Alternating Direction Projections onto Convex Sets for Super-Resolution Image Reconstruction

Image reconstruction is important in computer vision and many technologies have been presented to achieve better results. In this paper, gradient information is introduced to define new convex sets. A novel POCS-based model is proposed for super resolution reconstruction. The projection on the convex sets is alternative according to the gray value field and the gradient field. Then the local noise estimation is introduced to determine the threshold adaptively. The efficiency of our proposed model is verified by several numerical experiments. Experimental results show that, the PSNR and the SSIM can be both significantly improved by the proposed model.

Download Full-text

Spatial Scale Effect of a Typical Polarized Remote Sensor on Detecting Ground Objects

Sensors ◽

10.3390/s21134418 ◽

2021 ◽

Vol 21 (13) ◽

pp. 4418

Author(s):

Ying Zhang ◽

Jingyi Sun ◽

Rudong Qiu ◽

Huilan Liu ◽

Xi Zhang ◽

...

Keyword(s):

Image Reconstruction ◽

Spatial Scale ◽

Scale Effect ◽

Convex Sets ◽

Spatial Scales ◽

Super Resolution ◽

Reconstruction Method ◽

Mixed Pixel ◽

Resolution Image ◽

Remote Sensor

For polarized remote sensors, the polarization images of ground objects acquired at different spatial scales will be different due to the spatial heterogeneity of the ground object targets and the limitation of imaging resolution. In this paper, the quantitative inversion problem of a typical polarized remote sensor at different spatial scales was studied. Firstly, the surface roughness of coatings was inversed based on the polarized bidirectional reflectance distribution function (pBRDF) model according to their polarization images at different distances. A linear-mixed pixel model was used to make a preliminary correction of the spatial scale effect. Secondly, the super-resolution image reconstruction of the polarization imager was realized based on the projection onto convex sets (POCS) method. Then, images with different resolutions at a fixed distance were obtained by utilizing this super-resolution image reconstruction method and the optimal spatial scale under the scene can be acquired by using information entropy as an evaluation indicator. Finally, the experimental results showed that the roughness inversion of coatings has the highest accuracy in the optimal spatial scale. It has been proved that our proposed method can provide a reliable way to reduce the spatial effect of the polarized remote sensor and to improve the inversion accuracy.

Download Full-text

Projection onto convex sets super-resolution image reconstruction based on wavelet bi-cubic interpolation

Proceedings of 2011 International Conference on Electronic & Mechanical Engineering and Information Technology ◽

10.1109/emeit.2011.6022950 ◽

2011 ◽

Cited By ~ 7

Author(s):

Zhifei Tang ◽

Mike Deng ◽

Chuangbai Xiao ◽

Jing Yu

Keyword(s):

Image Reconstruction ◽

Convex Sets ◽

Super Resolution ◽

Resolution Image ◽

Projection Onto Convex Sets

Download Full-text

Research of Super-Resolution Image Reconstruction Methods

Applied Mechanics and Materials ◽

10.4028/www.scientific.net/amm.687-691.3782 ◽

2014 ◽

Vol 687-691 ◽

pp. 3782-3786

Author(s):

Ling Tang

Keyword(s):

Image Processing ◽

Computer Vision ◽

Image Reconstruction ◽

Development Process ◽

Super Resolution ◽

Resolution Image ◽

Reconstruction Methods ◽

Research Findings

The super-resolution image reconstruction has become a hot topic in the areas of image processing and computer vision because of its extensive theoretical and practical values. This paper described the concept of super-resolution reconstruction, reviewed the development process of the technique, common algorithms classification, the current research findings and other related issues. The characteristics of different algorithms are also analyzed.

Download Full-text