To address the challenge of bird species detection on transmission lines, this paper proposes a detection method based on dual data enhancement and an improved YOLOv8s model. The method aims to improve the accuracy of identifying small- and medium-sized targets in bird detection scenes on transmission lines, while also accounting for the impact of changing weather conditions. To address these issues, a dual data enhancement strategy is introduced. The model’s generalization ability in outdoor environments is enhanced by simulating various weather conditions, including sunny, cloudy, and foggy days, as well as halo effects. Additionally, an improved Mosaic augmentation technique is proposed, which incorporates target density calculation and adaptive scale stitching. Within the improved YOLOv8s architecture, the CBAM attention mechanism is embedded in the Backbone network, and BiFPN replaces the original Neck module to facilitate bidirectional feature extraction and fusion. Experimental results demonstrate that the proposed method achieves high detection accuracy for all bird species, with an average precision rate of 94.2%, a recall rate of 89.7%, and an mAP@50 of 94.2%. The model also maintains high inference speed, demonstrating potential for real-time detection requirements. Ablation and comparative experiments validate the effectiveness of the proposed model, confirming its suitability for edge deployment and its potential as an effective solution for bird species detection and identification on transmission lines.
To address the inherent limitations of single-modality approaches in fine-grained bird species recognition, this paper proposes an adaptive audiovisual fusion method based on prediction confidence. The proposed framework comprises three core components: an image classification branch, an audio classification branch, and a confidence–adaptive fusion module. The image branch employs EfficientNet-B3 to extract fine-grained visual features through compound scaling and squeeze-and-excitation (SE) attention. The audio branch utilizes ResNet-50 to classify Mel spectrograms converted from bird vocalizations, incorporating a dense sampling inference strategy to fully exploit complete audio information. For multimodal integration, a confidence–adaptive fusion strategy is introduced that jointly considers information entropy and probability gap to dynamically assess the reliability of each modality’s prediction, thereby assigning fusion weights at the sample level without any additional trainable parameters. Experiments on the SSW60 multimodal bird recognition dataset show that the image branch achieves a Top-1 accuracy of 91.55%, outperforming ResNet-50 (89.75%) and VGG-16 (83.81%); the audio branch reaches 68.20%, surpassing AST (63.29%) and VGG-16 (53.48%); and the fused model attains 95.30% Top-1 accuracy, a 3.75 percentage-point improvement over the image-only baseline and a 0.21 percentage-point gain over the learning-based TMC fusion baseline without introducing any trainable parameters, confirming the effectiveness of the proposed method.