Color Image Enhancement and Segmentation for Computer Vision

计算机视觉的彩色图像增强和分割

• 批准号: 9112350
• 负责人: Jeffrey Rodriguez
• 金额: $ 1.2万
• 依托单位: University of Arizona
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1991
• 资助国家: 美国
• 起止时间: 1991-07-01 至 1992-09-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9112350&HistoricalAwards=false
• 关键词: Color; Image; Enhancement; Segmentation; Computer

The field of computer vision has dealt mostly with monochrome images, leaving the use of color images largely unexplored. Because of advances in computer technology and in monochrome image processing and interpretation, it is reasonable to begin directing more effort toward the incorporation of color information into computer vision systems. New techniques for color image enhancement, analysis, and segmentation need to be developed that apply directly to the vector pixel values, rather than simply applying monochrome techniques to the red, green, and blue component images. In addition, theoretical models for color image processing and interpretation need to be developed so that a unified theory for color computer vision will exist. Planning activities will start with ascertaining which existing techniques can be readily extended to apply to color images and which new techniques must be developed. Also, methods for multispectral image analysis will be reviewed in order to determine which techniques can be adapted to color images. During the planning period, some preliminary analysis of various techniques will be performed through software implementation. This will help determine which methods offer the most promise and warrant further research. The result of these planning activities will be the formulation of a research project for the development of specific image enhancement, analysis, and segmentation procedures for a color computer vision system.

计算机视觉领域主要处理单色图像，而彩色图像的使用很大程度上尚未被探索。 由于计算机技术以及单色图像处理和解释的进步，开始将更多精力投入到将颜色信息纳入计算机视觉系统中是合理的。 需要开发直接应用于矢量像素值的彩色图像增强、分析和分割的新技术，而不是简单地将单色技术应用于红色、绿色和蓝色分量图像。 此外，需要开发彩色图像处理和解释的理论模型，以便存在统一的彩色计算机视觉理论。 规划活动将首先确定哪些现有技术可以轻松扩展以应用于彩色图像以及必须开发哪些新技术。 此外，还将回顾多光谱图像分析方法，以确定哪些技术可以适用于彩色图像。 在规划期间，将通过软件实现对各种技术进行一些初步分析。 这将有助于确定哪些方法最有希望并值得进一步研究。 这些规划活动的结果将是制定一个研究项目，用于开发彩色计算机视觉系统的特定图像增强、分析和分割程序。