Assessing the Accuracy and Feasibility of Using Close-Range Photogrammetry to Measure Channelized Erosion with a Consumer-Grade Camera

Round 1

Reviewer 1 Report

The aim of this study was to use close photogrammetry and estimate the accuracies on detecting soil erosion channels under different spatial, temporal and resource constraints.

 

The key words are not adequate because they repeat words of title. I recommend gully erosion, rill erosion, close photogrammetry, simulated DEM surface.

 

From the results is not clear if there are time consuming advantages from using this method instead of manual measurements (2.15 hours versus 1 hour for 2 persons), or from aerial photos from drone.

Also, cost advantages over the aerial photogrammetry method are not clear since it was not tested and compared a drone of same cost.

In line 494 the authors do not justify why obstacle detection in drone technology, does not help the operation of a drone in the forest. Why is impossible? You should mention a reference or you should present data from drone that you used.

In line 495 I agree that the weather conditions may limit use of drones but you should present references of your own data about this statement.

In line 503 was stated that there is also a trade-off between the detection area and accuracy. Meanwhile the authors did not test any aerial photogrammetry to compare with the current method and prove statistically significant differences.

The discussion chapter in general does not compare with accuracy measured in other studies and express merely the opinion of the authors about the advantages of their method. In Table 6 was presented limitations to accuracy of the method from different angles but there was no comparison with accuracy from drone in same altitude, angle 90 and speed of data collection due to minimum need to walk in the field.

Most conclusions are not based on the data presented. How can you conclude that manual method is less accurate that close photogrammetry? or aerial photogrammetry?

The lack of a built-in GPS system of their method has disadvantages that the authors do not mention in conclusions.

In general, the authors fail to show a comparison to the aerial photogrammetry and close-range photogrammetry, thus they cannot conclude that their method has relatively quicker data acquisition, higher accessibility, mobility and lower cost. Meanwhile there are a lot of references presenting the opposite results of what this paper wants to conclude.

From what was presented it was clear only the weather advantage of the close photogrammetry, but I do not see any reason for someone to go to the field for measurements during strong wind and under heavy rain.

 

Author Response

Please see the attachment.

 

Author Response File: Author Response.docx

Reviewer 2 Report

Dear Editor,

I had a thorough review of the Manuscript RS-777484 “Assessing the accuracy and feasibility of using close- range photogrammetry to measure channelized erosion with a consumer-grade camera ".

The manuscript fits the Journal topics, provides a solid methodology and quantifies an important process for generating a point cloud. It is not regarded as innovative (it is a rather simple idea more like turning back to the basics), but it can be useful for several applications. However, the authors need to communicate better these applications to the readers and they need to clarify some other points (see below). The authors actually assess the overlap % and which shooting angles are the best for SfM (Structure from Motion). Currently we do this empirically, so now they provide the reasoning and the quantification. I think that this manuscript is useful and can be cited by future papers that collect data with SfM in the field, when they present their methodology part.

Overall I think that the manuscript can be accepted following moderate revision. Since there is no such choice on the menu I would recommend minor revision, however most of my comments are questions that relate to their methodology that need to be clarified, rather than major weaknesses.

Several weaknesses or things that the authors need to clarify:

1) In such short distances the authors used zoom lens. As a result, some distortion is expected. How this is corrected/dealt with? This needs to be clarified.

2) The authors need to clarify what they used as a scale for correcting the distances on the point cloud. In other words, how they improved the relative accuracy when building their model, before starting a comparison with the real measurements. Normally I guess they should have used fixed distances for calibration, but not the ones used for validation.

3) How the authors used the reference points? They refer in both techniques that they insert plastic target plates, but of course they did not use GPS and they do not clarify whether they measured the relative distances. So how they utilized them? They refer within the text that "These plastic targets can also be automatically detected by the software, facilitating fast data processing.". However, this can not be the main role/reason for the reference points.

4) The comparison with the UAV may sounds a little extreme. I would not expect anyone trying to use a UAV for research purposes at such scales. Anyway, it is questionable, on whether this part should remain or deleted, particularly when they come up with a 3cm error with their camera. For example, 3cm error in such short distances (less than 2m) is relatively high in my opinion (even if it lies within the 4% error boundary). If someone introduces a DGPS then it can even go to the mm scale. A UAV with DGPS can obtain 1cm accuracy even at 20 or 30m height. Finally, if the authors wish to make a point cloud comparison, then they must refer to t-LiDAR as well.

5) The authors need to explain where their methodology can be applied. This is currently missing from the manuscript (with the exeption of refs 1 and 2) and it is a major weakness. Obviously the main application concerns soil erosion and conservation planning. Therefore, they need to clearly communicate this. Soil erosion is a major issue and several models try to quantify and assess its impact such as the Rusle (Revised Universal Soil Loss Equation, Wischmeier and Smith, 1978, van der Knijff et al., 2000 used mostly in meadows and fields for grazing) and the PESERA  model (Pan-European Soil Erosion Risk Assessment, Kirkby et al.,2003 mostly in agricultural plots). This methodology (monitoring erosion) is particularly important for postfire settings. Now with climate change and higher vulnerability trends wildfires are a major issue that result in high soil erosion rates (Shakesby, 2011, Karamesouti et al. 2016). For example, average post-fire soil loss value, compared to pre-fire conditions, is about six to nine times greater (Karamesouti et al. 2016 Geoderma). This is another major application that requires urgent measures and this methodology can clearly assist by quantifying and monitoring such phenomena. So the authors should dedicate some space on the above remarks and why not expanding them to other possible solid applications. Another application may involve the measurements of road rutting, but they have to increase their resolution (e.g. Gézero and Antunes 2019 ISPRS)

6) References are relatively ΟΚ, but except the above remarks (in 5) a few more citations can be inserted (e.g.  Rieke-Zapp, D. H., & Nearing, M. A. (2005). Digital close range photogrammetry for measurement of soil erosion. The Photogrammetric Record, 20(109), 69–87.

Butler, J. B., Lane, S. N., & Chandler, J. H. (1998). Assessment of DEM quality for characterizing surface roughness using close range digital photogrammetry. Photogrammetric Record, 16(92), 271–291.

In addition, t-Lidar is another common technique dealing  with similar problems. However, this is not referred within the text, so a couple of references dealing with t-Lidar applications might be needed.

 

Yours sincerely

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Dear Authors,

I have read the manuscript remotesensing-777484 "Assessing the accuracy and feasibility of using close-range photogrammetry to measure channelized erosion with a consumer-grade camera" by Zheng et collab., submitted for publishing in Remote Sensing Journal. The paper is well written, and describes an experimental approach of application of low cost digital imagery in the study of micro-landforms. In my opinion, some parts must be rewritten, in order to enrich the scientific value of the paper. I found that some descriptive frames are repetitions and they can be seen in almost every part of the paper. In the introduction part, I suggest to describe the applicability of the method in a general framework of the erosional processes, especially to those induced by water (with reference at a global scale). Instead a too much enlarged ideas from 77-97 lines, the authors should address a SfM background in the introduction part. Even the manuscript will be smaller with 1 page, should try to avoid repetitions among different parts of the paper. But, generally, the manuscript has a good scientific value, with an impressive accuracy of the results (e. g. ... based on a 1 mm DEM accuracy). Other specific comments and suggestions can be found in the *.pdf file attached. I recommend the publication after minor revisions. Best regards

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 4 Report

I find the work topic very interesting and important both for science and practice.

In my opinion, this is a valuable analysis in the field of optimizing the use of digital cameras to reflect the land surface and its shape. The results obtained this way may be subject to many errors, as the authors pointed out. This gives many possibilities of using the results of this work to optimize techniques of field works. For this reason, you should consider rebuilding your work goals and descriptions relevant to them in the entire paper. Unnecessarily, the results of this work are confronted with work with drones. They can also be useful for drone operators.

The results of the work in a fact are about the parameters of optimal camera placement with aim to minimize errors generated during data acquisition.

I put next comments directly in the pdf file as a comments.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 5 Report

Dear Authors

I have read the manuscript entitled: Assessing the accuracy and feasibility of using close- range photogrammetry to measure channelized erosion with a consumer-grade camera with great interest and I have reviewed it carefully. I believe that it is also of great interest to the readers of Remote sensing MDPI journal and is within the aims of the Journal.

I think it is suitable for publication with few refinements.

The manuscript is quite complete, clear and concise. However, I think it lacks some contextualization with respect to some crucial debates in the geomorphological community (aerial imagery / UAV or remote sensing imagery as a source for the construction of high resolution DEM).

In my opinion a brief statement regarding the structure from motion methodology and performances has to be mentioned. The reader need a term of comparison of the methodology proposed.

Regarding a “cheap” application for landforms mapping using SFM, the author may have a look at Kropáček et al 2016 Assessment of gully erosion in the upper awash, central Ethiopian highlands based on a comparison of archived aerial photographs and very high resolution satellite images.

As I mentioned before about the second objective of the paper, the comparison in term of costs with aerial image (orthophotos) is missing and it would be highly appreciated by the readers.

Nothing to say regarding the experimental approach.

Line 469, why the authors mentioned snowy condition for a channel measurement? Is it possible to use the proposed method to measuring channel erosion while snowing?

Though I trust the authors, I ask the editor to check for plagiarism, I cannot check on my own.

Therefore, I wish you best of luck.

Kind regards

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 6 Report

Dear Authors,

Congrats for a paper well-written and for the applicability of your study. At this stage, I would recommend minor revisions, because there are some minor issues to be corrected.

You need to substantially increase the number of references (as you have only 37 references), to at least 50. There is plenty of literature to be cited nowadays.

L208: is there any reason why the camera was set on a height of 1.92 m above the ground? Please, explain if it is random or is there any reason for that.

References: journal's name should be abbreviated and the use of DOI is highly indicated.

 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The paper was significantly improved and now is more clear. In general I agree that the paper can be published as it is, but in line 530 is mentioned that UAV cannot be used under forest canopy because cannot fly in the woods. I disagree and references 56-59 are relatively old for this topic. Exist drones with sensors that can avoid obstacles automatically. They can be used in closed spaces with many obstacles like is the case of precision agriculture in greenhouses. Thus, I propose to change the 2 words "cannot" with "not recommended" or "difficult" (or something similar). 

In reference 1 should have the name of the journal and not the initials. 

In reference 6 is not Fla, but FL.

In reference 18 UAV is with capital letters, and the same for DEM in reference 19. 

Reference 18 should be cited as a conference with date and location.

Reference 21 is incomplete in location, country, editors.

Reference 24, 26, 33, 34 and others with journal name as initials. 

Reference 32:  Unmanned Aerial Vehicle (UAV) 

Reference 56: Earth Surface Processes and Landforms and New Zealand.

Author Response

Please see the attachment. 

Author Response File: Author Response.docx

Reviewer 4 Report

The paper has been substantially improved. It can be published in present form.

Author Response

We greatly appreciate the reviewer’s kind help.