Challenges in CISE: Automatic Interpretation of High-Altitude Image Data for Eco-System Modeling

9726401 Riseman，Edward Finn，马萨诸塞州约翰大学，Amherst在CEISE中的挑战：用于生态系统建模的高海拔图像数据的自动解释马萨诸塞大学计算机科学系和林业和野生动物管理系的联合研究工作的目标是开发和演示一种实用的方法，使之能够通过航空视频和摄影对自然资源和土地利用进行高度自动化和准确的科学监测。第二个目标是使用同一系统的一部分进行野生动物调查和管理。该研究项目将几乎完全自动化并显著扩展目前使用的手动方法，涉及机载视频数据收集、耗时的手动图像配准以及构建针对植被和地形覆盖的定制专家系统分类器。因此，拟议的计算机科学研究涵盖了计算机视觉、机器学习和高性能计算方面的技术问题。

9726401 Riseman, Edward Finn, John University of Massachusetts, Amherst Challenges in CISE: Automatic Interpretation of High-Altitude Image Data for Eco-System Modeling The goal of the joint research effort between the University of Massachusetts Computer Science Department and the Department of Forestry and Wildlife Management is the development and demonstration of a practical methodology to enable highly automated and accurate scientific monitoring of natural resource and land use via aerial video and photography. A secondary goal is to use portions of the same system for wildlife surveys and management. The research project will almost completely automate and significantly extend the manual methodology currently in use involving airborne video data collection, time-consuming manual image registration, and construction of customized expert system classifiers for vegetation and terrain cover. Thus the proposed computer science research spans technical issues in computer vision, machine learning, and high performance computing.