CAJ | 학술논문

为实现视频监控设备对雾天天气现象的自动识别,提出了基于K均值聚类算法的雾天天气现象自动识别方法.该方法通过分析雾天天气现象对视频图像采集的影响,提取图像饱和度的均值、方差为特征参数,并利用K均值聚类算法对训练图像进行分类,得到不同图像类别的聚类中心,测试阶段计算不同图像与聚类中心的相异度即可完成分类.实验结果表明,该方法简洁高效,易于实现对大规模图像数据的处理,并能实现图像分类后类别的标注,对雾天的识别率高于90%.
위실현시빈감공설비대무천천기현상적자동식별,제출료기우K균치취류산법적무천천기현상자동식별방법.해방법통과분석무천천기현상대시빈도상채집적영향,제취도상포화도적균치、방차위특정삼수,병이용K균치취류산법대훈련도상진행분류,득도불동도상유별적취류중심,측시계단계산불동도상여취류중심적상이도즉가완성분류.실험결과표명,해방법간길고효,역우실현대대규모도상수거적처리,병능실현도상분류후유별적표주,대무천적식별솔고우90%.
To realize the foggy weather automatic recognition with video surveillance equipment,a method of foggy weather automatic recognition based on K-means clustering algorithm is put forward,in which the influence of foggy weather on video image acquisition is analyzed,and the mean value of the image saturability and variance are extracted as the characteristic pa-rameters. The training images are classified by using K-means clustering algorithm to obtain the clustering center of the different image classification. In the test stage,the classification can be completed by calculating the dissimilarity of different images and clustering centers. The experimental results show this method is simple and efficient,and easy to realize large-scale image data processing,and can realize the category labeling after image classification. The recognition accuracy is higher than 90%.