{"title":"基于特征融合的多背景伪装图案设计方法研究","authors":"Haibo Fang, R. Peng, Xianggang Sha, Yongsheng Lv","doi":"10.1109/ICVISP54630.2021.00056","DOIUrl":null,"url":null,"abstract":"In order to solve the problem that the single-background camouflage patterns designed by the existing methods cannot meet the camouflage requirements under the multi-terrain background, a multi-background camouflage pattern design method based on feature fusion is proposed, and formed a complete set of multi-background camouflage pattern design theory. Through similarity analysis of background images, camouflage color determination, patch extraction and other steps, the color and patch shape features are extracted, and then the features are merged to generate a multi-background camouflage pattern, which is suitable for multi-regional camouflage. The simulation experiment results show that the multi-background camouflage pattern is better integrated with each background in terms of brightness, color and texture characteristics. Compared with the single-background camouflage pattern, the camouflage similarity index (CSI) increased by 8.1% and 10.8% respectively.","PeriodicalId":296789,"journal":{"name":"2021 5th International Conference on Vision, Image and Signal Processing (ICVISP)","volume":"51 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Research on Multi-Background Camouflage Pattern Design Method Based on Feature Fusion\",\"authors\":\"Haibo Fang, R. Peng, Xianggang Sha, Yongsheng Lv\",\"doi\":\"10.1109/ICVISP54630.2021.00056\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"In order to solve the problem that the single-background camouflage patterns designed by the existing methods cannot meet the camouflage requirements under the multi-terrain background, a multi-background camouflage pattern design method based on feature fusion is proposed, and formed a complete set of multi-background camouflage pattern design theory. Through similarity analysis of background images, camouflage color determination, patch extraction and other steps, the color and patch shape features are extracted, and then the features are merged to generate a multi-background camouflage pattern, which is suitable for multi-regional camouflage. The simulation experiment results show that the multi-background camouflage pattern is better integrated with each background in terms of brightness, color and texture characteristics. Compared with the single-background camouflage pattern, the camouflage similarity index (CSI) increased by 8.1% and 10.8% respectively.\",\"PeriodicalId\":296789,\"journal\":{\"name\":\"2021 5th International Conference on Vision, Image and Signal Processing (ICVISP)\",\"volume\":\"51 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2021-12-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2021 5th International Conference on Vision, Image and Signal Processing (ICVISP)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/ICVISP54630.2021.00056\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2021 5th International Conference on Vision, Image and Signal Processing (ICVISP)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICVISP54630.2021.00056","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Research on Multi-Background Camouflage Pattern Design Method Based on Feature Fusion
In order to solve the problem that the single-background camouflage patterns designed by the existing methods cannot meet the camouflage requirements under the multi-terrain background, a multi-background camouflage pattern design method based on feature fusion is proposed, and formed a complete set of multi-background camouflage pattern design theory. Through similarity analysis of background images, camouflage color determination, patch extraction and other steps, the color and patch shape features are extracted, and then the features are merged to generate a multi-background camouflage pattern, which is suitable for multi-regional camouflage. The simulation experiment results show that the multi-background camouflage pattern is better integrated with each background in terms of brightness, color and texture characteristics. Compared with the single-background camouflage pattern, the camouflage similarity index (CSI) increased by 8.1% and 10.8% respectively.