Abstract
Abstract. Deep learning and convolutional neural networks (CNN) have obtained a great success in image processing, by means of its powerful feature extraction ability to learn specific tasks. Many deep learning based algorithms have been developed for dense image matching, which is a hot topic in the community of computer vision. These methods are tested for close-range or street-view stereo data, however, not well studied with remote sensing datasets, including aerial and satellite data. As more high-quality datasets are collected by recent airborne and spaceborne sensors, it is necessary to compare the performance of these algorithms to classical dense matching algorithms on remote sensing data. In this paper, Guided Aggregation Net (GA-Net), which belongs to the most competitive algorithms on KITTI 2015 benchmark (street-view dataset), is tested and compared with Semi-Global Matching (SGM) on satellite and airborne data. GA-Net is an end-to-end neural network, which starts from an stereo pair and directly outputs a disparity map indicating the scene’s depth information. It is based on a differentiable approximation of SGM embedded into a neural network, performing well for ill-posed regions, such as textureless areas, slanted surfaces, etc. The results demonstrate that GA-Net is capable of producing a smoother disparity map with less errors, particularly for across track data acquired at different dates.
Highlights
Dense image matching is a key topic in the community of computer vision for stereo reconstruction
In case of dense matching, many algorithms based on deep learning have been developed, which achieve state-of-the-art, such as matching cost based on convolutional neural networks (CNN) (MC-CNN), Guided Aggregation Net (GA-Net), Atrous Multiscale Network (AM-Net), etc. (Du et al, 2019) (Zbontar, LeCun, 2016) (Zhang et al, 2019)
The disparity maps obtained by the two algorithms are displayed in Figure 2
Summary
Dense image matching is a key topic in the community of computer vision for stereo reconstruction. Machine learning, deep learning and convolutional neural networks (CNN) (LeCun et al, 1998) are showing great success in the filed of image processing. With appropriate training datasets available, CNN based algorithms are capable of learning specific tasks to achieve very competitive results, by means of the powerful feature extraction ability. In case of dense matching, many algorithms based on deep learning have been developed, which achieve state-of-the-art, such as matching cost based on CNN (MC-CNN), Guided Aggregation Net (GA-Net), Atrous Multiscale Network (AM-Net), etc. As more high-quality datasets are available thanks to the recent airborne and spaceborne sensors, it is necessary to compare the wellperformed CNN based algorithms to classical dense matching methods on remote sensing data
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
