Lightweight Low-light Object Detection Algorithm Based on CDD-YOLO
Release time:2025-02-14
Hits:
- DOI number:
- 10.3778/j.issn.1002-8331.2410-0127
- Journal:
- Computer Engineering and Applications
- Key Words:
- low-light; YOLOv8; attention mechanism; loss function; lightweight network
- Abstract:
- To address the challenges of low detection accuracy, high computational costs, and excessive memory consumption encountered by target detection algorithms in low-light conditions, we propose a lightweight low-light target detection network model, CDD-YOLO, designed to enhance the performance of YOLOv8. Firstly, a multi-scale convolutional module based on a coordinate attention mechanism is proposed to extract texture features from different sensory fields and to capture long-range dependencies between spatial locations. Secondly, a dynamic head frame is integrated into the detection head to minimize the interference caused by complex backgrounds and scale variations. The bounding box regression loss function is designed using a dynamic non-monotonic focusing mechanism to enhance the regression path and quality of the anchor boxes, thereby improving the model's adaptability to variations in lighting and noise. Finally, redundant parameters in the model are pruned using a pruning algorithm to achieve model light weighting. Experimental validation using the self-constructed dataset, ExDark, and the VOC dataset. The experimental results show that the method in this paper has better detection effect compared with the mainstream algorithms, and achieves a better balance between computational complexity and detection accuracy.
- Co-author:
- liuxuechao,周星宇
- First Author:
- YANG Chao
- Indexed by:
- Journal paper
- Correspondence Author:
- Lichen Shi
- Discipline:
- Engineering
- Document Type:
- J
- Translation or Not:
- no
- Date of Publication:
- 2024-01-01
- Included Journals:
- EI
- Links to published journals:
- https://link.cnki.net/urlid/11.2127.tp.20241129.09