Wuhan University

China Three Gorges Construction Engineering Corporation

Changjiang Spatial Information Technology Engineering Co., Ltd. (Wuhan)

Zheng Wang

Chao Zhang

Bo Shi

Xiaoyang Guo

Yutian Lin

Hailong Huang

Shuangping Li

Yuanzhu Chen

Zuqiang Liu

Huawei Wang

Bin Zhang

Zhao Tang

Yiming Chen

Yonghua Li

The present disclosure provides a method and system for establishing a space-ground integrated real-time monitoring reference for dam deformation. The method includes: transmitting control instructions to a space-based Beidou system, a ground-based measurement robot system and a ground-based inverted plumb line system; obtaining a dynamic reference point position of a dam deformation area in real time through the space-based Beidou system; obtaining a spatial relative reference point position of the dam deformation area in real time through the ground-based measurement robot system; obtaining a static gravity reference point position of the dam deformation area in real time through the ground-based inverted plumb line system; and performing data fusion on the dynamic reference point position, the spatial relative reference point position, and the static gravity reference point position, and obtaining a monitoring reference value of the dam deformation area.

Method and system for establishing space-ground integrated real-time monitoring reference for dam deformation

This invention proposes a pedestrian re-identification method based on virtual samples, comprising following steps: s) obtaining virtual persons generated by game engine, and generating the virtual samples with person labels by fusing a background of a target dataset and a pose of real persons through a multi-factor variational generation network; s) rendering the generated virtual samples according to lighting conditions; s) sampling the rendered virtual samples according to person attributes of target dataset; s) constructing a training dataset according to virtual samples obtained by sampling to train a pedestrian re-identification model, and verifying identification effect of the trained model. The present invention uses a virtual image generation framework that integrates translation-rendering-sampling to narrow the distribution between virtual images and real images as much as possible to generate virtual samples, and conduct person re-identification model training, which can be effectively and effectively applied to pedestrian datasets in real scenes.

Pedestrian re-identification method based on virtual samples

Growing community of inventors

Hubei, China

Zheng Wang