Underground Water Level Prediction in Remote Sensing Images Using Improved Hydro Index Value with Ensemble Classifier

Round 1

Reviewer 1 Report (Previous Reviewer 1)

Comments to the authors

The current version of the manuscript presents important changes that strengthen the first version. We thank the authors for the changes made.

After the review carried out, I consider that the weakness persists in the presentation of results in figures 3 to 7, where the information interpreted from the numerical analysis is represented. It is recommended to add some color palette, filters or combination of bands or polarizations, as the case may be, to improve visualization. The results should be presented adequately for readers interested in the topic. Add more interpretation to the figures.

The authors provided important information on the sensors used, however the relationship/use of the passive sensor images used to calculate the indices with the active sensor information has not yet been established. For example, figures 3 to 7, what type of images do they belong to? Are passive images used to obtain indices and statistics, what do you get from active images? In both cases, what was the individual/combined contribution in the mathematical analyses? What were the main advantages and disadvantages found?

Performance Analysis Subsection (3.2). I think there is an imbalance between the text and graphics and Table 1. I think there are a lot of figures with a lot of information (I suggest removing them). Table 1 is more in line with the information in the text.

Author Response

Reviewer 1:

The current version of the manuscript presents important changes that strengthen the first version. We thank the authors for the changes made.

1. After the review carried out, I consider that the weakness persists in the presentation of results in figures 3 to 7, where the information interpreted from the numerical analysis is represented. It is recommended to add some color palette, filters or combination of bands or polarizations, as the case may be, to improve visualization. The results should be presented adequately for readers interested in the topic. Add more interpretation to the figures.

Response: Thank you for your important comments. As per the reviewer’s comment, we have updated the visualization and presentation of results in figures 3 to 7 at section 4.2.

 

2. The authors provided important information on the sensors used, however the relationship/use of the passive sensor images used to calculate the indices with the active sensor information has not yet been established. For example, figures 3 to 7, what type of images do they belong to? Are passive images used to obtain indices and statistics, what do you get from active images? In both cases, what was the individual/combined contribution in the mathematical analyses? What were the main advantages and disadvantages found?

Response: Thank you for your useful comments. From the figures 3 to 7, we have used passive type sensors, those information’s are clearly described at section 3.3.

This research works on various aspects with active and passive sensors. Active sensors can operate in a wide range of irradiation conditions but are restricted to a smaller range of wavelengths depending on the type and quantity of light sources. Unlike active sensors, which have their own light-emitting devices, passive sensors rely on sunlight as their light source. The ability of active and passive sensing systems to evaluate pertinent agronomic and physiological characteristics still needs to be understood. These types of images are belonging to passive remote sensing systems and those passive sensors are used to obtain indices and statistics. The advantage of passive sensors is that they depend on the Sun's light to illuminate the target; as a result, they don't need their own energy source, making them simpler machinery. However, it is not suitable for low light levels. Those information’s are acquired from https://earthexplorer.usgs.gov/. The proposed ensemble model is a combination of NN, SVM, and Improved DCNN. Initially, the total extracted 3290 HI features are subjected to the NN and SVM, and the predicted results from them are subjected to Improved DCNN to determine the final result.

The above-statement is updated at section 3.3.

3. Performance Analysis Subsection (3.2). I think there is an imbalance between the text and graphics and Table 1. I think there are a lot of figures with a lot of information (I suggest removing them). Table 1 is more in line with the information in the text.

Response: Thank you for your valuable comments. As per the reviewer’s suggestion, we have eliminated the graphical illustration (Figure 8, 9, 10) at section 4.3 (Performance Analysis).

 

Reviewer 2 Report (Previous Reviewer 2)

I did not find the validation of true underground water level with the image retrieved water level yet. Otherwise, authors improved the manuscript than previous one.

Author Response

Reviewer 2:

I did not find the validation of true underground water level with the image retrieved water level yet. Otherwise, authors improved the manuscript than previous one.

 

Response: Thank you for your useful comments. From the below mentioned links, the validation of true underground water level with the image retrieved water level are carried out in this research study.

http://cgwb.gov.in/ground-water/gw%20monitoring%20report_january%202016.pdf

https://pubs.usgs.gov/circ/circ1186/html/gen_facts.html

https://pib.gov.in/PressReleasePage.aspx?PRID=1742815

https://www.deccanherald.com/opinion/in-perspective/groundwater-reaching-dangerous-levels-816761.html

https://prsindia.org/files/policy/policy_analytical_reports/1455682937--Overview%20of%20Ground%20Water%20in%20India_0.pdf

http://cgwb.gov.in/Ground-Water/GW%20YEAR%20BOOK%202019-20%20ALL%20INDIA%20FINAL%20752021%20(1).pdf

 

Round 2

Reviewer 1 Report (Previous Reviewer 1)

Comments to authors.

The response of the authors is appreciated.

I consider that the answers to questions 1 and 2 are not satisfactory, since they do not respond to what was requested.

Required in question 1: The necessary modifications were not made in figures 3 to 7 to improve the interpretation of the results. It is not possible to relate the information in the text and figures.

Required in question 2: The usefulness of the products generated with satellite images of active sensors and their usefulness in the research, as well as the type of products generated and their procedure are not clearly explained. Nor is it indicated how they were integrated into the methodological process or what results were obtained.

Author Response

Reviewer 1:

The response of the authors is appreciated. 

I consider that the answers to questions 1 and 2 are not satisfactory, since they do not respond to what was requested.

Required in question 1: The necessary modifications were not made in figures 3 to 7 to improve the interpretation of the results. It is not possible to relate the information in the text and figures.

Response: Thank you for your useful suggestion.

As per the reviewer’s comment, we have provided the interpretation of the results (figure 3 to 7) at section 4.2.

Figure 3 and 4 shows the representation of underground water level images at critical and highly exploited conditions. Where, the pixel resolution level is poor and unable to gather more information from the particular image. While taking figure 5, Images of region is processed with Safe water level, where we can get huge volume of information with higher accuracy from the particular image. Figure 6 shows the representation of underground water level images with salinity conditions. Due to the salinity process, unable to get the clear portions from the image. Figure 7 shows the representation of underground water level images at semi-critical conditions, where we can get average data with acceptable accuracy.The location and its evaluation ranges of all the specified images are declared in the beginning of section 4.2.

Our main goal is to predict (classify) the quality, and results of the classification are described in section 4.3.

Required in question 2: The usefulness of the products generated with satellite images of active sensors and their usefulness in the research, as well as the type of products generated and their procedure are not clearly explained. Nor is it indicated how they were integrated into the methodological process or what results were obtained.

Response: Thank you for your valuable comments.

With due respect, I would like to substantiate that the present research is on exploiting the listed imagery, but not on acquisition procedures and methodologies of the listed imagery. 

The usefulness of the products generated with satellite images of active sensors and their usefulness in the research, as well as the type of products generated and their procedure are not clearly explained. 

The usefulness of these imagery in our research is to categorize the present underground water level. The usefulness can be extended to image processing research such as filtering, reconstruction, enhancement, and many more, and applications such as agriculture, soil fertility prediction, etc. However, they are all not our research interests. Hence, we have limited our explanation up to, how we have adopted our methodology in this imagery, rather than how these images are acquired, because the acquisition procedure can be rightfully explained by the researchers who acquired these images. 

Nor is it indicated how they were integrated into the methodological process or what results were obtained.

As our research is on exploiting the images to predict the underground water level, we have provided all the performance analysis in terms of predicting the water level using the images. You are humbly requested to refer Table 1-3. 

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Comments to the authors

The interest in the subject is appreciated. A great effort is observed in the manuscript. However, there are several elements that should be considered, as they limit the importance and contribution of the research.

Best of luck.

Some considerations to the manuscript are:

A.- It is recommended to improve the wording in all sections. In addition, it is important to reduce the literature review section, since it is long and generates disinterest. Less is more.

B.- The methods section presents an important scientific weakness in the development of the different sections. At times it is closer to a technical manual than a scientific article. This limits the contribution and understanding of the results and conclusions.

C.- The authors mention the use of satellite images, but do not specify the type of sensor used and the importance of the biophysical variables used to calculate the predictive model. It is important to detail the particular characteristics (resolutions or polarization if active), since the advantages or disadvantages of the information provided by the images are related to their characteristics.

D.- It is not specified what type of additive noise is sought to be eliminated by the Wiene filter, and at what level of image processing said filter was applied.

E.- In section 3.1. DFT preprocessing is applied. This means that a long time series of satellite images (data cube) was used. It is important to specify.

F.- The importance of the images of active and passive sensors in the study is not established. What is your individual and combined contribution?

G.- There is a great numerical contribution to analyze the satellite information, however, sub-section 3.2 (indices and statistics) is weak, which limits understanding the efficiency of the indicators within the predictive model.

 

H.- It is recommended to deeply improve the sections of results and discussion and conclusions. Figures 3 to 7 presented in the text should improve in quality and presentation. In its current state, it is difficult to observe its contribution.

I.- The subsections of performance analysis, error statistics and analysis of characteristics are difficult to understand. It is recommended to improve the wording.

Reviewer 2 Report

I am happy to review the manuscript, but expecting the ground water level output with some ground validation. Anyway, I have some comments in the following issues:

1.       Figures 3-7 do not really understandable about the ground water level, those are showing some improvement of visualization after filtering. Detail explanations are needed in the captions and marked the changes what happened actually after filtering.

2.       I did not see any ground water level information of the studied areas here, and need to validate with ground truth data. Authors just compare the performance of the method statistically, but really need ground validations of the improved method they develop or formulated.

3.       Data availability statement is missing here, it is needed to reproduce the output if anybody needed.

With these observations, I would like to recommend a major revision of the manuscript.

Comments for author File: Comments.pdf