Hadoop下煤矿图像PCA-SIFT算法改进及研究  

PCA-SIFT Feature Extraction Algorithm of Coal Mine Image in Hadoop Platform

作　　者：史昊[1] 曹建芳[2] 陈立潮[1] 潘理虎[1,3] SHI Hao;CAO Jian-fang;CHEN Li-chao;PAN Li-hu(School of Computer Science and Technology,Taiyuan University of Science and Technology,Taiyuan 030024,China;Department of Computer Science and Technology,XinZhou Teachers’University,XinZhou 034000,China;Institute of Geographic Science and Natural Resource Research,Chinese Academy of Science,Beijing 100101,China)

机构地区：[1]太原科技大学计算机科学与技术学院,太原030024 [2]忻州师范学院计算机科学与技术系,山西忻州034000 [3]中国科学院地理科学与资源研究所,北京100101

出　　处：《太原科技大学学报》2020年第2期124-129,共6页Journal of Taiyuan University of Science and Technology

基　　金：山西省自然基金项目(2014011019-3);山西省科技重大专项(20121101001);山西省中科院科技合作项目(20141101001)。

摘　　要：大数据技术是当下热点问题,其中,Hadoop技术在煤矿领域的应用也引起了广泛关注。针对传统煤矿视频监控系统图像采集点多,单机环境煤场景矿图像特征提取效率低下等问题,提出了一种基于Hadoop的PCA-SIFT图像特征并行提取算法,该算法采用PCA算法对并行提取的煤矿图像的SIFT特征进行了并行降维处理。使用汾西矿务局煤矿图像井下数据集进行实验,提出的算法SIFT特征点检测效果好,运行耗时少,图像数量加剧后,系统加速比几乎呈线性增长的趋势,验证了算法处理大规模煤矿图像数据的有效性。Big data technology is the hot issue,and the Hadoop technologies applied in the field of coal mine has obtained wide public concern. In the case of the traditional mining of video,the image of a single-plane environment coal scene is an inefficiently extracted image,and this is a kind of a PCA-SIFT image based on Hadoop,which is an algorithm that uses PCA’s algorithm to do parallel to the SIFT of the mining image in parallel. Using images in coal mine underground mining bureau FenXi experiment data set,the detecting effect of algorithm SIFT feature point is good,the operation time is less,the systematic acceleration ratios shows almost linear growth trend with the number of intensified image,and the effectiveness of algorithm dealing with large-scale coal image data is verified.

关 键 词：煤矿 HADOOP SIFT特征 PCA 降维 

分 类 号：TD301.6[矿业工程—矿井建设]