一种基于KNN的天气现象连续观测识别方法  

Continuous Observation and Identification of Weather Phenomena via KNN

作　　者：刘泗照 徐冠雷[1] 

机构地区：[1]大连舰艇学院,辽宁大连

出　　处：《图像与信号处理》2018年第1期36-42,共7页Journal of Image and Signal Processing

摘　　要：多级天气分类是视频监控和智能交通等诸多潜在应用的基础和重要技术。大多数现有的天气分类方法只考虑天气晴朗或晴天多云的天气条件,并且主要集中在一个固定的场景,如大众旅游和交通情景。本文基于HSV图像色彩模型,主要对各天气现象图片的亮度进行分析和研究,进行亮度间的对比,比较其亮度值的平均值与方差,采用最近邻方法,对所选图像进行分类与识别,达到自动识别各类天气的目的。实验结果证实了此研究方法的可行性。Multi-level weather classification is the basis and important technology of many potential applications such as video surveillance and intelligent transportation. Most of the existing weather classification methods only consider weather conditions that are sunny or sunny cloudy and mainly concentrate in a fixed scene, such as mass tourism and traffic scenarios. Based on the HSV image color model, this paper mainly analyzes and studies the brightness of each weather image, com-paring the brightness, and the average and variance of its brightness value. Using the nearest neighbor method, the selected images are classified and identified, so as to automatically identify various types of weather. The experimental results confirm the feasibility of this method.

关 键 词：天气分类 HSV图像色彩模型 最近邻方法 图像识别 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]