Feature attention gated context aggregation network for single image dehazing and its application on unmanned aerial vehicle images

• Published in: IET cyber-physical systems 2024-09, Vol.9 (3), p.218-227
• Main Authors: Wu, Yongquan, Zhao, Xuan, Zhang, Xinsheng, Long, Tao, Luo, Ping
• Format: Article
• Language: English
• Subjects: computer vision; learning (artificial intelligence)
• ISSN: 2398-3396; 2398-3396
• DOI: 10.1049/cps2.12076

Single‐image dehazing is a highly challenging ill‐posed task in the field of computer vision. To address this, a new image dehazing model with feature attention, named feature attention gated context aggregation network (FAGCA‐Net), is proposed to tackle the issues of incomplete or over‐dehazing caused by the original model's inability to handle non‐uniform haze density distributions. A feature attention module that combines channel attention and spatial attention is introduced. Additionally, the authors propose a new extended attention convolutional block, which not only addresses the grid artefacts caused by the extended convolution but also provides added flexibility in handling different types of feature information. At the same time, in addition to the input image itself, incorporating the dark channel and edge channel of the image as the final input of the model is helpful for the model learning process. To demonstrate the robustness of the new model, it is applied to two completely different dehazing datasets, and it achieves significant dehazing performance improvement over the original model. Finally, to verify the effectiveness of the model in practical production processes, the authors apply it as an image preprocessing step to a set of UAV (Unmanned Aerial Vehicle) images of foreign objects. The result shows that the UAV images after being processed by FAGCA‐Net for haze removal have a better impact on subsequent usage. To address the issue of incomplete or over‐dehazing caused by previous models inability to handle non‐uniform hazy images, a new dehazing model named Feature Attention Gated Context Aggregation Network (FAGCA‐Net) is presented. A new dilated attention convolution block is proposed, which not only deals with the grid artefacts caused by dilated convolution but also has additional flexibility in handling different types of feature information. To demonstrate the model’s robustness, it is applied to two completely different dehazing datasets and achieves significant performance improvement over previous models. Furthermore, to verify the effectiveness of the model in practical production processes, it is applied as an image preprocessing step to a set of Unmanned Aerial Vehicle (UAV) images. The results show that the UAV images processed by FAGCA‐Net for haze removal are more suitable for subsequent usage.