基于岭回归的毫米波辐射图像的复原研究

Research on Millimeter Waves Radiometric Image Restoration Based on Ridge Regression Algorithm

毫米波辐射计具有全天时全天候的工作能力,可提供红外与可见光探测器不能获取的特殊信息。在分析低分辨率毫米波辐射成像主要退化因素的基础上,采用基于最小二乘的岭回归算法反演地物目标的视在温度分布,利用水平和垂直方向的视在温度分布来获取毫米波辐射图像的细节纹理。以光学图像为参考3mm波段的实际辐射成像复原结果表明,基于岭回归的3mm波段的辐射图像纹理细节清晰,空间分辨率获得了较大的改善。

Millimeter wave radiometer working at all day and all weather can acquire specific information that infrared or optical detectors can＇t acquire.On basis of analyzing the degeneration factor of low spatial resolution milimeter wave radiometric image,a least square ridge regression image restoration method is presented adopting horizontal or vertical brightness temperature distiibution.The results with contrast to the optical image show 3mm radiometric image texture and detail is vivid and acquires higher spatial resolution by the ridge regression algorithm.