基于多信息融合的籽棉异纤识别分拣系统设计与实现

Sorting System for Seed Cotton Fiber Based on Multi-information Fusion

目前国内棉花分拣技术较国外而言还有很大差距,棉花在加工过程中异纤检测效率不高,杂质残留较多。提出籽棉异纤识别分拣的概念,通过彩色线阵相机、黑白相机、高光谱成像仪在不同光谱环境下对棉花信息进行采集并通过卷积神经网络进行分类识别。实验结果表明,此分拣系统能够有效地识别异纤,解决棉花异纤含量高的问题。

In this paper,the concept of seed cotton fiber sorting is proposed.The cotton information is collected by color linear array camera,black and white camera and hyperspectral imager in different spectral environment and sorted by convolution neural network.The experimental results show that this sorting system can effectively identify the different fiber,to solve the problem of high content of cotton fiber.