首页|高光谱结合改进CARS和SSA-XGBoost的鸡蛋品质快速检测方法

高光谱结合改进CARS和SSA-XGBoost的鸡蛋品质快速检测方法

扫码查看
[目的]实现鸡蛋品质的无损、准确和快速检测.[方法]在高光谱检测技术的基础上,提出了一种将樽海鞘群算法与XGBoost算法相结合的鸡蛋品质快速检测方法.通过樽海鞘群算法优化XGBoost模型的多个超参数,提高XG-Boost模型的预测性能.高光谱采集图像通过数据预处理和特征波长选择后输入优化的XGBoost模型进行品质检测.通过试验验证了所提无损检测方法的性能.[结果]试验方法可实现鸡蛋品质的快速无损检测,具有较高的识别精度和效率,决定系数为0.942,平均检测时间为0.032 s.[结论]高光谱检测技术结合试验方法可以实现鸡蛋品质的快速、准确、无损检测.
A rapid detection method for egg quality using CARS and SSA-XGBoost improved by combining hyperspectral analysis
[Objective]To realize non-destructive,accurate,and rapid detection of egg quality.[Methods]On the basis of hyperspectral detection technology,a rapid egg quality detection method combining the bottle sea squirt group algorithm and XGBoost algorithm has been proposed.Optimizing multiple hyperparameters of the XGBoost model through the Tartary Sea Salp Swarm Algorithm to improve the predictive performance of the XGBoost model.The quality of hyperspectral images was detected by inputting optimized XGBoost models after data preprocessing and feature wavelength selection.The performance of the proposed non-destructive testing method was verified through experiments.[Results]The experimental method could achieve rapid non-destructive testing of egg quality,with high recognition accuracy and efficiency,a coefficient of determination of 0.942,and an average detection time was 0.032 seconds.[Conclusion]The combination of hyperspectral detection technology and experimental methods can achieve rapid,accurate,and non-destructive testing of egg quality.

eggqualityhyperspectral detectionsalp swarm algorithmXGBoost algorithmnon destructive testing

王淋铱、邹倩颖、孙强

展开 >

电子科技大学成都学院,四川 成都 611731

吉利学院,四川 成都 641423

成都乐因生物科技有限公司,四川 成都 610043

鸡蛋 品质 高光谱检测 樽海鞘群算法 XGBoost算法 无损检测

四川省自然科学基金项目四川省住房和城乡建设厅项目

22SCK349806SCJSKJ2022-05

2024

食品与机械
长沙理工大学

食品与机械

CSTPCD北大核心
影响因子:0.89
ISSN:1003-5788
年,卷(期):2024.40(8)