基于YOLOv8和PSP-Ellipse的火龙果成熟度识别
Ripeness identification of pitaya fruit based on YOLOv8 and PSP-Ellipse
刘昕璞 1赵春雷 2李志锋 3冯超3
作者信息
- 1. 秦皇岛工业职业技术学院,河北 秦皇岛 066000
- 2. 华北理工大学,河北 唐山 063210
- 3. 河北农业大学,河北 保定 071001
- 折叠
摘要
[目的]提高火龙果成熟度检测准确率及鲁棒性.[方法]采用YOLOv8目标检测模型与PSP-Ellipse分割算法相结合的策略,提出一种高效且准确的火龙果成熟度自动识别方法.先利用YOLOv8的实时目标检测功能对火龙果进行初步定位和识别,再通过PSP-Ellipse的形状识别能力,对火龙果的形状和成熟度进行进一步的精细分类.[结果]火龙果成熟度分类准确率为97.6%,鲁棒性较强.[结论]该方法在复杂背景和多种光照条件下能够显著提高火龙果的自动化分级效率.
Abstract
[Objective]Improve the accuracy and robustness of maturity detection of pitaya fruit.[Methods]Combining the YOLOv8 object detection model with the PSP-Ellipse segmentation algorithm,an efficient and accurate automatic identification method for pitaya fruit maturity was proposed.First,the real-time target detection function of YOLOv8 was used to locate and identify the pitaya fruit initially.Then the shape recognition capability of PSP-Ellipse was used to further fine classify the shape and maturity of the pitaya fruit.[Results]The accuracy rate of maturity classification of pitaya fruit was 97.6%,and the robustness was strong.[Conclusion]This method can significantly improve the automatic classification efficiency of pitaya fruit under complex backgrounds and various lighting conditions.
关键词
火龙果/成熟度识别/YOLOv8/PSP-Ellipse/目标检测/形状识别Key words
pitaya fruit/maturity identification/YOLOv8/PSP-Ellipse/target detection/shape recognition引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据