Depth ranging-based moving target detection method and system

基于深度测距的运动目标检测方法及系统

Abstract

本发明是有关一种基于深度测距的运动目标检测方法及系统,其包括如下步骤:视频摄取,通过位于拍摄区域上方的摄像设备获取包含彩色图像和深度图像的视频信息;运动目标确定,分别对所述视频信息中的彩色图像和深度图像进行前景提取得到包含与运动目标相应运动块的二值图像,对所述二值图像进行融合处理,得到融合后的二值图像;运动目标数量和位置确定,对所述深度图像中在融合后的二值图像判定为运动块区域内的像素信息进行梯度下降聚类处理,得到运动目标的数量和位置,所述像素信息包含像素的像素值与像素位置。本发明实现了运动目标的实时检测、确定,在此基础上分析运动目标的轨迹实现更加精确的运动目标的检测和统计。
The invention relates to a depth ranging-based moving target detection method and a system. The method comprises a step of video acquisition in which a photographing device located above a photographing area is used for acquiring video information with a color image and a depth image included, a step of moving target determination in which foreground extraction is respectively carried out on the color image and the depth image in the video information to obtain a binary image of moving blocks corresponding to the moving target, fusion processing is carried out on the binary image, and a binary image after fusion is obtained, a step of moving target number and position determination in which gradient descent clustering processing is carried out on pixel information in a moving block area judged by the binary image after fusion in the depth image, and the number and positions of the moving targets are obtained, wherein the pixel information comprises pixel values and pixel positions. The moving target can be detected and determined in real time, the trajectory of the moving target is analyzed on the basis, and more accurate moving target detection and statistics can be realized.

Claims

Description

Topics

Download Full PDF Version (Non-Commercial Use)

Patent Citations (4)

    Publication numberPublication dateAssigneeTitle
    CN-101588445-ANovember 25, 2009宁波大学一种基于深度的视频感兴趣区域提取方法
    CN-102288613-ADecember 21, 2011北京科技大学一种灰度和深度信息融合的表面缺陷检测方法
    CN-102737370-AOctober 17, 2012株式会社理光Method and device for detecting image foreground
    CN-103150559-AJune 12, 2013南京理工大学Kinect three-dimensional depth image-based head identification and tracking method

NO-Patent Citations (4)

    Title
    IOANNIS STAMOS 等: "3D Model Construction Using Range and Image Data", 《COMPUTER VISION AND PATTERN RECOGNITION》
    MAURÍCIO PAMPLONA SEGUNDO 等: "Automatic Face Segmentation and Facial Landmark Detection in Range Images", 《SYSTEMS, MAN, AND CYBERNETICS-PART B: CYBERNETICS》
    许素萍,苏松志: "深度图像下基于特征学习的人体检测方法", 《福建电脑》
    钱健: "自由视点视频系统中绘制技术研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》

Cited By (0)

    Publication numberPublication dateAssigneeTitle