{"title":"Image enhancement for point feature detection in built environment","authors":"V. Lehtola, P. Rönnholm","doi":"10.1109/ICSAI.2014.7009389","DOIUrl":null,"url":null,"abstract":"Image preprocessing is required for texture poor, low-contrast images to enable the functionality of local feature detection algorithms. This is especially true in built environment. We propose a novel image enhancement procedure that conserves the core `signal' in images, while enriching its environment by damping non-essential parts of the signal. Proposed image enhancement method is examined with SIFT, MSER and Harris corner detectors. As a result all detectors, which were originally not able to recover the essential features, improved their performance. Such features were detected also in areas that were not optimally illuminated.","PeriodicalId":143221,"journal":{"name":"The 2014 2nd International Conference on Systems and Informatics (ICSAI 2014)","volume":"17 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2014-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"The 2014 2nd International Conference on Systems and Informatics (ICSAI 2014)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICSAI.2014.7009389","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 2
Abstract
Image preprocessing is required for texture poor, low-contrast images to enable the functionality of local feature detection algorithms. This is especially true in built environment. We propose a novel image enhancement procedure that conserves the core `signal' in images, while enriching its environment by damping non-essential parts of the signal. Proposed image enhancement method is examined with SIFT, MSER and Harris corner detectors. As a result all detectors, which were originally not able to recover the essential features, improved their performance. Such features were detected also in areas that were not optimally illuminated.