颜色传感器TCS230采用BP神经网络技术用于颜色识别,同人眼的颜色感觉大体一致,能较好地反映人对物体颜色的心理感受,取得了良好效果,在食品加工行业有较好的应用前景。

Based on the fast speed and high accuracy color identification of TCS230 color sensor,a color identification system was proposed combined with BP neural network.A color detection circuit has been designed according to the sensor,as well as measuring 2,200 color card frequency value.By using the measured frequency of the color,the BP neural network was established by the corresponding RGB space.In order to integrate the target color class and the fuzzy results observed by eyes,CIE1976Lab uniform color space and color difference formula has been used for the target color card which includes black,blue,green blue,red,purple,yellow,white.The TCS230 color sensor technology combined with BP neural network used in color recognition largely has the same effect of eye-observation which has already been provide by experiment.Besides,it can reflect people＇s psychological feelings to objects well and has good results.Thus it has good prospect in food processing industry.