Design of an Airborne Low-Light Imaging System Based on Multichannel Optical Butting
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsFacing the challenge of airborne low-light imaging under 0.01lux illuminance conditions, the paper conducted parameter analysis and designed an optical system with adjustable apertures from F1.6 to F20 to accommodate low-light imaging and a wide dynamic range. The highlight of the optical design lies in the arrangement of six prism junctions split into eight channels for seamless stitching of a large field of view. The paper carried out theoretical analysis of signal-to-noise ratio under low-light conditions, laboratory testing, as well as external imaging tests at 0.01lux and flight tests at 1 ~ 10lux, demonstrating the ability to achieve high-contrast color imaging under 0.01lux conditions. The related research and workflow have significant reference value.
1 There are certain errors in the signal-to-noise ratio analysis in Section 2.3.3. The formulas (3) to (8) have illuminance values represented in lux, while the rest of the units in the formulas are in the International System of Units. In this case, lux must be converted to W/m^2, usually based on a specific wavelength, such as 1 W/m^2 = 683 lux (λ=550nm) or 1 W/m^2 = 180 lux (λ=380~680nm), or others. After calculation, it is confirmed that the article used 0.01 (lux), resulting in Nsig=154.2e, whereas when using W/m^2 as the unit, Nsig=0.2257~0.8567e, leading to subsequent errors in signal-to-noise ratio calculations. The optical system at F1.6 imaging under 0.01lux with a 4.5ms integration time has an extremely low photon signal-to-noise ratio limit, raising significant theoretical concerns about achieving imaging at 0.01lux.
2 There is a flaw in the first part of formula (6) on line 205, as integrating over quantum efficiency is not meaningful. Integration would typically be necessary when illuminance is in units of W/m^2/um. The correct form should be Nsig=sum E’(λ)tmAsλ/hcη(λ) dλ= E’tmAsηλ/hc; the first part of the formula can be omitted, retaining only the latter part.
3 In Table 5, when illuminance is 0.01lux, the average image gray level is 113.3DN. However, when the illuminance is increased by a factor of 1 million to 10,000 lux, the gray level only increases by approximately 7 times to 786.2DN, which is incorrect for a linear response sCMOS sensor. Clarification is needed if there were changes in test conditions such as the F-number, integration time, gain, or whether dark frame subtraction was applied to the images.
4 Section 2.3.3 analyzed the signal-to-noise ratio of the camera at 0.01lux and conducted laboratory imaging experiments at that illuminance level as shown in Table 5. It is suggested to include the actual measured signal-to-noise ratio at 0.01lux and compare it with the theoretical values.
5 The paper conducted external field tests at 0.01lux and flight tests at 1~10lux conditions, obtaining color low-light images. To enhance credibility, it would be beneficial to perform quantitative analysis on image quality metrics such as signal-to-noise ratio, contrast, or dynamic range, validating them against theoretical analyses or comparing them horizontally with other cameras.
Author Response
Response to Reviewer 1 Comments
Point 1: There are certain errors in the signal-to-noise ratio analysis in Section 2.3.3. The formulas (3) to (8) have illuminance values represented in lux, while the rest of the units in the formulas are in the International System of Units. In this case, lux must be converted to W/m^2, usually based on a specific wavelength, such as 1 W/m^2 = 683 lux (λ=550nm) or 1 W/m^2 = 180 lux (λ=380~680nm), or others. After calculation, it is confirmed that the article used 0.01 (lux), resulting in Nsig=154.2e, whereas when using W/m^2 as the unit, Nsig=0.2257~0.8567e, leading to subsequent errors in signal-to-noise ratio calculations. The optical system at F1.6 imaging under 0.01lux with a 4.5ms integration time has an extremely low photon signal-to-noise ratio limit, raising significant theoretical concerns about achieving imaging at 0.01lux..
Response 1: Thank you for pointing out the significant errors in the manuscript and providing us with the correct calculation method. We have recalculated based on the conversion relationship between photometric and radiometric measurements.
Point 2: There is a flaw in the first part of formula (6) on line 205, as integrating over quantum efficiency is not meaningful. Integration would typically be necessary when illuminance is in units of W/m^2/um. The correct form should be Nsig=sum E’(λ)tmAsλ/hcη(λ) dλ= E’tmAsηλ/hc; the first part of the formula can be omitted, retaining only the latter part..
Response 2: We have made revisions to the manuscript according to your feedback.
Point 3: In Table 5, when illuminance is 0.01lux, the average image gray level is 113.3DN. However, when the illuminance is increased by a factor of 1 million to 10,000 lux, the gray level only increases by approximately 7 times to 786.2DN, which is incorrect for a linear response sCMOS sensor. Clarification is needed if there were changes in test conditions such as the F-number, integration time, gain, or whether dark frame subtraction was applied to the images.
Response 3: In laboratory testing, for low illumination conditions with an illuminance of 0.01 Lux, a longer exposure time and greater gain were adopted. For high illumination environments, the exposure time and gain are reduced, otherwise the image will show obvious overexposure. I have added clarification in the manuscript.
Point 4: Section 2.3.3 analyzed the signal-to-noise ratio of the camera at 0.01lux and conducted laboratory imaging experiments at that illuminance level as shown in Table 5. It is suggested to include the actual measured signal-to-noise ratio at 0.01lux and compare it with the theoretical values.
Response 4: I have consulted many articles and materials in order to obtain methods for laboratory testing signal-to-noise ratio. To test the signal-to-noise ratio of an imaging system using an integrating sphere, it is necessary to collect a large number of images under different illumination conditions and obtain noise grayscale through comparison. In the preliminary research work, this part of the work was not carried out, so the actual signal-to-noise ratio of the imaging system was not obtained. In the future, I will carry out this part of the work in our new manuscript.
Point 5: The paper conducted external field tests at 0.01lux and flight tests at 1~10lux conditions, obtaining color low-light images. To enhance credibility, it would be beneficial to perform quantitative analysis on image quality metrics such as signal-to-noise ratio, contrast, or dynamic range, validating them against theoretical analyses or comparing them horizontally with other cameras.
Response 5: Indeed, due to factors such as weather and aircraft control, flight tests were not conducted under 0.01 Lux environmental conditions. The manuscript conducted quantitative analysis on image quality and signal-to-noise ratio. In the laboratory, some tests were conducted on the actual image quality indicators of the imaging system, but these indicators may not reflect the true performance and imaging quality of the imaging system. In the future, we will conduct new flight tests to test the performance of this imaging system.
Author Response File: 
Author Response.doc
Reviewer 2 Report
Comments and Suggestions for AuthorsThe authors developed and tested a camera system that can capture high-resolution, wide-angle images in low-light conditions from long distances. They used 8 SCOMs detectors to splice the optical field of view without adding too much bulk or weight to the system. The design of the reflective prism was innovative and was calculated in the manuscript. The research in this manuscript fits with the theme of the special issue "Optical Imaging and Measurements." The study's findings and design approach can be useful for creating low-light imaging equipment for airborne use.
I have taken note of the following points for your manuscript:
1. There are noticeable seams in the images in Figures 15 and 16, and the gray-scale of the images appears to be different. Please explain this issue and propose a solution to rectify it.
2. It would be beneficial to include an explanation of the relationship between image gray values and SNR in Chapter 4.
3. Consider bolding the explanatory text in the images to ensure clarity. This should be done for Figures 1, 4, 5, 6, 8, 9, 10, and 11.
4. Check the citation format of Figure 3 and Figure 16 (Line 157, 349).
5. Use "km" instead of "Km" to ensure uniformity in the units of measurement in the manuscript (Table 3).
6. Please review and correct any minor typos, such as those on line 6.
7. Verify the accuracy of your references and ensure that they adhere to citation standards.
Comments on the Quality of English LanguageMinor editing of English language required
Author Response
Response to Reviewer 2 Comments
Point 1: There are noticeable seams in the images in Figures 15 and 16, and the gray-scale of the images appears to be different. Please explain this issue and propose a solution to rectify it.
Response 1: The main reason for the seam in the image is that there are some differences in the photoelectric response performance and quantum conversion efficiency of the 8 SCOMS detectors, which are more pronounced when obtaining low illumination images. These differences result in differences in the grayscale of the 8 images, and result in more noticeable seams. The improvement method is to calibrate the response outputs of 8 SCOMS in the laboratory and configure parameters to maintain consistent grayscale of the output images.
Point 2: It would be beneficial to include an explanation of the relationship between image gray values and SNR in Chapter 4.
Response 2: The signal-to-noise ratio (SNR) of a low-light imaging system is defined as the ratio of the system's output signal to the output noise, and the specific calculation formula has been provided in Chapter 2. The SNR of the system can be calculated by collecting multiple frames of integrating sphere image data in the laboratory, extracting the grayscale values of signals and noise. The ratio of noise to signal grayscale value is the SNR of the system.
Point 3: Consider bolding the explanatory text in the images to ensure clarity. This should be done for Figures 1, 4, 5, 6, 8, 9, 10, and 11.
Response 3: Revised according to suggestions.
Point 4: Check the citation format of Figure 3 and Figure 16 (Line 157, 349).
Response 4: Revised according to suggestions.
Point 5: Use "km" instead of "Km" to ensure uniformity in the units of measurement in the manuscript (Table 3).
Response 5: Revised according to suggestions.
Point 6: Please review and correct any minor typos, such as those on line 6.
Response 6: Revised according to suggestions.
Point 7: Verify the accuracy of your references and ensure that they adhere to citation standards.
Response 7: Revised according to suggestions.
Author Response File: Author Response.docx
Reviewer 3 Report
Comments and Suggestions for AuthorsThe author designed and tested a camera that can achieve long-range, high-resolution, and ultra wide field of view imaging capabilities in low light environments. Through laboratory and flight testing, it has been proven that this system can meet the requirements of low illumination environments of 0.01 lux, and its method and design were successful.I have just a few suggestions for the authors.
1. Please explain the basis for selecting a 13×13 μm pixel size in Chapter 2.3.
2. The figures are a bit blurry. Please consider replacing them with clearer ones.( Figures 3, 8,9,10)
3. In section 2.3.2 of the manuscript, extending the exposure time will result in forward image motion caused by flight. However, the conclusion from Table 4 is that "exposure time can be further increased to ensure excellent image fidelity." These two conclusions seem somewhat contradictory.
4. Coordinate system explanations should be added in Figures.9 and Figures.10.
5. Pay attention to the units after the numbers and remove spaces.
Comments on the Quality of English LanguageThe author of this article possesses good English language skills, which enable their innovative ideas to be clearly presented to the readers.
Author Response
Response to Reviewer 3 Comments
Point 1: Please explain the basis for selecting a 13×13 μm pixel size in Chapter 2.3.
Response 1: Generally speaking, the larger the pixel size of SCOMS detectors, the more photons they receive, and the stronger their imaging ability in low-light environments. We compared commonly used detectors with pixel sizes of 6.5um, 11um, and 13um, considering image resolution, frame rate, quantum efficiency, dark current, and procurement channels, and ultimately chose this SCOMS detector with a pixel size of 13um.
Point 2: The figures are a bit blurry. Please consider replacing them with clearer ones.( Figures 3, 8,9,10).
Response 2: Revised according to suggestions.
Point 3: In section 2.3.2 of the manuscript, extending the exposure time will result in forward image motion caused by flight. However, the conclusion from Table 4 is that "exposure time can be further increased to ensure excellent image fidelity." These two conclusions seem somewhat contradictory.
Response 3: Excessive exposure time can result in forward image motion caused by the flight, and the forward image motion needs to be less than half the pixel size to ensure clear imaging of the ground. Therefore, Formula 2 in the manuscript is the longest exposure time calculated by forward image motion less than 1/2 pixel size. By using this formula, the exposure time for oblique forward viewing can be greater than that for vertical downward viewing. These two conclusions are not contradictory.
Point 4: Coordinate system explanations should be added in Figures.9 and Figures.10.
Response 4: Revised according to suggestions.
Point 5: Pay attention to the units after the numbers and remove spaces.
Response 5: Revised according to suggestions.
Author Response File: Author Response.docx
