Taking photos while diving with the artifacts of water color and turbidity makes underwater photography one of the most challenging tasks. Often times you either have to bring your own daylight flashlights underwater, and or correct the ‘white balance’ while editing the images topside. There’s a similar although more artificial problem in the reef aquarium hobby industry right now where coral vendors areintentionally以蓝光,黄色的橙色摄像机过滤器的形式引入珊瑚色的颜色讨人喜欢的文物,并在照片编辑中吹出设置。
This has resulted in some pretty dishonest, misleading representations of corals for sale, what corals and reef tanks can actually look like in real life – some Instagram pages and coral vendors are literally built on this surreptitious trick. Thankfully, a new tool developed by researchers to ‘remove water’ from subsurface photographs may both be able to dramatically help improve underwater photography, and might even be developed into a tool to re-balance oversaturated or ‘juiced’ coral photos.
Normally in underwater photos we would just white-balance a photo using a neutral black, white or grey chart or chip for a given depth, and this has been enough to correct the internal settings of a camera while diving. The ‘Sea-Thru’ algorithm was developed by oceanographers and engineers from Haifa university using Artificial Intelligence to basically remove the effects of water from underwater photographs.
为了使用Sea-Thru算法,一系列照片中包含了具有精确校准色和灰色尺度的专门色彩图表。这就像白色平衡一样,但不仅仅是灰色阴影,它可以用作大多数颜色和roygbiv阴影的内置参考秤。通过使用此颜色图表拍摄一系列照片并将其插入AI驱动的算法中,图像可以校正,好像没有水。
这种新的成像技术不仅可以成为水下潜水员和摄影师的福音,而且现在比以往任何时候都更需要在水族馆珊瑚照片中看到这种照片真相表。我们长期以来一直将蛋白或人手的明显颜色用作供应出售的照片中的一种晴雨表或颜色参考,但这还不够。如果珊瑚供应商在其珊瑚照片中包括颜色芯片,它可以用作一种“验证”协议真的你得到什么。
For now the Sea-Thru AI imaging algorithm is only in the published phase so we’re a long way off from being able to use it in our own underwater photos or videos, let alone in casual use for aquarium coral photos. In the meantime, you can read our writeup on how to manually“反向光毛”水族馆过饱和水族馆coral photos. [Computer Vision Foundation]