High-speed 3-dimensional space recognition method based on moving images of a monocular camera

如果机器人要在日常生活中真正帮助人们，三维坐标的识别是必不可少的。我们的方法来自于常见的视觉特征，即近的东西移动很多，而远的东西移动很少。我们提出了一种方法，可以从视频图像中检测到的相对运动速度假设一个三维空间。通过直线段和匹配点的检测，实现了快速的处理。线段比像素信息更简单、更紧凑。我们描述了一个高速系统，假设机器人的环境，从线段。首先，在2004年，当它是第一年，我们计划建立转换技术的图像信息表示为一条线，并检查三维空间感知技术的线段表示。我们计划建设他们使用该技术的系统，以及一系列实时处理计划建设可能的实验环境。2005年是第二年，空间感知算法的高速精度的数量检查了伙伴机器人的系统化，每个项目的多-我认识到一个高质量的采集图像的三维信息的目的使用能够放在这个系统中。我们最后包括一个问题的精度，但建立了基本的算法来执行三维空间感知的机器人。此外，为了实用化，我还进行了改进，构建了自动移动机器人的系统。今后，为了提高精度，我们将通过多次测量来验证技术的建立。此外，在实际使用方面，我们将计划构建我聚集在一起的系统，并使用通用相机和机器人，以便实现低成本。

The recognition of 3-dimensional coordinates is essential if robots are to truly help people in daily life. Our approach comes from the common visual characteristic that near things move a lot while far things move little. We propose a method that can hypothesize a 3-dimensional space from the relative movement speeds detected in video images. High processing speeds are achieved through the detection of line segments and matching points.The method proposed here extracts line segments from the images. Line segments are simpler and more compact than pixel information. We describe a high-speed system that hypothesizes the robot's environment from the line segments.At first, in 2004 when it was the first year, we planned establishment of conversion technique to expression for image information for a line and examined three dimensions space perception technique by segment of a line expression. And we planned construction of the system which they used this technique for, and a series of real-time processing planned construction of possible experiment environment.In 2005 when it was the second year, amount of high speed precision of space perception algorithm examined systematization of a partner robot, each item of the multi-purpose use of the three dimensions information that I recognized a high quality of an acquisition image to be able to put in this system.We finally included a problem of precision, but established basic algorithm to perform three-dimensional space perception by a robot. In addition, I advanced implementation for practical use and built a system for an auto moving robot.We are going to examine establishment of technique aiming at improvement of precision in future by measuring multiple times. In addition, in practical use side, we are going to plan construction of the system which I get together, and used a general-purpose camera and robot so that low cost can become realize.