Superresolution Of Millimeter-wave Radiation Image And Edge Detection Algorithm

Millimeter-wave radiometric imaging technology is widely used. Apart from military and medical uses, the millimeter-wave radiometric technology has found applications in new civil fields closely related to national economy in recent years. However, in the procedures of capturing millimeter-wave radiometric images, the degradation of image quality is caused by a lot of factors, such as system noises, motion blurring and non-ideal sampling. Demands on the quality of images sometimes can not be reached and characteristics of target scenes can not be reflected clearly and accurately, what is worse, the proceedings of scientific research are badly affected.In practical situation, the quality of images could be improved by ameliorating hardware conditions of imaging system, but this way is technically difficult and costly. The quality could also be improved by using signal processing methods to restore the collected low quality images. Compared with the former way, without changing hardware conditions, millimeter-wave radiometric images restoration technology comes to be a research hotspot.Firstly, the basic principal and system model of millmeter-wave radiometric imaging were introduced in this paper. Secondly, the image restoration theory and normal methods, especially the super-resolution image restoration technology were introduced. The research on the POCS super-resolution restoration algorithm was the major part of this papar. With respect to the performance of super-resolution restoration algorithm, the significant effect of motion estimate was discussed. Taking MSE and PSNR as the evaluation criterion to judge image restoration effect, a series of simulation experiments were made to indicate the advantages of POCS algorithm on restoration effect compared with some traditional image restoration algorithms and other super-resolution algorithms. After analyzing how the parameters influenced the performance of POCS algorithm, solutions were made to answer how to make parameter preferences in practical application. At last, the POCS algorithm was applied onto millimeter-wave radiometric images in the field of security control, and an adaptive binarization edge detection method based on hierarchical clustering for millimeter-wave radiometric image was discussed. The hidden objects can be detected from the body background image effectively by this new method.