Refining a rational functional model (RFM) by subdividing the multiple images into multiple groups of sub-images according to the imaging sensor, then performing the RFM refinement for each group separately and estimating a 3D transformation using the sub-images from the central sensor as a reference.

Image acquisition and analysis systems for efficiently generating high resolution geo-referenced spectral imagery of a region of interest. In some examples, aerial spectral imaging systems for remote sensing of a geographic region, such as a vegetative landscape are disclosed for monitoring the development and health of the vegetative landscape. In some examples photogrammetry processes are applied to a first set of image frames captured with a first image sensor having a first field of view to generate external orientation data and surface elevation data and the generated external orientation data is translated into external orientation data for other image sensors co-located on the same apparatus for generating geo-referenced images of images captured by the one or more other image sensors.

The present disclosure relates to a method for generating a relative depth map image. The method comprises feeding at least two input images to a neural network, the input images relating to a region of interest at different time periods, the input images being obtained at corresponding arbitrary positions and attitudes with respect to the region of interest, and predicting, using the neural network, the relative depth map image.

A pan-sharpening method based on multimodal texture correction and adaptive edge detail fusion is provided, including: fusing upsampled low-resolution multispectral (LRMS) images with panchromatic images to obtain fused images; respectively extracting intensity components of the LRMS image and the fused image; inputting the intensity components and the panchromatic images into a multimodal texture correction model, and performing optimization solution on the multimodal texture correction model through optimization method to obtain texture-corrected images; extracting details of the texture-corrected images and applying edge protection to obtain first image details; extracting details of the upsampled LRMS image and applying edge protection to obtain second image details; performing adaptive fusion on the first image details and the second image details to obtain detail information; and adding the detail information to the upsampled LRMS image to obtain final high-resolution multispectral (HRMS) images.