The Knowledge of thermal imaging temperature camera (part 2)
3. Temperature compensation
Can we start thermal imaging temperature camera after finishing the above two steps? The process described above is just the ideal situation.
The body temperature of the refrigeration equipment is basically constant, and external influences can be ignored. However, due to considerations such as cost and craftsmanship, most of the current civilian detectors are basically non-refrigeration detectors. The temperature of the uncooled detector itself is affected by the environment, and the temperature affects the response characteristics of the detector.
In addition to the target radiation, the radiation induced by the detector also has a part of the barrel radiation, which also affects the response of the detector.
The above two factors are greatly affected by the ambient temperature. In practical applications, a certain temperature compensation is generally performed according to the ambient temperature to correct the measured value. Compensation is not easy to do. To ensure a better use effect, the simplest recommendation is to use the thermal imaging device in an indoor environment.
In addition to these factors, combined with the atmospheric transmission problem mentioned in our previous article, the distance of the object will also affect the accuracy of temperature camera, and distance compensation can be considered in practical applications.
There are errors in every link of the whole process, and the accumulation of errors makes temperature camera difficult. Therefore, the nominal error range in the general human temperature camera industry is ±0.5°C (how much moisture is here, I don’t know, I actually think It's still hard to think). Ideally, if the use environment and the calibration environment are exactly the same, the temperature camera can still be done very accurately.
Add blackbody temperature camera
Can the actual use environment be turned into a calibration environment? Some manufacturers have taken the more extreme path and put the black body directly in the application scene for calibration.
To put it simply, the traditional temperature camera can be called "offline calibration", and the temperature camera with a black body added to the scene is called "online calibration".
The characterization of the detector is relatively easy. For example, the mapping curve of the detector response and temperature is difficult to quantify. The black body in the scene directly solves the quantitative problem. In this case, the error can generally be ±0.3°C. However, this method also brings more problems, such as inconvenient installation, sharp increase in cost, and limited application environment.
If temperature camera is a difficult problem, adding boldface is a problem-solving method, but I don't like this method of problem-solving too much, it is too simple and rude, and lacks a bit of elegance.
AI temperature camera
In addition to the above two traditional methods, the so-called "AI temperature camera" has also emerged in the market recently. Because the public's perception of AI is still biased, it seems that bringing AI things is unfathomable. But we believe that the clearer the public's understanding of AI, the more conducive to the healthy development of the industry's ecology, so today we will also talk about AI temperature camera.
To keep it simple, use one sentence to make it clear that AI temperature camera: treat people as black bodies.
The emissivity of the human skin is determined, and the average temperature of the population tends to be stable. It is too suitable for blackbody!!! The selection of the temperature area is done on the algorithm, and the online calibration strategy is determined. The actual effect will not be too bad.
That's all about AI temperature camera, isn't it simple?
Accurate temperature camera is difficult. The standard for fever is generally set at 37.3 degrees Celsius. Even if a black body is added to the environment, there will be a large number of false positives and under-reports within the error level.
Thermal imaging is used for temperature camera. It is difficult to accurately quantify (±0.1°C accuracy), but it is easy to qualitatively (judge who has the highest temperature in the crowd).
Dear customers and friends, forehead temperature itself will also be affected by many factors. Infrared temperature camera is good for preliminary temperature screening, and it must not be completely relied on.
The Knowledge of thermal imaging temperature camera (part 2) - Safeagle