Abstract
Many objects in the naturalenvironment are generated from the background and even transformed by nature or human beings. Thus, they do not have closed and well-defined boundaries in remote sensing imagery. Recently, convolutional neural network (CNN) based object detection achieved great success in the remote sensing field. However, there is no investigation in the literature about the detection of objects with ambiguous boundaries. In this article, taking the case of the potential loess landslide detection, we designed massive experiments to evaluate convolutional neural networks for detecting objects with ambiguous boundaries in remote sensing imagery. We analyzed the evaluated methods comprehensively by comparing the performance on objects with ambiguous boundaries in remote sensing imagery with the performance on ordinary objects in visual imagery. Furthermore, drawing from these analyses, we provided a fundamental principle of object representation and a meaningful suggestion of information learning to detect objects with ambiguous boundaries. We finished this article by presenting several promising directions for detecting objects with ambiguous boundaries to facilitate and spur future research. This article would provide a significant reference and guidance to develop detectors for objects with ambiguous boundaries in remote sensing imagery.
Highlights
I N the past few decades, remote sensing observations have accomplished tremendous advances, and a large number of remote sensing images are collected each year
Convolutional neural network and attention mechanism were widely employed in object detection in visual imagery, with numerous CNN-based object detectors [5], [6], [7] and attention modules [8], [9] emerging recently
A variety of CNN-based object detection methods and attention mechanism modules were proposed in the computer vision community and obtained tremendous success in both visual imagery and remote sensing imagery
Summary
I N the past few decades, remote sensing observations have accomplished tremendous advances, and a large number of remote sensing images are collected each year. Convolutional neural network and attention mechanism were widely employed in object detection in visual imagery, with numerous CNN-based object detectors [5], [6], [7] and attention modules [8], [9] emerging recently. In recent years, they were extended to object detection in remote sensing imagery. Zhou et al [15] proposed a local attention network and made the network focus on individual parts of objects to improve the detection of occluded objects in remote sensing images.
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
