论文部分内容阅读
针对以离散余弦变换为核心的人类视觉模型舰船检测算法受数据类型限制的问题(即对复数类型的数据检测效果不好),该文提出了一种改进的人类视觉模型SAR图像舰船检测算法。该算法是以快速傅里叶变换代替离散余弦变换,将SAR图像从空间域变换到频率域;快速傅里叶变换对数据类型要求较低,只要求数据是离散的,并且运行效率更高。然后,采用3种星载SAR数据——ENVISAT ASAR(25m)、Sentinel-1(10m)和Cosmo-Skymed(2.5m)进行对比实验。结果表明,以快速傅里叶变换为核心的人类视觉模型舰船检测算法的检测性能和效率优于以离散余弦变换为核心的算法、双参数恒虚警率(CFAR)算法和K分布恒虚警率算法。
Aiming at the limitation of data type limitation of human vision model ship detection algorithm based on Discrete Cosine Transform (ie, the data detection of complex type is not effective), this paper presents an improved human vision model SAR image ship detection algorithm. The algorithm uses fast Fourier transform instead of discrete cosine transform, the SAR image from the spatial domain to the frequency domain transform; fast Fourier transform less demanding on the data type, only requires that the data is discrete, and run more efficiently. Then, three kinds of spaceborne SAR data were used - ENVISAT ASAR (25m), Sentinel-1 (10m) and Cosmo-Skymed (2.5m) for comparative experiments. The results show that the detection performance and efficiency of the human vision model ship detection algorithm based on Fast Fourier Transform are superior to those based on discrete cosine transform. The CFAR algorithm and K-distribution constant Alarm algorithm.