基于DSP技术的钨矿初选系统的研究

Design of Primary Sorting System of Tungsten Ore Based on the DSP

针对钨矿初选技术的现状,搭建以TMS320DM642为核心的嵌入式钨矿初选系统。利用机器视觉和DSP技术,设计以改进的全局阈值分割算法为主体的图像处理以及矿苗定位算法,完成原矿图像采集、图像特征提取、矿苗识别及定位的钨矿初选系统。实验结果表明,该系统可以实现实时、非接触智能化的钨矿初选,同时具有体积小、稳定性高、功耗低、成本低廉的特点,具备较高的实际应用价值。

In view of the current situation of primary sorting technology of tungsten ore, a Primary sorting system of tungsten ore was built based on the embedded TMS320DM642. The image processing and outcropping lo- calization algorithm were proposed by using machine vision and DSP technology, which was based on the improved global threshold segmentation algorithm. Ore image acquisition, image feature extraction and outcropping recognition and lo- cation were realized. Experimental results showed that the proposed system can realize the real-- time, non-- contact in- telligent primary sorting of tungsten ore, and also has small size,high stability,low power consumption, low cost, so has a high practical value.