how to find the difference between fire and sun rays bcoz , both contains red and yellow color
조회 수: 3 (최근 30일)
이전 댓글 표시
how to find the difference between fire and sun rays bcoz , both contains red and yellow color.How to find the pixel intensities between those two.
댓글 수: 0
채택된 답변
Image Analyst
2016년 12월 16일
You need to look at the spectrum. There you will see lots of differences, though they're both somewhat similar to black body radiation. Just Google solar spectrum images and you'll see it. The spectrum of fire will depend on what's burning and at what temperature it's at. Basically different things are burning so their color will be slightly different. Sunlight peaks in the green of course because it is very very hot - much hotter than your typical campfire, which will have a spectrum shifted more towards the yellow and red with less in the blue. Additionally sunlight will have several absorption bands for things in the atmosphere like Co2 and water.
댓글 수: 1
Walter Roberson
2016년 12월 16일
You probably cannot perform an appropriate spectral analysis if all you have is ordinary photographs.
Looking up Wien's Law and the temperatures of wood fires, remembering to convert to Kelvin from Fahrenheit, we find that peak output from a wood fire, according to black body laws, would be about 5730 nm to 3145 nm, both in the far infrared, with un-scattered sunshine peaking around 506 nm (green). Without special equipment you are simply not going to be able to measure over that entire range.
추가 답변 (1개)
Walter Roberson
2016년 12월 16일
You cannot determine the difference between the two based upon intensity. A fire photographed in the dark might saturate the sensor intensity (depending on the settings), whereas sunlight shining on something during the day might not saturate the sensors (depending on the settings.) At the absolute minimum you would need information about the sensor settings at the time of the photograph in order to make any kind of assessment of what high intensity meant in the situation.
Working by color instead of by intensity has a number of difficulties:
- What if it was evening sunlight near sunset, where the sunlight is being defracted the atmosphere more than usual, and possibly having its frequency modified by atmospheric dust (such as a red sunset)?
- Is the photograph of the Sun directly or is it of sunshine on something? If it is sunshine on something then the resulting reflected color depends upon the color and the absorption profile of what is being illuminated
- converting from sRGB to frequency is complicated and requires a bunch of assumptions that are probably false unless you design the photography circumstances specifically
- if you are working with reflected light, then estimating the temperature of illumination by doing a Black Body Radiation estimation is probably going to fail
- peak Black Body temperatures for regular fires are in the infrared range, where your sensors probably are not measuring. And you would need to measure in multiple infrared ranges in order to figure out where the peak is
So... if you are just working with ordinary photographs, you probably cannot really determine the difference, not in a scientific way. But if the task is "classification" by neural networks or something similar, where you are permitted inaccuracy, then you might be able to work something out that was "good enough" -- and the "good enough" might come down to just measuring average brightness because pictures of fires tend to be darker than pictures of objects in sunshine.
댓글 수: 3
Walter Roberson
2016년 12월 16일
편집: Walter Roberson
2016년 12월 16일
Good point. If you were specifically monitoring for fires then an IR camera set for the range from roughly 8000 nm (70C or so) to about 3000 nm (650C or so) would find abnormalities (70C to 200C) as well as fires (leather will self-combust as low as 200C it appears.)
참고 항목
카테고리
Help Center 및 File Exchange에서 Annotations에 대해 자세히 알아보기
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!