The accurate identification of crop aphids is an important aspect of improving agricultural productivity. Aphids are characterised by small targets and a body colour similar to their surroundings. Even the most advanced detectors can experience problems such as low detection accuracy and a high number of missed detections. In this paper, a multi-stream target detection model is proposed for fast and accurate detection of crop aphids in complex backgrounds. First, inspired by the human visual system, we propose a bionic attention (BA) approach. Unlike previous strategies, we do not improve the model but input additional category labels as bionic information streams into the network at the network input stage to support mainstream recognition, which improves the detection effect. In addition, through web crawling and manual screening, we construct an aphid dataset containing 2059 images, named IP_Aphids, based on IP102. In this paper, we combine the proposed BA with a number of classical target detection models, including YOLOv5s, YOLOv7-tiny, YOLOv8n, SSD, and faster R-CNN, and we also compare the BA with the classical attention mechanisms CBAM and SE. This approach allows the effectiveness of the method to be evaluated from multiple perspectives. The results show that combining the bionic information flow of BA with the original mainstream information flow as inputs to the five target detection models improves the mAP by 2.2%, 7%, 2.7%, 1.78%, and 1.12%, respectively, compared to the models using only the original inputs. In addition, the mAP of the YOLOv8n_BA model is higher than that of the YOLOv8n, YOLOv8n_CBAM, and YOLOv8n_SE models by 4.6%, 3.3%, and 2.7%, respectively. This indicates that the BA proposed in this study is significantly better than the classical attention to improve crop aphid detection, which provides a reference for crop aphid-related research.

The pecking phenomenon in poultry can lead to stress responses, feather loss, and even death. Therefore, precise and timely monitoring of pecking behavior in poultry is crucial for modern precision animal husbandry. Existing deep learning-based object detection models often face challenges such as slow processing speed, large parameter quantities and bulky model sizes when dealing with complex environments characterized by high density and multiple targets. To address these issues, this paper proposes a cherry valley duck feather pecking site detection model based on YOLOv8. Firstly, a novel lightweight multi-scale feature extraction module, NeoMSM-C2f, is integrated into the backbone network to enhance the model's multi-scale feature extraction capability. Secondly, DyHead is employed as the detection head, which adaptively adjusts the detection strategy based on the dynamic variations of input features. Finally, using YOLOv8_nS as the student model and YOLOv8_sT as the teacher model, knowledge distillation is applied to further improve the detection accuracy of YOLOv8_nS, resulting in the YOLOv8-DuckPluck model, which balances both detection speed and accuracy. Experimental results show that the YOLOv8-DuckPluck model significantly outperforms the baseline model YOLOv8_n, with the mAP increasing from 85.51 % to 90.24 %, and the detection speed rising from 66.98 f/s to 76.3 f/s. This demonstrates that the model meets the real-time processing requirements for feather pecking site detection task of Cherry Valley ducks, achieves high detection accuracy, reduces deployment costs and complexity, and exhibits strong practical applicability.