摘要
准确检测成熟度对香菇智能化采摘具有重要意义,因此提出一种基于改进YOLOv5实例分割香菇的成熟度检测方法。该方法在骨干网络的C3模块中添加挤压和激发模块(SENet),增强了对香菇具体特征的学习能力,同时将颈部网络中的2个卷积模块替换为可变形卷积模块(Deformable Convnets v2, DCN v2),使网络更好地适应目标香菇的形状和位置变化,提高成熟度检测的准确率和鲁棒性。实验表明,改进后的模型识别香菇成熟度的检测精度达到91.7%,较原模型提高了6.1%,检测的准确性与可靠性均优于原模型,为香菇智能化种植推广提供了技术支撑。
It is of great significance to accurately detect the maturity of mushroom for the promotion of intelligent cultivation of mushroom. In order to achieve accurate detection of mushroom maturity, a segmentation method for mushroom maturity detection based on improved YOLOv5 example was proposed. This algorithm added Squeeze and Excitation module (SENet) to C3 module in the backbone network to enhance the learning ability of specific characteristics of shiitake mushrooms. The two convolution modules in the neck network were replaced by Deformable Convnets v2 (DCN v2), which made the network better adapt to the shape and position changes of the target, and improved the accuracy and robustness of the target detection. The experiment showed that the accuracy of the improved model was 91.7%, 6.1% higher than that of the original model. The accuracy and reliability of the improved model were better than that of the original model, which provided technical support for the promotion of intelligent planting of mushroom.
作者
李俊成
徐增丙
孙茂基
LI Juncheng;XU Zengbing;SUN Maoji(School of Mechanical Automation,Wuhan University of Science and Technology,Wuhan 430081,Hubei,China;Key Laboratory of Metallurgical Equipment and Its Control,Ministry of Education,Wuhan University of Science and Technology,Wuhan 430081,Hubei,China)
出处
《农业装备与车辆工程》
2024年第6期18-22,共5页
Agricultural Equipment & Vehicle Engineering
基金
国家自然科学基金面上项目“广度自适应的深度迁移度量学习方法及其在复杂机械设备故障诊断与预测中的应用研究”(51775391)。
