公开(公告)号：US20030198378A1

公开(公告)日：2003-10-23

申请号：US10125028

申请日：2002-04-18

Applicant: STMICROELECTRONICS, INC.

Inventor： Kim Chai Ng

IPC: G06K009/00

CPC classification number: G06K9/20 ,  G06K2209/40 ,  G06T7/55

Abstract: An image processing system and method for smoothing irregularities from 3D image information that was reconstructed from a plurality of 2D views of a scene, and particularly from homogeneous surfaces of objects in a scene. The method defines a window that overlaps a plurality of pixels of one of a plurality of 2D image views of a scene. Each pixel is associated with predefined 3D depth information, and further is associated with a matching curve. A subject pixel is located within the plurality of pixels overlapped by the window. The method calculates an average 3D depth information associated with the plurality of pixels overlapped by the window, and assigns the calculated average 3D depth information to the 3D depth information of the subject pixel, if the calculated average 3D depth information is within an error region of a matching curve associated with the subject pixel.

Abstract translation: 一种用于从3D场景的多个2D视图重建的3D图像信息中平滑不规则的图像处理系统和方法，特别是从场景中的物体的均匀表面。 该方法定义了与场景的多个2D图像视图中的一个的多个像素重叠的窗口。 每个像素与预定义的3D深度信息相关联，并且还与匹配曲线相关联。 被摄体像素位于由窗口重叠的多个像素内。 该方法计算与由窗口重叠的多个像素相关联的平均3D深度信息，并且如果所计算的平均3D深度信息在所述目标像素的误差区域内，则将计算出的平均3D深度信息分配给对象像素的3D深度信息 与对象像素相关联的匹配曲线。

公开(公告)号：US20030198377A1

公开(公告)日：2003-10-23

申请号：US10125004

申请日：2002-04-18

Applicant: STMICROELECTRONICS, INC.

Inventor： Kim Chai Ng

IPC: G06K009/00

CPC classification number: G06K9/20 ,  G06K2209/40 ,  G06T7/593

Abstract: An image processing system and method reconstructs 3D image information corresponding to a scene from a plurality of 2D images of the scene. The method receives a plurality of image features corresponded between different 2D views of the scene, the corresponded image features deviating between different views as a result of camera relative motion. The method determines image features of the received plurality of image features that are occluded views, determines image features of the received plurality of image features that are confident seeds associated with 3D depth information, and propagates 3D depth information from the confident seeds to neighboring image features, while avoiding image features that have been determined to be occluded views. The 3D image information can be rendered and displayed such as for a virtual walkthrough of the scene.

Abstract translation: 图像处理系统和方法从场景的多个2D图像重建与场景相对应的3D图像信息。 该方法接收场景的不同2D视图之间对应的多个图像特征，作为相机相对运动的结果，对应的图像特征在不同视图之间偏离。 该方法确定接收到的被遮挡视图的多个图像特征的图像特征，确定接收到的与3D深度信息相关联的确定种子的多个图像特征的图像特征，并将3D深度信息从有信心的种子传播到相邻图像特征 同时避免已被确定为闭塞视图的图像特征。 可以渲染和显示3D图像信息，例如用于场景的虚拟演练。