Research on Intelligent Mechanized Mushroom Harvesting Device Based on AI Vision and Multi-Sensor Fusion

Large-scale mushroom production faces persistent challenges including low manual harvesting efficiency, high damage rates, and rising labor costs, creating an urgent demand for intelligent agricultural equipment. To address these issues, this study develops an intelligent mechanized mushroom harvesting device based on AI vision and multi-sensor fusion, forming a closed-loop “perception–decision– execution” control system with a modular architecture. Field experiments were conducted on oyster mushrooms, shiitake mushrooms, and enoki mushrooms in a 500 m2 greenhouse, with manual harvesting used as the comparison. The results show that the device achieves a harvesting efficiency of 38.6– 45.3 kg/h, which is 3–4 times higher than manual harvesting, with an average damage rate of no more than 4.2% and recognition accuracy of at least 93.6% under low-light, high-humidity, and dense-growth conditions. The modular design supports rapid adaptation to different mushroom varieties, while the IoT platform enables real-time monitoring and remote scheduling. This device provides technical support for intelligent agricultural production and offers engineering references for machine vision, wireless sensing, and multi-sensor information fusion.