本研究致力于开发一种智能铁屑清理移动机器人的视觉寻迹方法,旨在提高机器人在复杂工业环境中对铁屑的自动清理效率。研究重点在于图像处理和分析,特别是图像阈值的自动选择,这是实现精确图像分割的关键。本文综合分析了直方图波谷法、...本研究致力于开发一种智能铁屑清理移动机器人的视觉寻迹方法,旨在提高机器人在复杂工业环境中对铁屑的自动清理效率。研究重点在于图像处理和分析,特别是图像阈值的自动选择,这是实现精确图像分割的关键。本文综合分析了直方图波谷法、Otsu法、最大熵方法等现有技术,并针对单摄像头视觉伺服系统的挑战,提出了一种改进的图像分割算法。针对现有技术的局限性,如光照变化敏感性和对静态图像的依赖,本文设计了一种基于单目视觉的稳定可靠的智能铁屑清理机器人寻迹方法。该方法通过RGB色彩空间处理,利用动态阈值提取技术,有效解决了光照变化和栅格遮挡下的实时彩色图像分割问题。实验部分验证了所提方法的有效性。通过对智能铁屑清理移动机器人在实际行进过程中的寻迹图像进行分割,结果表明该方法能够适应不同光照条件,且分割速度快,满足了实时性要求。This study is dedicated to the development of a visual tracking method for intelligent iron shaving cleaning mobile robots, aiming to enhance the efficiency of robots in automatically cleaning iron shavings in complex industrial environments. The research focuses on image processing and analysis, particularly the automatic selection of image thresholds, which is key to achieving accurate image segmentation. This paper comprehensively analyzes existing technologies such as histogram valley method, Otsu’s method, and maximum entropy method, and proposes an improved image segmentation algorithm in response to the challenges of single-camera visual servo systems. Addressing the limitations of existing technologies, such as sensitivity to lighting changes and reliance on static images, this paper designs a stable and reliable tracking method for iron removal robot based on monocular vision. The method processes images in the RGB color space, using dynamic threshold extraction technology to effectively solve the real-time color image segmentation problem under varying lighting conditions and grid obstructions. The experimental part verifies the effectiveness of the proposed method. By segmenting the tracking images of the intelligent iron shaving cleaning mobile robot during actual travel, the results show that the method can adapt to different lighting conditions and has a fast segmentation speed, meeting the real-time requirements.展开更多
文摘本研究致力于开发一种智能铁屑清理移动机器人的视觉寻迹方法,旨在提高机器人在复杂工业环境中对铁屑的自动清理效率。研究重点在于图像处理和分析,特别是图像阈值的自动选择,这是实现精确图像分割的关键。本文综合分析了直方图波谷法、Otsu法、最大熵方法等现有技术,并针对单摄像头视觉伺服系统的挑战,提出了一种改进的图像分割算法。针对现有技术的局限性,如光照变化敏感性和对静态图像的依赖,本文设计了一种基于单目视觉的稳定可靠的智能铁屑清理机器人寻迹方法。该方法通过RGB色彩空间处理,利用动态阈值提取技术,有效解决了光照变化和栅格遮挡下的实时彩色图像分割问题。实验部分验证了所提方法的有效性。通过对智能铁屑清理移动机器人在实际行进过程中的寻迹图像进行分割,结果表明该方法能够适应不同光照条件,且分割速度快,满足了实时性要求。This study is dedicated to the development of a visual tracking method for intelligent iron shaving cleaning mobile robots, aiming to enhance the efficiency of robots in automatically cleaning iron shavings in complex industrial environments. The research focuses on image processing and analysis, particularly the automatic selection of image thresholds, which is key to achieving accurate image segmentation. This paper comprehensively analyzes existing technologies such as histogram valley method, Otsu’s method, and maximum entropy method, and proposes an improved image segmentation algorithm in response to the challenges of single-camera visual servo systems. Addressing the limitations of existing technologies, such as sensitivity to lighting changes and reliance on static images, this paper designs a stable and reliable tracking method for iron removal robot based on monocular vision. The method processes images in the RGB color space, using dynamic threshold extraction technology to effectively solve the real-time color image segmentation problem under varying lighting conditions and grid obstructions. The experimental part verifies the effectiveness of the proposed method. By segmenting the tracking images of the intelligent iron shaving cleaning mobile robot during actual travel, the results show that the method can adapt to different lighting conditions and has a fast segmentation speed, meeting the real-time requirements.
