A platform for research: civil engineering, architecture and urbanism
Roadbed compaction quality evaluation method and equipment based on multi-mode diplopia
The invention relates to a roadbed compaction quality evaluation method and device based on multi-mode diplopia, a camera group is installed to continuously obtain roadbed compaction images, and the camera group comprises an infrared camera and at least two RGB cameras; the infrared camera is used for acquiring a clearly shot infrared image of the roadbed, one RGB camera is used for acquiring a vertically shot low-position mesoscopic image of the roadbed, and the other RGB cameras are used for acquiring a squint shot high-position macroscopic image of the roadbed; performing frequency domain conversion on the acquired infrared image to obtain an infrared frequency domain image; forming an image data set by the low-order mesoscopic image, the high-order macroscopic image, the infrared image and the infrared frequency domain image, and forming a parameter data set by the mechanical parameters of the road roller and the filling parameters of the roadbed; and using the image data set, the parameter data set and the corresponding roadbed compaction degree to train a multi-modal double-view roadbed compaction degree prediction network for roadbed compaction quality evaluation. The method improves the accuracy and robustness of the model, and can be applied to different construction scenes.
本发明为一种基于多模态复视的路基压实质量评价方法及装备,安装相机群连续获取路基压实图像,所述相机群包括红外相机和至少两个RGB相机;红外相机用于获取清晰拍摄的路基的红外图像,一个RGB相机用于获取垂直拍摄的路基的低位细观图像,其余RGB相机用于获取斜视拍摄的路基的高位宏观图像;对采集的红外图像进行频域转换,得到红外频域图像;将低位细观图像、高位宏观图像、红外图像和红外频域图像构成图像数据集,压路机机械参数和路基填土参数构成参数数据集;利用图像数据集和参数数据集及对应的路基压实度训练多模态复视路基压实度预测网络,用于路基压实质量评价。提高模型准确性和鲁棒性,可应用于不同施工场景。
Roadbed compaction quality evaluation method and equipment based on multi-mode diplopia
The invention relates to a roadbed compaction quality evaluation method and device based on multi-mode diplopia, a camera group is installed to continuously obtain roadbed compaction images, and the camera group comprises an infrared camera and at least two RGB cameras; the infrared camera is used for acquiring a clearly shot infrared image of the roadbed, one RGB camera is used for acquiring a vertically shot low-position mesoscopic image of the roadbed, and the other RGB cameras are used for acquiring a squint shot high-position macroscopic image of the roadbed; performing frequency domain conversion on the acquired infrared image to obtain an infrared frequency domain image; forming an image data set by the low-order mesoscopic image, the high-order macroscopic image, the infrared image and the infrared frequency domain image, and forming a parameter data set by the mechanical parameters of the road roller and the filling parameters of the roadbed; and using the image data set, the parameter data set and the corresponding roadbed compaction degree to train a multi-modal double-view roadbed compaction degree prediction network for roadbed compaction quality evaluation. The method improves the accuracy and robustness of the model, and can be applied to different construction scenes.
本发明为一种基于多模态复视的路基压实质量评价方法及装备,安装相机群连续获取路基压实图像,所述相机群包括红外相机和至少两个RGB相机;红外相机用于获取清晰拍摄的路基的红外图像,一个RGB相机用于获取垂直拍摄的路基的低位细观图像,其余RGB相机用于获取斜视拍摄的路基的高位宏观图像;对采集的红外图像进行频域转换,得到红外频域图像;将低位细观图像、高位宏观图像、红外图像和红外频域图像构成图像数据集,压路机机械参数和路基填土参数构成参数数据集;利用图像数据集和参数数据集及对应的路基压实度训练多模态复视路基压实度预测网络,用于路基压实质量评价。提高模型准确性和鲁棒性,可应用于不同施工场景。
Roadbed compaction quality evaluation method and equipment based on multi-mode diplopia
一种基于多模态复视的路基压实质量评价方法及装备
WANG XUEFEI (author) / LI JIALE (author) / WANG TINGKAI (author)
2023-08-29
Patent
Electronic Resource
Chinese
Roadbed dynamic compaction reinforcement effect evaluation method
European Patent Office | 2015
|Compaction equipment for collapsible loess roadbed construction
European Patent Office | 2023
|