Learning-based single image super-resolution (SISR) aims to learn a versatile mapping from low resolution (LR) image to its high resolution (HR) version. The critical challenge is to bias the network training towards continuous and sharp edges. For the first time in this work, we propose an implicit boundary prior learnt from multi-view observations to significantly mitigate the challenge in SISR we outline. Specifically, the multi-image prior that encodes both disparity information and boundary structure of the scene supervise a SISR network for edge-preserving. For simplicity, in the training procedure of our framework, light field (LF) serves as an effective multi-image prior, and a hybrid loss function jointly considers the content, structure, variance as well as disparity information from 4D LF data. Consequently, for inference, such a general training scheme boosts the performance of various SISR networks, especially for the regions along edges. Extensive experiments on representative backbone SISR architectures constantly show the effectiveness of the proposed method, leading to around 0.6 dB gain without modifying the network architecture.

Light-field (LF) camera holds great promise for passive/general depth estimation benefited from high angular resolution, yet suffering small baseline for distanced region. While stereo solution with large baseline is superior to handle distant scenarios, the problem of limited angular resolution becomes bothering for near objects. Aiming for all-in-depth solution, we propose a cross-baseline LF camera using a commercial LF camera and a monocular camera, which naturally form a 'stereo camera' enabling compensated baseline for LF camera. The idea is simple yet non-trivial, due to the significant angular resolution gap and baseline gap between LF and stereo cameras.

The capability to identify damage level of building due to destructive earthquake from RS image depends on the sensor type (for example optical or SAR) and spatial resolution. It is usually estimated by considering the object size and the pixel size, the number of pixels the object covers. Such comparison does not show effective capability of damage grade identification. The authors have developed quantitative methods to determine earthquake damage grade by using a RS seismic damage index (RSDI) which describes the seismic damage level in an area scale of street block or village. RSDI differs from the seismic damage index determined through ground survey (GDI). The comparison between RSDI and GDI will suggest difference in damage identification. In this paper, the estimation of building damage grade is introduced based on the quantitatively extraction of building damage grade from various RS images acquired after 2010 Ms7.1 Yushu earthquake, Qinghai, China. Then a series of quantitative models between the RSDI and GDI for each source of RS image are obtained. The capability to identify building damage from RS image with different resolution will be analyzed in the paper.

The SAR image can be acquired without the affection of cloud or rain and the limitation of sunlight, so it is widely used in many fields, especially in the disaster management. The paper is concentrated on the research of the method to extract disaster information quickly from SAR images, which is important and useful for emergency remote sensing processing. The fractal analyzing methods are described and applied to extract the building damage of 2010 Yushu Ms 7.1 earthquakes from SAR images. The results are compared with the interpretation of building damage from the high resolution airborne and satellite optical images.