摘要
针对数字图像处理课程中存在理解困难、理论模糊不清和应用实践匮乏等问题,基于智能采摘竞赛进行设计标识物检测与识别教学案例,旨在增强学生的学习兴趣并提升教学效果。利用摄像头拍摄照片并存储图像,结合Jetson nano平台和YOLOv5S多目标检测框架,训练智能采摘标识物检测模型。基于已训练模型预测目标,并输出预测结果与物体坐标,从而获得舵机角度并执行抓取功能。实验结果显示,平均检测精度达99.3%,在Jetson nano平台上展现出优异的性能,未来有望借助更多模型检测实现更精准的检测和广泛应用。实验过程中,通过目标检测与抓取效果向学生形象地展示图像目标检测的应用,加深其对数字图像处理的理解,增强学生进行深度学习与图像智能处理的能力。
Aiming at the problems of understanding difficulty,theory ambiguity and lack of application practice that often exist in digital image processing courses,a teaching case of designing markers detection and recognition is carried out based on the smart picking competition to enhance students'learning interest and improve the effect of this teaching.In the experiment,camera is used to take photos,images are stored and used to train the intelligent picking marker detection model by combining the Jetson nano platform and the YOLOv5S multi-target detection framework.The target is predicted based on the trained model,and the prediction results and object coordinates are output to obtain the helm angle and perform the grasping function.Eventually,the average detection accuracy of the experimental results is as high as 99.3%,which shows excellent performance on the Jetson nano platform,and is expected to realize more accurate detection and wide application with the help of more model detection in the future.During the experiment,the target detection and grasping results show students the application of image target detection in a visual way,deepen their understanding of digital image processing,and enhance their ability to carry out deep learning and image intelligent processing.
作者
翁士状
李士昌
李迎松
WENG Shizhuang;LI Shichang;LI Yingsong(School of Electronic and Information Engineering,Anhui University,Hefei 230601,China)
出处
《实验室研究与探索》
北大核心
2025年第8期48-54,共7页
Research and Exploration In Laboratory
基金
安徽省级质量工程(2023jyxm0131,2023cxtd019)
安徽省新时代育人质量工程项目(2022zyxwjxalk021,2023zyxwjxalk012,2023shsjsfkc004)。
