ALSS-YOLO: An Adaptive Lightweight Channel Split and Shuffling Network for TIR Wildlife Detection in UAV Imagery

Ang He,Xiaobo Li,Ximei Wu,Chengyue Su,Jing Chen,Sheng Xu,Xiaobin Guo

2024-09-12

Abstract:Unmanned aerial vehicles (UAVs) equipped with thermal infrared (TIR) cameras play a crucial role in combating nocturnal wildlife poaching. However, TIR images often face challenges such as jitter, and wildlife overlap, necessitating UAVs to possess the capability to identify blurred and overlapping small targets. Current traditional lightweight networks deployed on UAVs struggle to extract features from blurry small targets. To address this issue, we developed ALSS-YOLO, an efficient and lightweight detector optimized for TIR aerial images. Firstly, we propose a novel Adaptive Lightweight Channel Split and Shuffling (ALSS) module. This module employs an adaptive channel split strategy to optimize feature extraction and integrates a channel shuffling mechanism to enhance information exchange between channels. This improves the extraction of blurry features, crucial for handling jitter-induced blur and overlapping targets. Secondly, we developed a Lightweight Coordinate Attention (LCA) module that employs adaptive pooling and grouped convolution to integrate feature information across dimensions. This module ensures lightweight operation while maintaining high detection precision and robustness against jitter and target overlap. Additionally, we developed a single-channel focus module to aggregate the width and height information of each channel into four-dimensional channel fusion, which improves the feature representation efficiency of infrared images. Finally, we modify the localization loss function to emphasize the loss value associated with small objects to improve localization accuracy. Extensive experiments on the BIRDSAI and ISOD TIR UAV wildlife datasets show that ALSS-YOLO achieves state-of-the-art performance, Our code is openly available at <a class="link-external link-https" href="https://github.com/helloworlder8/computer_vision" rel="external noopener nofollow">this https URL</a>.

Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

The paper aims to address the challenges of wildlife detection in thermal infrared (TIR) drone images, particularly in the context of nighttime anti-poaching applications. The main issues include:

1. **Blurred and overlapping targets**: TIR images captured by drones are often affected by jitter and target overlap, making it difficult to extract features of small targets.
2. **Low resolution and noise**: TIR images typically have lower resolution and higher noise, complicating feature extraction.
3. **Limitations of existing lightweight networks**: Existing lightweight networks perform poorly in handling blurred small targets, especially in overlapping scenes.

To tackle these challenges, the research team proposed ALSS-YOLO, an efficient and lightweight object detector optimized specifically for TIR images. Its main contributions include:
- Proposing the Adaptive Lightweight Channel Split and Reorder module (ALSS) and the Lightweight Coordinate Attention module (LCA) to enhance feature extraction capabilities.
- Designing the FineSIOU loss function, which emphasizes the speed and accuracy of small target bounding box regression.
- Experimental results on the BIRDSAI and ISOD datasets show that ALSS-YOLO achieves higher mAP scores and parameter efficiency compared to other lightweight detectors.