Infrared Aircraft Detection Algorithm Using Semantic Segmentation
Author Information
Author(s): Liu Gang, Xi Jiangtao, Ma Chao, Chen Huixiang, Yi Wei
Primary Institution: Henan University of Science and Technology
Hypothesis
Can a high-resolution feature-enhanced semantic segmentation network effectively detect infrared aircraft under interference conditions?
Conclusion
The proposed algorithm achieves effective detection of infrared aircraft in the presence of interference, outperforming classic segmentation algorithms.
Supporting Evidence
- The proposed algorithm achieved a mean intersection over union (mIoU) of 92.74%.
- It outperformed classic segmentation algorithms such as DeepLabv3+, Segformer, HRNetv2, and DDRNet.
- The mean pixel accuracy (mPA) was 96.34%, and the mean recall (MR) was 96.19%.
Takeaway
This study created a smart computer program that helps find airplanes in thermal images, even when there are distractions like clouds or other objects.
Methodology
The study used a high-resolution feature-enhanced semantic segmentation network that combines location attention feature fusion, hybrid atrous spatial pyramid pooling, and a dice loss function.
Limitations
The algorithm may require significant computational resources and storage space.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website