基于视觉注意机制的复杂背景下红外小目标检测

Infrared Small Target Detection Under Complex Background Based on Visual Attention

针对视觉注意机制Itti模型对复杂背景下红外小目标检测易受到图像背景杂波影响,检测结果不理想的情况,对传统算法加以改进,在Itti模型中引入背景预测算子。对图像背景进行预测,与原图像进行差减,以达到突出目标区域的目的,消除背景区域对目标显著性的影响。再提取滤除背景后的图像视觉差异,找出图像的显著性区域,实现对红外小目标的检测。将改进后的模型应用于复杂背景下红外小目标检测中,实验结果表明,相对于传统的Itti模型的检测算法,新提出的算法具有更高的检测率。

For the condition of infrared small target detection is vulnerable to background clutter and the results are undesirable by visual attention Itti model, the traditional Itti model is improved and background prediction is introduced. Firstly, image background is predicted and then subtracted from the original image to highlight the target area and eliminate back- ground effects on target region. Secondly, the visual difference of filtered background is ex- tracted to find out the significant target area and achieve the infrared small target detection. The improved model is applied to complex background infrared small target detection, the experimental results show that compared with traditional Itti model detection algorithm, the new proposed algorithm has a higher detection rate.