摘要
高效率的目标检测是视觉应用的重要技术,但运动目标的提取易受环境的影响。关联成像能够解决特殊环境下难以获得清晰图像和一些常规成像技术不易解决的问题。在目标检测中,利用关联成像采集图像信息并运用背景差分法在压缩域中获得目标图像的测量值,直接通过压缩感知重构出目标图像。这种方法可以解决在特殊情况下无法检测到目标的问题,同时检测到的目标图像清晰,采样次数少,信噪比也较高。
Efficient target detection is an important technology for many vision applications,but the extraction of moving targets is easily affected by the environment.Ghost imaging can solve the problem of capturing a clear image in a special situation.It can solve the problems for conventional imaging techniques.In target detection,the image information is firstly captured by ghost imaging.The measured values of target image are obtained in the compressed domain.The target image is reconstructed by compressive sensing.The method can solve the problem that the conventional method can't detect the target in a special environment.The reconstructed target image is clear and has less number of samples.The signal noice ratio is higher.
出处
《光学仪器》
2017年第6期1-6,共6页
Optical Instruments
关键词
压缩感知
关联成像
背景差分
目标检测
compressive sensing
ghost imaging
background subtraction
target detection
