Missing Plant Detection in Vineyards Using UAV Angled RGB Imagery Acquired in Dormant Period

Round 1

Reviewer 1 Report

For a research paper, in my opinion, more experiments needs to be conducted.  Especially during the growing season of different vines. Because, for different growing period, the background of the vineyard and the growth of branches and leaves are different, and this is the difficulty of this technology. Please take it into consideration.

NA

Author Response

Dear Reviewer,

in attachment you can find the response to your comments.

Best regards,

Riccardo Dainelli

Author Response File: Author Response.docx

Reviewer 2 Report

Twelve out of forty-six links are self-citations, which is 21.74%. In addition, there are nine references to the publications of researchers with whom the authors have joint publications, that is, this also applies to the ethical violation, which consists in the mutual promotion of citation of closely collaborating persons. Total, ethically unfounded links 45.65%. This is an unacceptably high rate.

These links include links under the numbers: 1, 3, 4, 6, 8,9, 10, 11, 12, 15, 20, 25, 29, 30, 31, 33, 36, 37, 38, 44, 45.

References seem unfounded: [1,2] [3,4], [4–9], [10], [11,12], [13–19], [25–28], [20–24], [ 29.30]; [8,10,31–34], [15].

In the presentation, there are no references to publications under the numbers [43], [44], [45], in the presentation, after the reference to the article [42], there is a direct link to the article [46]. Apparently, the authors decided to provide additional citation of these articles, and inserted them into the bibliographic list, but did not give a link to them in the text. It is quite obvious that in this case you should not have inserted these references. In general, all references to their own publications and to publications of their co-authors in other general publications are completely redundant, they are not required for the presentation of materials, except, perhaps, references under the numbers [36], [37], [38], [44], [ 45]. With regard to these five publications, the question is doubtful, it is necessary to understand. But at this stage of reviewing the article, when it turns out that there is confusion with all the references in general, this question of the need for these five references can not be considered for the time being.

Section 4.3 cites [26] with the first author of "Primicerio", but this author is the first in the publication under the number [42] and not under the number [26].

Further in the same section, the authors refer to their own publication [20], but the section is called “Evaluation of results vs other studies in literature”! How so? Do the authors compare their results with their own results, and call this comparison with the results of other authors? Thus, they deceive readers!

Both of these cases should be classified as "inaccurately provided information in the article." The following is a reference to De Castro et al. [27]”, but judging by the name of the first author, this should be a reference to the publication under the number [29].

Such confusion is everywhere, for example, "Jurado et al. [29]" - should be [45]. Su et al. [13]” – should be [41]. "Padua et al. [28]" - should be [36]. "Di Gennaro and Matese No. [19]" - should be [33], or 37, or 38, and again these are articles by the authors of this article themselves, so they should not be cited as the results of other authors! "Hajar et al. [31]" - should be [46]

 In other words, this article is written so sloppy, carelessly, with disregard for all ethical standards and with deception of readers, since the authors give references to their own results for world achievements, both in the staging of the article and in the section entitled so that it should comparisons are made with the results of other researchers, and not the authors themselves. Indeed, the fact that the authors obtained better results than they themselves had obtained before does not yet prove that these results are significant and deserve a separate publication. Otherwise, every small personal success could be passed off as a global achievement.

Regarding the results. The correct definition of 797 pillars out of 806 pillars gives an indicator of 98.88%. It is not clear where two more estimates came from, which are higher, namely: 99.50%, and 99.19%?

It goes on to say that the best accu- racy was 90.14% and 82.06% across aggregated metrics, which is significantly worse than previously reported values.

After all, apparently, 9 missed vines out of 806 are not such a significant error, but if we are talking about an error of about 18% (see the indicator of 82.06% accuracy), then we are already talking about 145 missed vines. This is a very substantial loss of potential yield, and in this case such a method cannot be applied at all due to its inefficiency. It is possible that the authors themselves understand what indicators are being discussed and how acceptable the indicated errors are, but they did not write the article in order to read it themselves, but so that readers understand what is at stake and how useful and published results are reliable.

In general, the article is submitted to a journal called "Drones", it is expected that it reports something from the field of drone control, perhaps from the field of processing photos and video images received by drones, but the article does not discuss the reasons for which these errors and how the drone program can be changed to reduce these errors to negligible values.

After all, it is completely incomprehensible whether these errors can be eliminated due to an additional flight, or are they fundamentally unremovable, since the drone camera is basically unable to distinguish the absence of a vine in those places where the system gave erroneous results? That is, the article does not disclose methodological approaches and does not suggest ways to improve the results.

The illustrative material is very poor in the sense that the photographs of the vineyards with labels are not explained and therefore provide little in the sense of understanding image processing algorithms, and the two flowcharts presented (Figure 2 and the bottom of Figure 4) offer the most general information, in particular , for example, an indication that there are four signal processing methods.

One can only guess that if more information is used in methods 1 and 4, then they can give more reliable results. But in this case, it is not clear why it is necessary to consider method 2 and method 3 at all? Maybe it would be better to describe and compare method 1 and method 4 in more detail, or even suggest synthesizing a new method from them, for example, because you can process images using both of these methods and compare the results, if they do not match, then these points should be considered with extra attention, and either complicate the processing of these points, or send a drone for new information? All these are hypotheses that the reader is forced to put forward, since the presentation in the article is not clear enough.

In the presented form, the article cannot be published in any way.

Author Response

Dear Reviewer,

in attachment you can find the response to your comments.

Best regards,

Riccardo Dainelli

Author Response File: Author Response.docx

Reviewer 3 Report

Dear Editor,

the manuscript describes some interesting approaches used to count vines and identify missing vines using point clouds generated from RGB images. The article is well organized and reports interesting results and, in my opinion, requires only a few minor revisions to be acceptable for publication.

Author Response

Dear Reviewer,

in attachment you can find the response to your comments.

Best regards,

Riccardo Dainelli

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Does canopy leaves have an effect on plant detection during the vine flourishing period?

Author Response

Dear Reviewer,
in the attachment, you can find our reply to your comment.
Best regards,
Riccardo Dainelli

Author Response File: Author Response.docx

Reviewer 2 Report

The revised paper is OK.

Author Response

Dear Reviewer,
in the attachment, you can find our reply to your comment.
Best regards,
Riccardo Dainelli

Author Response File: Author Response.docx