微型电脑应用

由于采集的室内规则物体信息不完整，导致三维建模精度低，无法真实地反映室内规则物体的形态。为此，提出倾斜摄影测量下的室内规则物体精细三维建模。规划多相机系统的拍摄位置，采用倾斜摄影测量技术采集室内规则物体的精细信息。设计组件感知的三维物体重建网络，将室内规则物体精细信息输入网络中，采用组件位置预测模块获取组件在室内规则物体中的位置，利用特征提取模块获取室内规则物体的联合特征，在重建模块中完成室内规则物体的三维建模。采用纹理映射技术实时渲染物体三维模型，实现物体的精细三维建模。实验结果表明，所提方法的三维建模平面误差以及高程误差较低，能够有效提高建模精度，加快建模速度。

Due to incomplete information collected on indoor regular objects, the accuracy of 3D modeling is low and it cannot truly reflect the shape of indoor regular objects. Therefore, a precise 3D modeling of indoor regular objects under oblique photogrammetry is proposed. By planning the shooting position of a multi-camera system the fine information of indoor regular objects is collected by oblique photo grammetry technology. A component aware 3D object reconstruction network is designed, fine information of indoor rule objects is inputted into the network, a component position prediction module is used to obtain the position of components in indoor rule objects, a feature extraction module is used to obtain joint features of indoor rule objects, thus, the 3D modeling of indoor rule objects in the reconstruction module is completed. The texture mapping technology is used to render the 3D model of the object in real time to achieve the fine 3D modeling of the object. The experimental results show that the 3D modeling plane error and elevation error of the proposed method are low, which can effectively improve the modeling accuracy and accelerate modeling speed.