中文摘要：

传统RGB图像容易受到光照、姿态、遮挡等因素的影响，深度信息的研究为计算机视觉的发展提供了新思路。该文首先研究了Kinect深度图像成像原理，仿真分析了深度图像的性质及其分割优势，并在Ubuntul2．04系统下基于openCV设计了利用深度信息实现快速目标检测与分割的系统。

英文摘要：

Traditional RGB image is vulnerable to the influence of illumination, pose and occlusion factors, etc. Research on image depth information provides a new idea for the development of computer vision. This paper first studies the Kinect imaging principle, and then analyzes the characters and the segmentation advantage of depth image by simulation. At the end of this paper, a fast depth-information-combined object detection and segmentation system based on the OpenCV library run on the Ubuntu 12.04 is designed and realized.