基于深度学习的航空发动机损伤识别实验设计  

Experimental Design for Aircraft Engine Damage Recognition Based on Deep Learning

作　　者：吕巨建 曾运标 林凯瀚 黎嘉文 陈艳美 陈荣军 LÜJujian;ZENG Yunbiao;LIN Kaihan;LI Jiawen;CHEN Yanmei;CHEN Rongjun(School of Computer Science,Guangdong Polytechnic Normal University,Guangzhou 510665,China;School of Computer Science and Technology,Guangdong University of Technology,Guangzhou 510006,China)

机构地区：[1]广东技术师范大学计算机科学学院,广州510665 [2]广东工业大学计算机学院,广州510006

出　　处：《实验室研究与探索》2025年第4期97-102,共6页Research and Exploration In Laboratory

基　　金：国家自然科学基金青年基金项目(11801097);教育部产学合作协同育人项目(241003632084003,241005211144020);2023年广东省高等职业教育教学质量与教学改革工程项目(2023JG296);广东技术师范大学2022年博士点建设单位科研能力提升项目(22GPNUZDJS31)。

摘　　要：针对高校人工智能课程教学中缺乏实际应用案例的问题,设计并实现一种基于深度学习的航空发动机损伤识别实验。实验采用YOLOv9-c作为基线模型提取损伤特征,在主干网络中集成空间通道注意力模块,以增强模型对关键特征的识别能力。通过Python编程语言和PyTorch深度学习框架,开发并实现一个基于改进YOLOv9模型的航空发动机损伤检测系统。实验过程中,学生可通过基线模型和改进模型的识别对比实验,验证改进模型的有效性,还可自行编程设计完成改进YOLO模型的创新性实验。实验为实践导向的教学提供深度学习在实际工程中的应用案例,有助于培养学生的实践技能和创新思维。To address the issue of a lack of comprehensive real-case demonstrations in artificial intelligence course in universities,an experiment of aero-engine damage identification based on deep learning is designed.This experiment employs the YOLOv9-c as the baseline model for feature extraction of engine damages,and incorporates a spatial channel attention module into the backbone network to enhance the model’s ability to recognize key features.It utilizes Python to construct a detection system for aero-engine damage within the Pytorch framework,based on the enhanced YOLOv9 model.During the experimental process,students are not only able to conduct recognition comparison experiments between the pre-trained YOLOv9 and the enhanced model to verify the effectiveness of the enhancement,but also have the opportunity to program and refine the YOLO model for innovative experiments.This pratical-oriented experiment provides students with a practical engineering application case of deep learning,and helps to cultivate their practical skills and innovative thinking.

关 键 词：深度学习 航空发动机 损伤识别 空间通道注意力 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TP391.4[自动化与计算机技术—控制科学与工程] V241.07[航空宇航科学与技术—飞行器设计] V231.95