{"title":"论水下图像彩色重建的可能性","authors":"Y. Rzhanov, K. Lowell","doi":"10.23919/OCEANS40490.2019.8962735","DOIUrl":null,"url":null,"abstract":"Color is an important cue in object recognition and classification problems. In underwater imagery, colors distort due to light propagation through an absorbing and scattering medium. Distortions depend on a number of complex phenomena, the most important being wavelength-dependent absorption and sensitivity of sensors in trichromatic cameras. Using a simplified model of spectral profile for various colored objects, we show by means of numerical simulation that for fixed hardware and medium properties there exist many spectral profiles that lead to different recorded colors in air (without absorption) but the same color in water (with absorption). The above implies that the restoration of true and unique colors from trichromatic underwater imagery is not possible.","PeriodicalId":208102,"journal":{"name":"OCEANS 2019 MTS/IEEE SEATTLE","volume":"1 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2019-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"On the (im)possibility of color reconstruction in underwater images\",\"authors\":\"Y. Rzhanov, K. Lowell\",\"doi\":\"10.23919/OCEANS40490.2019.8962735\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Color is an important cue in object recognition and classification problems. In underwater imagery, colors distort due to light propagation through an absorbing and scattering medium. Distortions depend on a number of complex phenomena, the most important being wavelength-dependent absorption and sensitivity of sensors in trichromatic cameras. Using a simplified model of spectral profile for various colored objects, we show by means of numerical simulation that for fixed hardware and medium properties there exist many spectral profiles that lead to different recorded colors in air (without absorption) but the same color in water (with absorption). The above implies that the restoration of true and unique colors from trichromatic underwater imagery is not possible.\",\"PeriodicalId\":208102,\"journal\":{\"name\":\"OCEANS 2019 MTS/IEEE SEATTLE\",\"volume\":\"1 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2019-10-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"OCEANS 2019 MTS/IEEE SEATTLE\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.23919/OCEANS40490.2019.8962735\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"OCEANS 2019 MTS/IEEE SEATTLE","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.23919/OCEANS40490.2019.8962735","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
On the (im)possibility of color reconstruction in underwater images
Color is an important cue in object recognition and classification problems. In underwater imagery, colors distort due to light propagation through an absorbing and scattering medium. Distortions depend on a number of complex phenomena, the most important being wavelength-dependent absorption and sensitivity of sensors in trichromatic cameras. Using a simplified model of spectral profile for various colored objects, we show by means of numerical simulation that for fixed hardware and medium properties there exist many spectral profiles that lead to different recorded colors in air (without absorption) but the same color in water (with absorption). The above implies that the restoration of true and unique colors from trichromatic underwater imagery is not possible.