A Multi-Spectral Thermal Gas Detection Imager Using Uncooled Infrared Camera

Round 1

Reviewer 1 Report

The authors present in this paper a simple but useful setup for ambient gas detection, and SF₆ release experiment is taken to primarily verify effectiveness. The paper is interesting, well written and perfectly adapted for a publication in Optics. Nevertheless I suggest more or less minor modifications to improve the quality of the article.

Introduction:

I invite the authors to give some examples of advantages of OGI compared to traditional technical methods;

Multispectral gas detection principle:

Caption Fig.1: "Background" instead of "Background";

General remarks: please provide units in your physical parameters (Radiance, path lengths...)

Replace "up" by "high" according to the "H" letter in Eq. 3, 4 & 7

To obtain Eq. 7 the substitution of (6) in (4) is not straightforward. The authors should add an intermediate equation facilitating the obtention of Eq. 7.

I suggest also to give the expressions of Of and Gf which appear in Eq. (8)

To obtain Eq. (9), do you estimate Tf=Sf/Of? and Rf=lambda_H-lambda_L? Please clarify this point?

A multi-spectral uncooled camera for gas imaging and radiation correction method

Fig.2 and Fig. 3 should be gathered in a unique Fig. 2 (a) and (b);

Give the unit of spatial resolution in Table 1;

LWIR in page 5 should be defined;

The materials, the thickness and the measured spectral responses of the 5 filters should be precised in this section. We don't know if the spectral transmittance curves un Fig.4 are theoretical or experimental? 

This is not an "absorption peak" for SF6 but an "absorption vibrational band" centered around 10.5 micrometer. I suggest to precise what kind of vibrational band is associated with the absorption shown in Fig.4;

The color's scale should be given in Fig. 5 and its caption should be more detailed: signification of the 5 row figures? What kind of correction? What kind of thermal emissions are observed?

The authors should explain how the calibration performed in Fig. 6 affect the temperature image retrieval explained in section 2?

Gas classification and SF6 experiment

According to Fig. 4, I don't understand "SF6 has weak absorption in filter 1-3, slightly strong absorption in filter 4, and none absorption in filter 6." filter 6 mean filter 5? Explain how the absorption of filters 3 to 1 decrease in non-saturation conditions?

It will be interesting to have an idea of the SF6 targeted concentration in ppm/ppb unit. Moreover in Fig.8 "l" is missing in "molecules"

SVM should be defined. Why this method? Which advantage?

For Fig. 9, the authors should give  the orgin of the data used to simulate the atmosphere transmittance (HITRAN?) and how the multispectra was simulated with this atmosphere. The authors should refer again to the eq. of section 2.

P. 9 "substruction" instead of "subtraction";

Fig. 7 and 8 and Fig. 9 and 10 could be gathered in one figure Fig. 7a and 7b/ Fig. 8a and 8b;

Results shown on Fig. 11 are very impressive but I think that its construction (i.e the yellow part associated with the SF6 absorption optimized with filter 4) is not well explained. The quantity of gas release and the time frame from 1 to 6 should be given. 

Conclusion

It would have been interesting to experiment with another gas (another GHG) to illustrate the selectivity of the method? The sentence "More performance experiments are needed to verify system detection ability." is a little light for a conclusion???

Author Response

Introduction:

I invite the authors to give some examples of advantages of OGI compared to traditional technical methods;

A: This setup is used in SF6 leakage detection in electric grids, so the example is given by SF6 detection in paragraph 1.

Multispectral gas detection principle:

Caption Fig.1: "Background" instead of "Background";

A: Mistakes are fixed.

General remarks: please provide units in your physical parameters (Radiance, path lengths...)

A: Mistakes are fixed.

Replace "up" by "high" according to the "H" letter in Eq. 3, 4 & 7

A: Mistakes are fixed.

To obtain Eq. 7 the substitution of (6) in (4) is not straightforward. The authors should add an intermediate equation facilitating the obtention of Eq. 7.

A: Equation (6)-(9) are added. See in section 2.

I suggest also to give the expressions of Of and Gf which appear in Eq. (8)

To obtain Eq. (9), do you estimate Tf=Sf/Of? and Rf=lambda_H-lambda_L? Please clarify this point?

A: Equation 9 is rewritten and Gf is explained in text.  is system gain coefficient representing filter’s attenuation effect.

A multi-spectral uncooled camera for gas imaging and radiation correction method

Fig.2 and Fig. 3 should be gathered in a unique Fig. 2 (a) and (b);

A: Figures are rearranged.

Give the unit of spatial resolution in Table 1;

A: Mistakes are fixed.

LWIR in page 5 should be defined;

A: Mistakes are fixed.

The materials, the thickness and the measured spectral responses of the 5 filters should be precised in this section. We don't know if the spectral transmittance curves un Fig.4 are theoretical or experimental? 

A: Filters are made of silicon with 5mm thick 25mm diameter. The curve is theoretical display according to wave band cutoff.

This is not an "absorption peak" for SF6 but an "absorption vibrational band" centered around 10.5 micrometer. I suggest to precise what kind of vibrational band is associated with the absorption shown in Fig.4;

A: Mistakes are fixed.

The color's scale should be given in Fig. 5 and its caption should be more detailed: signification of the 5 row figures? What kind of correction? What kind of thermal emissions are observed?

A: The unevenness in figure 4 are shown in rectangle areas. The correction are performed by subtraction of ambient blackbody. Thermal emissions embody as unevenness in thermal picture.

The authors should explain how the calibration performed in Fig. 6 affect the temperature image retrieval explained in section 2?

 A: The calibration fiting in Figure 5 is aimed to adjust temperature value altered by filter installment, which embodies as coefficient  in equation . Each filter has different coefficient because of different waveband cutoffs. After this calibration procedure, the output image is real temperature that removal of filters’ uneven reflection and attenuation.

Gas classification and SF6 experiment

According to Fig. 4, I don't understand "SF6 has weak absorption in filter 1-3, slightly strong absorption in filter 4, and none absorption in filter 6." filter 6 mean filter 5? Explain how the absorption of filters 3 to 1 decrease in non-saturation conditions?

A: The reason that decrease of filters 3 to 1 in Figure 6 is SF6 absorption peak integration varies with filter’s waveband. The small proportion of SF6 peak area in filter’s waveband corresponds to small decrease value.

It will be interesting to have an idea of the SF6 targeted concentration in ppm/ppb unit. Moreover in Fig.8 "l" is missing in "molecules"

A: Units change to ppm*m.

SVM should be defined. Why this method? Which advantage?

A: SVM explanations are added in the end of 4.1.

For Fig. 9, the authors should give  the orgin of the data used to simulate the atmosphere transmittance (HITRAN?) and how the multispectra was simulated with this atmosphere. The authors should refer again to the eq. of section 2.

A: SF6 transmittance is calculated according equation , and atmosphere transmittance is simulated by MODTRAN. The combined transmittance of atmosphere and SF6 is multiplication of both.

1. 9 "substruction" instead of "subtraction";

Fig. 7 and 8 and Fig. 9 and 10 could be gathered in one figure Fig. 7a and 7b/ Fig. 8a and 8b;

A: Mistakes are fixed.

Results shown on Fig. 11 are very impressive but I think that its construction (i.e the yellow part associated with the SF6 absorption optimized with filter 4) is not well explained. The quantity of gas release and the time frame from 1 to 6 should be given. 

A: SF6 flow rate is 2 L/min. Every picture in figure8 is taken in 8 seconds interval.

Conclusion

It would have been interesting to experiment with another gas (another GHG) to illustrate the selectivity of the method? The sentence "More performance experiments are needed to verify system detection ability." is a little light for a conclusion???

A: Some additional information is added.

Author Response File: Author Response.pdf

Reviewer 2 Report

This paper considers the detection of harzardous and toxic gases using IR camera and wide band filters. SF6 is used as an example. The image uneveness can be solved with the use of a blackbody at ambient temperature. It has interesting application for industry concerning the harzadous and toxic gas leakage accidents.

Before publication, I have the suggestions below:

The blackbody radiance B in equation (5) is function of only one variable, the wavelengh (or the temperature if we consider the Wien displacement law). So, it is not necessary a partial derivative. Or if you want to use partial derivatives, it would be better write the radiance B as function of 2 varaibles, B(\lambda, T).

It is not so clear why when we substitute eq. (6) in eq. (4) we get eq. (7) as claimed in the text right before eq. (7).

I also have the following suggestions regardng the writting:

- Just after section 3.2 it is written The output of commercial IR camera is temperature image… Maybe it is a typo and you tried to write The output of commercial IR camera is the temperature image… or The output of commercial IR camera is temperature imager…

- Before eq. (6), I suggest Substitute (5) into (1)… omitted and equation (1) is written as…

- In first paragraph of section 4.1, last phrase says, Intensity maybe lower… and it should be Intensity may be lower...

Comments for author File: Comments.pdf

Author Response

This paper considers the detection of harzardous and toxic gases using IR camera and wide band filters. SF6 is used as an example. The image uneveness can be solved with the use of a blackbody at ambient temperature. It has interesting application for industry concerning the harzadous and toxic gas leakage accidents.

Before publication, I have the suggestions below:

The blackbody radiance B in equation (5) is function of only one variable, the wavelengh (or the temperature if we consider the Wien displacement law). So, it is not necessary a partial derivative. Or if you want to use partial derivatives, it would be better write the radiance B as function of 2 varaibles, B(\lambda, T).

A: Mistakes are fixed.

It is not so clear why when we substitute eq. (6) in eq. (4) we get eq. (7) as claimed in the text right before eq. (7).

A: This part is rewritten. All equations are rededuced.

I also have the following suggestions regardng the writting:

- Just after section 3.2 it is written The output of commercial IR camera is temperature image… Maybe it is a typo and you tried to write The output of commercial IR camera is the temperature image… or The output of commercial IR camera is temperature imager…

- Before eq. (6), I suggest Substitute (5) into (1)… omitted and equation (1) is written as…

- In first paragraph of section 4.1, last phrase says, Intensity maybe lower… and it should be Intensity may be lower...

A: Mistakes are fixed.

Author Response File: Author Response.pdf