针对海天场景复杂干扰情况下多尺度检测红外偏振图像中舰船目标困难的问题, 本文提出一种基于引导滤波和自适应尺度局部对比度的舰船目标检测方法。首先将强度信息作为引导信息对红外偏振图像利用引导滤波, 得到目标背景对比度、局部信噪比更高的融合图像; 然后基于融合图像显著的海天线垂直梯度特征, 提出一种检测海天线方法, 再对融合图像进行海天线加权抑制海杂波干扰; 最后基于单尺度局部对比度算法与舰船目标比例特征, 提出自适应尺度局部对比度方法, 当尺度与目标匹配时响应最大, 通过不同尺度对目标的响应结果确定最大尺度, 得到舰船目标检测结果。实验结果表明, 引导滤波融合方法的提高图像的目标背景对比度和局部信杂比, 与典型检测方法对比, 本文方法能够有效抑制干扰并能够检测海天场景不同尺度舰船目标, 具有较高的鲁棒性和准确性, 检测率、虚警率分别为95.0%, 3.5%, 为红外偏振图像目标检测提供了新的方法。

Aiming at the difficulty of multi-scale detection of ship targets in infrared polarization images under complex interference conditions of sea-sky scene, a ship target detection method was proposed based on guided filtering and adaptive scale local contrast method. First, by using the intensity information as the guiding information to guide and filter the infrared polarization image obtained the fusion image with higher local contrast and local signal-to-clutter ratio. Then, based on the vertical gradient characteristics of the fusion image, detecting the sea-sky line method was proposed so that sea-sky line weighting suppressed the sea clutter on the fused image. Finally, based on the single-scale local contrast method and the ship target proportional feature, the adaptive scale local contrast method was proposed. When the scale matched the target, the response was the largest, and the maximum scale was determined by the response of the target at different scales. The experimental results show that the local contrast and local signal-to-noise ratio of the image are improved by the guided filtering fusion method. Compared with the typical detection methods, our method can effectively suppress interference and detect different scales of ship targets in sea-sky scene with high robustness and accuracy. The detection accuracy ratio and false alarm ratio are 95.0% and 3.5%, respectively, which provides a new method for infrared polarization image target detection.