This paper proposes a traffic sign detection method based on YOLOv8, named TSD-YOLO, to address the challenges of multi-scale variations and complex backgrounds in traffic sign detection. The method introduces the Space-to-Depth (SPD) module to compress spatial information into depth channels, expanding the receptive field and enhancing detection capabilities for objects of varying sizes. Additionally, the Select Kernel (SK) attention mechanism is employed to dynamically adjust the model's focus and more effectively concentrate on key features. The WIoUv3 loss function is also adopted to optimize loss calculation through a weighted approach, improving the model's detection performance across various sizes and frequencies of instances. The proposed methods were validated on the CCTSDB and TT100K datasets, achieving substantial improvements of 3.2% and 5.1% on the mAP50 metric compared to YOLOv8s, while maintaining high detection speed. The code for this paper is available at https://github.com/dusongjie/TSD-YOLO-Small-Traffic-Sign-Detection-Based-on-Improved-YOLO-v8. The study demonstrates that the proposed method effectively enhances the detection capabilities of the model, particularly for small objects, while maintaining real-time performance. The results show that the method outperforms other mainstream single-stage detection models, including YOLOv5, YOLOv7, and SSD, in terms of detection accuracy and performance. The method also shows promising results in handling complex backgrounds and multi-scale variations in traffic sign detection. The study concludes that the proposed method is effective in enhancing the detection capabilities of the model, particularly for small objects, while maintaining real-time performance.This paper proposes a traffic sign detection method based on YOLOv8, named TSD-YOLO, to address the challenges of multi-scale variations and complex backgrounds in traffic sign detection. The method introduces the Space-to-Depth (SPD) module to compress spatial information into depth channels, expanding the receptive field and enhancing detection capabilities for objects of varying sizes. Additionally, the Select Kernel (SK) attention mechanism is employed to dynamically adjust the model's focus and more effectively concentrate on key features. The WIoUv3 loss function is also adopted to optimize loss calculation through a weighted approach, improving the model's detection performance across various sizes and frequencies of instances. The proposed methods were validated on the CCTSDB and TT100K datasets, achieving substantial improvements of 3.2% and 5.1% on the mAP50 metric compared to YOLOv8s, while maintaining high detection speed. The code for this paper is available at https://github.com/dusongjie/TSD-YOLO-Small-Traffic-Sign-Detection-Based-on-Improved-YOLO-v8. The study demonstrates that the proposed method effectively enhances the detection capabilities of the model, particularly for small objects, while maintaining real-time performance. The results show that the method outperforms other mainstream single-stage detection models, including YOLOv5, YOLOv7, and SSD, in terms of detection accuracy and performance. The method also shows promising results in handling complex backgrounds and multi-scale variations in traffic sign detection. The study concludes that the proposed method is effective in enhancing the detection capabilities of the model, particularly for small objects, while maintaining real-time performance.