基于多尺度局部对比度和多尺度梯度一致性的红外小弱目标检测算法

针对复杂背景和强杂波干扰下红外小弱目标检测虚警率高的问题，提出了一种基于多尺度局部对比度方法与多尺度梯度一致性方法的红外小弱目标检测算法。利用多尺度局部对比度方法对红外图像中红外小弱目标进行增强，利用多尺度梯度一致性方法剔除复杂背景和强杂波干扰造成的虚警。从信噪比（SNR）增益、平均残留背景绝对值、检测率、虚警率及ROC曲线方面将新算法与max-mean算法、max-median算法、top-hat算法、IPI算法及MGDWIE算法进行了对比。实验显示：新算法相较于对比算法具有更高的SNR增益、更低的平均残留背景绝对值、更高的检测率及更低的虚警率。对比结果表明：新算法在复杂背景和强杂波干扰下具有良好的红外小弱目标检测准确性和鲁棒性，有效改善了复杂背景和强杂波干扰下红外小弱目标检测虚警率高的问题。

Infrared Dim Small Target Detection Based on Multi-scale Local Contrast and Multi-scale Gradient Coherence

A novel algorithm based on multi-scale local contrast method and multi-scale gradient coherence method is proposed for the high false alarm of infrared dim small target caused by complex background and heavy clutter. The multi-scale local contrast method is used to enhance the infrared dim small target in infrared image, and the multi-scale gradient coherence method is used to avoid the false alarms caused by complex background and heavy clutter. The proposed algorithm is compared with max-mean, max-median, top-hat, IPI and MGDWIE algorithms in terms of signal-to-noise ratio（SNR）gain, mean absolute value of residual background, detectivity and false alarm rate. The proposed algorithm achieves higher SNR gain, lower mean absolute value of residual background, higher detectivity and lower false alarm rate compared to the baseline algorithms. The experimental results show that the proposed algorithm can effectively reduce the false alarm rate under the disturbance of complex background and heavy clutter.