Abstract Considering the premise of animal welfare, floor‐rearing duck houses have become widely adopted for raising laying ducks. However, due to the limitation in the number of nesting boxes, laying ducks may oviposit in other areas of the house. According to field statistics, approximately 25% of the eggs require manual collection. To improve the efficiency of duck farming work, this study has developed an intelligent automated egg collection robot that integrates an autonomous mobile robot (AMR) with an egg collection and pushing mechanism. The AMR combines autonomous navigation technology with a coverage path planning (CPP) algorithm to ensure that the robot can perform a thorough, dead‐zone‐free inspection of the entire duck house. In addition, we used the YOLOv8 deep learning‐based image recognition method to successfully train a ground egg detection model, with optimal detection performance achieved at a distance of one meter. Subsequently, the vision‐based servo control technique developed uses image feedback to guide the AMR's movements, while the innovatively designed egg collection and pushing mechanism completes the retrieval of ground eggs and transports them to the central collection area. In addition, based on navigation cost map analysis, the system determines whether detected eggs are near obstacle edges and uses these data to identify the most optimal collection position for edge eggs, allowing the AMR to successfully pick up edge eggs while avoiding collisions. Ultimately, experimental validation was performed on a test field , where the robot inspected, collected, and deposited a total of 180 ground‐laid eggs, successfully collecting 172 eggs, corresponding to a success rate of 95.6%, thus demonstrating the effectiveness of the proposed system. In the future, the automated egg collection robot will be deployed in actual duck farms, with the aim of meeting animal welfare needs, reducing manual labor, and enhancing production efficiency through the introduction of intelligent automation in poultry egg production environments.
Jean et al. (Tue,) studied this question.