Year: 2024
Type: conference-paper
Abstract: High resolution remote sensing images contain rich ground information, and identifying key categories of objects in the images through algorithms is of great significance in both military and civilian fields. With the rapid development of artificial intelligence and deep learning technology, the use of semantic segmentation technology for pixel level segmentation of surface objects has been widely applied, with the characteristics of clear object contours and good visualization effects. This article proposes a lightweight semantic segmentation network structure based on convolutional neural networks to address the problems of high computational complexity and poor real-time performance in current semantic segmentation networks. Experimental results on the open-source high-resolution remote sensing image dataset Potsdam show that it can significantly reduce computational complexity while maintaining high segmentation accuracy, and has the foundation for application on low-power devices such as embedded systems.
Authors: Jiayin Teng, Tong Zhang, Yuhao Wang
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W4405088112