中文摘要：

针对工业现场中复杂背景下弱小目标检测的难题,提出了一种基于DSP的双层流水线的弱小目标检测新方法。该方法以双层流水线算法为背景,以半因果支持域背景预测为辅助,以DSP硬件图像处理系统平台为核心。利用C语言编写检测弱小目标的双层流水线算法,通过DSP集成开发软件Code Composer Studio 3.3进行移植、编译、调试及优化,将调试无误的算法程序烧写到DSP中,应用于高温镁熔液中弱小目标的检测。实验结果表明基于DSP的双层流水线方法可以有效检测出弱小目标,为实现工业现场弱小目标的实时检测奠定了基础。

英文摘要：

It is difficult to detect dim and weak target in complex background on industrial scene, a new method with double-pipeline based on DSP was proposed in order to solve the problem. It was set in double-pipeline algorithm, based on semi causal support domain background prediction and centered on DSP image processing system platform. Firstly, a double-pipeline algorithm used for dim and weak target detection was programmed with C language. Then, it was operated by transplantation, compilation, debugging and optimization on Code Composer Studio 3.3, which was an integrated development environment for DSP. Lastly, debugged algorithm program was rewritten into DSP, which was applied on the dim and weak target detection in magnesium melt. Experimental result shows that double-pipeline algorithm based on DSP could detect the dim and weak target. It also lays the foundation of real-time detection of dim and weak target on industrial scene.