公开(公告)号：US09465484B1

公开(公告)日：2016-10-11

申请号：US13793276

申请日：2013-03-11

Applicant: Amazon Technologies, Inc.

Inventor： Vijay Kamarshi ,  Qiang Liu

IPC: G06F3/042

CPC classification number: G06F3/042 ,  A63F13/213 ,  A63F13/215 ,  A63F13/26 ,  A63F2003/0418 ,  A63F2003/0428 ,  A63F2009/2461 ,  G06F3/017 ,  G06F3/0425

Abstract: A vision system associated with a projection system includes multiple optical pathways. For instance, when the projection system projects an image onto a generally vertical surface, the vision system may operate in a rear sensing mode, such as for detecting one or more gestures made by a user located behind the projection system. Alternatively, when the projection system projects the image onto a generally horizontal surface the vision system may operate in a front sensing mode for detecting gestures made by a user located in front of the projection system. One or more thresholds may be established for switching between the front sensing mode and the rear sensing mode based on orientation information. As another example, the vision system may be operated in both the front sensing mode and the rear sensing mode contemporaneously.

Abstract translation: 与投影系统相关联的视觉系统包括多个光学路径。 例如，当投影系统将图像投影到大致垂直的表面上时，视觉系统可以在后部感测模式中操作，例如用于检测位于投影系统后面的用户做出的一个或多个手势。 或者，当投影系统将图像投影到大致水平的表面上时，视觉系统可以以前感测模式操作，以检测位于投影系统前面的用户做出的手势。 可以建立一个或多个阈值，用于基于取向信息在前感测模式和后感测模式之间切换。 作为另一示例，视觉系统可以同时在前感测模式和后感测模式中操作。

公开(公告)号：US09361698B1

公开(公告)日：2016-06-07

申请号：US14539238

申请日：2014-11-12

Applicant: Amazon Technologies, Inc.

Inventor： Ji liang Song ,  Qiang Liu

IPC: G06T7/00

CPC classification number: G06T7/0057 ,  G06T7/521

Abstract: A system and method for determining depth information of an object is provided. The system projects dots on an object and captures an image of the object with captured dots. The system identifies the captured dots based on a brightness of the captured dots and identifies coordinates of the captured dots. The system processes the captured image by removing brightness information of pixels outside of the captured dots. The system processes the captured dot by overlaying synthesized dot image on the captured dot. The system divides the processed captured image into captured sections and compares the captured sections to reference sections. The system selects individual reference sections having a highest correlation score for each of the captured sections and correlates a captured dot from each of the captured sections to a reference dot in respective corresponding reference sections. Depth information may be calculated based on coordinates of the captured dots and correlated reference dots.

Abstract translation: 提供了一种用于确定对象的深度信息的系统和方法。 系统在对象上投射点，并用捕获的点捕获对象的图像。 该系统基于捕获的点的亮度来识别所捕获的点，并且识别捕获的点的坐标。 该系统通过去除捕获点之外的像素的亮度信息来处理捕获的图像。 系统通过在捕获的点上覆盖合成的点图像来处理捕获的点。 系统将处理的捕获图像划分为捕获的部分，并将捕获的部分与参考部分进行比较。 系统选择每个捕捉的部分具有最高相关分数的单个参考部分，并将来自每个捕获部分的捕获点与相应的对应参考部分中的参考点相关联。 深度信息可以基于捕获点的坐标和相关参考点来计算。