Machine Learning Techniques for Estimating Soil Moisture from Smartphone Captured Images
Round 1
Reviewer 1 Report
I have the following concerns.
1. It is necessary to clarify the statement of the problem regarding the use of ML.
2. Pay attention to the ratio and size of training and test and validation samples. Nothing is said about this.
3. Comparison of the accuracy of different ML models should be carried out under the same experimental conditions, namely distance from the ground, lighting angle, etc.
4. I do not see the scientific novelty of the article. Without it, it is already clear that the use of ML for images obtained from a smartphone to determine soil moisture will be faster and cheaper.
5. Why bring images from satellites when the technology for determining soil moisture based on a series of images obtained from a smartphone is proposed.
6. It is necessary to show not only the advantages but also the limitations of the proposed technology.
7. The development of a smartphone application will have practical application in the future.
8. The article lacks a comparison of the accuracy of determining soil moisture using a smartphone and ML with standard equipment.
Author Response
Dear Sir / Madam,
Thank you for allowing a resubmission of our manuscript, with an opportunity to address the reviewers’ comments. We are uploading (a) our point-by-point response to the comments (attached) (response to reviewers).
Best regards,
Muhammad Riaz Hasib Hossain et al.
Author Response File: Author Response.docx
Reviewer 2 Report
The manuscript presents very interesting and useful results on the estimation of soil moisture using images and machine learning techniques.
line 10: Please indicate why precise soil moisture (SM) assessment is essential in agriculture.
line 15: Four groups should be indicated.
The novelty of the study should be indicated in more detail in the Abstract and Introduction.
Keywords should be more informative.
Table 3: Did the different number of total soil samples and collection date affect the results?
The resolution of Figure 2 is too low.
lines 242-243: What were the differences between iPhone 242 11 Pro and an iPhone 6s? Only the resolution of the soil images?
Considering the obtained results presented in Tables 7-14, it is not fully understood why these 3 metrics were selected as classification accuracy metrics and the authors did not include others. The obtained values of R2 are not enough explained and discussed.
The discussion of the results obtained against the background of the available literature is insufficient.
Directions for further research should be indicated in more detail.
Author Response
Dear Sir / Madam,
Thank you for allowing a resubmission of our manuscript, with an opportunity to address the reviewers’ comments. We are uploading (a) our point-by-point response to the comments (attached) (response to reviewers).
Best regards,
Muhammad Riaz Hasib Hossain et al.
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
I am almost satisfied with the responses to my comments.
Author Response
Dear Sir / Madam,
Thank you for your comments. However, if you need any modifications to our manuscript, please let me know.
Best regards,
Muhammad Riaz Hasib Hossain et al.
Reviewer 2 Report
line 19: "four groups (Group 01, Group 02, Group 03, and Group 04)" - It is not informative
Author Response
Dear Sir / Madam,
Thank you for allowing a resubmission of our manuscript, with an opportunity to address the reviewers’ comments. We are uploading our point-by-point response to the comments (attached).
Best regards,
Muhammad Riaz Hasib Hossain et al.
Author Response File: Author Response.docx