为了提高复杂背景下红外小目标的检测能力,提出了一种改进加权增强局部对比算法(weighted strengthened local contrast measure,WSLCM)的红外小目标检测方法。在WSLCM的预处理阶段采用自适应曲率滤波在多尺度下对图像进行处理,抑制背景过程中保持真实目标不被淹没。在背景抑制计算中选取背景块中的灰度最大值作为背景估计,降低虚警率。同时引入目标增强因子与背景抑制因子,加强算法鲁棒性,消除背景噪声影响,增强对红外小目标的检测能力;通过嵌入式ZYNQ平台进行算法IP核定制,采用软硬件协同方式实现算法对特定场景下小目标的识别检测。实验表明,相比于传统WSLCM算法BSF和SCRG指标都有明显提高,连续帧检测率为93.2%,嵌入式平台检测效率比PC端提升17.6%,验证了算法和嵌入式系统的有效性。

In order to strengthen the detection capability of small infrared targets under complex background, an improved WSLCM(weighted local contrast measure) detection algorithm is proposed. In the pre-processing stage of WSLCM, adaptive curvature filtering is used to process the image in multi-scale, and the real object is not submerged in the process of background suppression. In the background suppression calculation, the maximum gray value of the background block is selected as the background estimation to reduce the false alarm rate. At the same time, target enhancement factor and background suppression factor are introduced to enhance the robustness of the algorithm, eliminate the influence of background noise, and enhance the detection ability of infrared small targets. The algorithm IP verification is carried out by the embedded ZYNQ platform, and the algorithm can recognize and detect the small target in the specific scene by using the hardware and software cooperation. Experiments show that compared with the traditional WSLCM algorithm, BSF and SCRG indexes are significantly improved, the continuous frame detection rate is 93.2%, and the detection efficiency of embedded platform is 17.6% higher than that of PC, which verifies the effectiveness of the algorithm and embedded system.