提出了考虑整体灰度和局部轮廓信息的图像匹配新算法,直方图粗匹配结合模板精匹配,具有计算速度快,匹配精度高的特点.设计并实现了嵌入式模具安全保护系统的硬件结构与软件架构.为验证该系统的准确性与可靠性,开展了控制实验及实际注射成型实验研究.实验结果表明,该系统可分辨出小于1 cm2的异物,计算匹配时间小于200 ms,在实际成型中可检测出肉眼难以分辨的透明塑料零件.%A novel image matching algorithm was proposed,according to the information of integral grayscales and partial contours.Combining a rough histogram matching method and an accurate template matching method has advantages of high calculation speed and high matching precision.A hardware structure and a software architecture for the embedded protecting system of mold were designed and developed.For the verification of the accuracy and the reliability of the developed system,a control experiment and an injection molding experiment were performed.Experimental results show that the system can distinguish a less than 1 cm2 abnormal object,and it only takes less than 200 ms.In addition,the system can detect a transparent plastic part,which can hardly distinguish by naked eyes.
展开▼