在基于单幅红外图像自动识别并重构标准几何体的方法基础上,利用正侧两幅红外图像对带有遮挡的标准几何体的组合进行自动识别和分类,提取图像的清晰边缘后通过提取遮挡几何体未被遮挡的对称区域来提取被遮挡几何体几何特征;利用空间直角变换获得尺寸参数,最后应用三维重构模型实现多目标有遮挡的红外图像的几何体系统的三维模型重构,为复杂的红外目标的自动重构打下基础。
A new method based on the identification and 3D reconstruction of single infrared image is proposed, using two infrared images which are taken in two different angle to identify and classify the concealed standard geometrical solids system automatically. The edge features based on the temperatures of solids are extracted. Then, the key feature of the concealed standard geometrical solid are obtained by the symmetry area extraction, The 3D standard geometrical solids system models are reconstructed by used the key parameters of solids, which are obtained by orthogonal coordinates transform. The 3D reconstruction of the concealed standard geometrical solids system is the base of reconstruction of complicated solids systems.