CAJ | 학술논문

针对红外单帧图像的小目标检测问题，提出基于全变分理论的背景杂波抑制的改进算法。构建以能量函数极值最优化为主体的全变分模型，结合自适应步长和非单调性线搜索条件，形成改进的全变分算法，在红外图像背景抑制方面取得良好效果。经过仿真实验测试几种算法的抑制效能，分析结果表明改进算法的检测性能较传统算法在杂波抑制、减少边缘虚警等方面均有优势，能够满足实时性要求，且适用于复杂背景的工程实现。
침대홍외단정도상적소목표검측문제，제출기우전변분이론적배경잡파억제적개진산법。구건이능량함수겁치최우화위주체적전변분모형，결합자괄응보장화비단조성선수색조건，형성개진적전변분산법，재홍외도상배경억제방면취득량호효과。경과방진실험측시궤충산법적억제효능，분석결과표명개진산법적검측성능교전통산법재잡파억제、감소변연허경등방면균유우세，능구만족실시성요구，차괄용우복잡배경적공정실현。
An improved algorithm of background clutter suppression based on total variation theory is proposed for infrared small target detection in a single frame image under complicated background. Combined with adaptive step size and non-monotone line search condition, the improved model whose main body is an optimization problem of energy function can achieve good results for background clutter suppression of infrared image. The simulation results about inhibition efficiency show that such method is remarkably improved compared with traditional methods in cluttering suppression and reducing false-alarm on edges. It can not only satisfy real-time requirement, but also its implementation is suitable for engineering fulfillment of infrared image under complicated background.