场景图像分类技术综述

Abstract

Scene classi cation is one of the primary problems in robotics research, as well as an important challenge in the eld of computer vision. With the rapid development of machine vision technologies in recent years, numerous scene classi cation methods have been developed. These methods cover a wide range of computer vision tasks. To build a relatively complete system of scene recognition, we acquired a deep knowledge of fundamental scene classi cation technologies; the resulting research review is provided in this article. First, a very brief review of scene classi cation developments is provided. The main domestic and international scene classi cation methods are then described in detail, beginning with feature extraction. A systematic analysis, comparison, and summary are then presented, and potential future developments in scene classi cation technology are discussed in the conclusion.