Review Reports

Round 1

Reviewer 1 Report

I appreciated the effort made by the authors in order to improve the quality of the manuscript. Some points remains to be improved :

1) Develop the arguments in the introduction because in my opinion it still sounds a descriptive. It could have some more strong arguments concerning the advantage and drawbacks of the different methods.

2) the legends are still too short for some figures. The authors modified only one.

3) Why the previous figure 7 was deleted ?

4) the discussion should refer to previous studies.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

The comments/suggestions made by this reviewer were fulfilled successfully. Paper deserves publication.

Author Response

Thank you very much for your valuable suggestions.

Reviewer 3 Report

The authors added some elements to enrich manuscript and I think this paper can now be accepted for publication.

Author Response

Thank you very much for your valuable suggestions.

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

The article deals with the presentation of a method to estimate wheat scab occurring on ear. This method is based mainly on image analysis and learning approaches. The method is compared with different other methods among which there is a traditional one.

Broad comments

The topic of this article is clearly in the scope of remote sensing. Although, the study seems to show an interested method, the way it is presented makes difficult the comprehension also because there are some English language problems. So the manuscript is not so easy to read and especially the discussion is not enough deep with very few references.

In its current form, I would not recommend this study for publication in Remote Sensing Journal because it looks as a working draft that is a good base to be improved. A point should be improved concerning the use and the signification of the acronyms which are numerous but not always explicated. In addition, most of the legends are too short and not enough explanatory.

 

Specific comments

Introduction:

I find that the introduction is very descriptive but not with enough argued. The references are numerous which is good point but they could be more discussed also in the discussion part.

L69 : in references [11], the first names of the authors are used as names in the list of reference.

 

Materials and methods:

L104 : the effect of the weather conditions could be discussed. For instance what is the limit between a fine weather and a bad weather ?

L118 : I did not understand the status of these 1600 images comparing to the 120 images.

L130 : “Unet” is not defined ?

L142 : learning rate, steps_per_epoch and epoch are not defined ?

L152 : ReLU is not defined

L156 : the legends are too short !!!

L163 : rephrase this sentence.

L218 : definition of the target image

L222 : what do you mean about location update equation ? Which equation ?

L268 : the description should be in the materials and methods part.

L290 : the maximum of the Y axis should be 1 instead of 1.02

L291 : it could be interesting to present the standard deviation

L314 : this sentence is not clear

L332-335 : it is not a paragraph of discussion but rather presentation of results.

L354 : suppress the “that that” repetition.

Reviewer 2 Report

See the attached file.

Comments for author File: Comments.pdf

Reviewer 3 Report

Authors proposed a new segmentation-based algorithm named IABC-K-PCNN to identify six classes of severity of wheat scab in China´s Jianghuai River Basin. The performance was evaluated by different evaluation indicators. English writing is fine, methods are described in detail. Nevertheless, the following issues need clarification:

Title is misleading. The paper does not present what is the severity of wheat scab in the study area. Instead, it presents a new optimized approach to identify wheat scab. L51-57. Authors state that current method to identify wheat scab visually in the field is time consuming and that the disease degree cannot be judged accurately. Next, image processing technology is highly effective in disease identification and low in cost. I am not sure how IABC-K-PCNN approach can be less time consuming than the visual method. Farmers need more information other than presence/absence of scab to take quick actions to control the disease? How higher number of severity classes can be useful for farmers? Convince me! Regarding the accuracy of the segmentation algorithm, what was the ground truth data? This information is missing in the text. Other minor comments are:

L47: Citations should be numbered in sequence.

L101: Please, check the dates of the experiment.

L313-314: Please, check for mistyping.

Reviewer 4 Report

Dear Authors,
Although your manuscript is, in general, an interesting contribution for the remote sensing community, the advantages of your study were not sufficiently appealed and a restructuring might be required in order to improve the scientific level.

L.46
What kind of bacteria can cause scab? Please show us more details on that disease.

L.65
Could you clarify 'relatively complex environments'?

LL.83-90
Although you only mentioned about ABC algorithm, there are vast number of hyperparameter optimization methods.
You should review some of them.

1. Introduction
Why did you use imagery from a digital camera? What kind of data have been used for this purpose? You need to review data types in this section.

L.100
Did you conduct this experiment in 2018 or 2019?

LL.102-103
Could you show us the photographic conditions such as focal length and exposure time?

LL.141-142
Why did you use these setting? Didn't you use dropout?

LL.264-266
Did you conduct a pixel-based evaluation?

4 Discussion
You said that the experimental period was from April 26 to May 13. Did you confirm the differences in accuracies between flowering and filling period?