以数字信号处理(Digital Signal Processing, DSP)平台实现图像处理的应用开发一般较复杂, 不利于复杂背景下弱小目标的检测。为提高检测程序执行的高效性和稳定性, 提出了基于 DSP/BIOS图像处理的弱小目标检测方法。该方法以高性能数字媒体处理器 DM642为核心, 以 DSP/BIOS内核和动态调用 API函数为辅助, 以工业现场复杂镁熔液中弱小目标检测为载体, 并通过数字 I/O口的输出信号来判断检测的目标。经过不同条件下 400次实验结果表明, 综合检测率高达 90.5%, 与单线程 DSP检测算法相比, 检测率提高了 11.04%, 检测程序的执行效率和稳定性都得到提高。

The application and development of image processing methods that are based on the digital signal processing platform are generally complex processes; these are also not feasible for the detection of weak and small targets in a complex background. In order to improve the efficiency and stability of a detection program, a method for the detection of weak and small targets based on DSP/BIOS image processing was proposed. In this method, the high performance of the digital media processor DM642 was considered to be the core, and the DSP/BIOS kernel and dynamic call API function were the auxiliary modules. Weak and small targets were detected in an industrial scenario where magnesium melt was used as the carrier in a complex background; the estimated target was determined by observing the output signal of the digital I/O. The results of 400 experiments under different conditions showed that detection rate was 90.59%, which is 11.04% higher than that of a single thread DSP detection algorithm; moreover, the efficiency and stability of the detection program also improved.