The invention discloses a quantitative analysis method of a visibility prediction model based on improved VGG16. The quantitative analysis method comprises the following steps: image preprocessing: extracting a key frame image from an original video; constructing a visibility prediction model based on the improved VGG16; carrying out network training and evaluation analysis on the visibility prediction model based on the improved VGG16; and importing the processed image into a visibility prediction model based on the improved VGG16 for prediction, and outputting a visibility analysis result. The invention discloses a quantitative analysis method of a visibility prediction model based on improved VGG16, which achieves universality, high safety and high reliability, and provides powerful support and effective guarantee for application in the fields of actual traffic management and safety.
本发明公开了一种基于改进VGG16的能见度预测模型的量化分析方法,包括:图像预处理:对原始视频提取关键帧图像;构建基于改进VGG16的能见度预测模型;对基于改进VGG16的能见度预测模型进行网络训练和评估分析;将处理后的图像导入基于改进VGG16的能见度预测模型进行预测,输出能见度分析结果。本发明公开了一种基于改进VGG16的能见度预测模型的量化分析方法,达到普适性、高安全性和高可靠性,为在实际交通管理和安全领域的应用提供有力支持和有效保障。
Quantitative analysis method of visibility prediction model based on improved VGG16
基于改进VGG16的能见度预测模型的量化分析方法
2024-08-30
Patent
Electronic Resource
Chinese
A Hybrid YOLO-VGG16 for Fallen Motorcycle Detection
IEEE | 2025
|Classification and Detection Method of Blood Lancet based on VGG16 Network
British Library Conference Proceedings | 2021
|