Abstract
With the development of effective deep learning algorithms, it became possible to achieve high accuracy when conducting remote sensing analyses on very high-resolution images (VHRS), especially in the context of building detection and classification. In this article, in order to improve the accuracy of building detection and classification, we propose a Faster Edge Region Convolutional Neural Networks (FER-CNN) algorithm. This proposed algorithm is trained and evaluated on different datasets. In addition, we propose a new method to improve the detection of the boundaries of detected buildings. The results of our algorithm are compared with those of other methods, such as classical Faster Region Convolution Neural Network (Faster R-CNN) with the original VGG16 and the Single-Shot Multibox Detector (SSD). The experimental results show that our methods make it possible to obtain an average detection accuracy of 97.5% with a false positive classification rate of 8.4%. An additional advantage of our method is better resistance to shadows, which is a very common issue for satellite images of urban areas. Future research will include designing and training the neural network to detect small buildings, as well as irregularly shaped buildings that are partially obscured by shadows or other occlusions.
Highlights
High-resolution remote sensing satellite imagery can provide the geometric features, spatial features and textures of many objects, including various types of buildings
Comparing the results obtained by using Faster R-CNN, Faster Edge Region Convolution Neural Network (FER-CNN) and Single-Shot Multibox Detector (SSD), it can be concluded that (1) models based on the Adam algorithm achieved good results only for Faster R-CNN, while they generated errors when used in SSD networks
The capabilities of neural networks in the detection and classification of buildings located in satellite images were examined
Summary
High-resolution remote sensing satellite imagery can provide the geometric features, spatial features and textures of many objects, including various types of buildings. The dynamic technological development of satellite systems has made it possible to acquire images with better spatial resolution, which has led to the possibility of extracting more details of objects contained in the images, i.e., easier and more effective detection of objects in the image Because of their range and temporal resolution, they provide large amounts of information in a short time, and they are playing an increasingly important role in updating, controlling and analyzing the spatial development of many studied areas [1,2,3]. Methods of extracting object features can be distinguished according to whether they are based on data or based on a model [6] The former relies on mathematical operations applied to a given image, without prior knowledge of what it may contain. Neither of these methods provides an unambiguous answer about the location of buildings that are in the image, let alone their purpose
Sign-in/Register to view highlights & summary 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 callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
