Intelligent Identification of Pine Wilt Disease Infected Individual Trees Using UAV-Based Hyperspectral Imagery

Round 1

Reviewer 1 Report

This manuscript uses unmanned aerial vehicle hyperspectral remote sensing and proposes a method for diagnosing PWD diseases based on deep learning methods, which has certain innovation. Suggestions are as follows: 

1. It is recommended to expand the review of PWD monitoring in the introduction, including all methods and remote sensing platforms, analyze the existing problems and advantages of drone remote sensing, and propose specific issues targeted in this article.

2. L67-70，How did the manuscript address the issue mentioned in the introduction? In the method section I did not see how to distinguish PWD from other diseases.

3. Mask R-CNN and Integrated Framework Combing Prototypical Network Classification Model，The relationship between the two methods you proposed? Are the two solving a problem together or separately?

4. One more conclusion should be summarized from the comparison between the two proposed methods.

Author Response

Dear Reviewer, Thanks very much for taking your time to review this manuscript. We really appreciate all your comments and suggestions! 

Author Response File: Author Response.docx

Reviewer 2 Report

The Manuscript: "Intelligent Identification of Pine Wilt Disease Infected Individual Trees Using UAV-Based Hyperspectral Imagery" (Manuscript ID: remotesensing-2425773) aim to enhance the detection of Pine Wilt Disease (PWD), caused by an invasive alien species, the Pine Wood Nematode. This is a significant issue in China, resulting in considerable economic and ecological impact. The authors use UAV-based hyperspectral imagery and deep learning techniques to identify infected trees at different stages of PWD. Two methods are explored: an improved Mask R-CNN instance segmentation method and an integrated approach, showing good performance for both methods in PWD identification.

The topic is interesting and fits within the scope of the journal, but the manuscript requires substantial revisions for clarity, justification of the methods, and improved presentation of the results.

The introduction could be significantly strengthened by providing a stronger justification for the choice of UAV-based remote sensing over other sources such as satellites. Additionally, the reasons for choosing hyperspectral data over other types of data (thermal, ms, lidar, etc.) should be clarified. The authors should provide more context by referencing other studies that have used UAVs and imagery (not only hyperspectral) for disease detection in trees. Then, the authors could emphasize the benefits of hyperspectral technology for disease detection. Some examples that the authors might find it useful to review for structure ideas, and/or include them to enrich their manuscript (this advice is not mandatory, the authors can consult other papers):

·         “Exploring the Potential of UAV-Based Hyperspectral Imagery on Pine Wilt Disease Detection: Influence of Spatio-Temporal Scales.” https://0-doi-org.brum.beds.ac.uk/10.3390/rs15092281

·         “UAV-Based Forest Health Monitoring: A Systematic Review.” https://0-doi-org.brum.beds.ac.uk/10.3390/rs14133205

·         “Mapping the Spatial Variability of Botrytis Bunch Rot Risk in Vineyards Using UAV Multispectral Imagery.” https://0-doi-org.brum.beds.ac.uk/10.1016/j.eja.2022.126691

·         “Comparison of Canopy Shape and Vegetation Indices of Citrus Trees Derived from UAV Multispectral Images for Characterization of Citrus Greening Disease.” https://0-doi-org.brum.beds.ac.uk/10.3390/rs12244122

Regarding the Methodology, in Figure 1, it is noted that 4 of the 6 sample sites are outside the Region of Interest (ROI) for the hyperspectral flight, which could potentially affect the accuracy and relevance of the results. The process of generating the UAV hyperspectral orthomosaic could be made clearer to enhance the reproducibility of the study. For instance, the meaning of the "corresponding image element" and the "brightness value of the scattered light" in equation 1 (line 130) are not properly explained. Additionally, the authors should provide information on how they employed ground control points (gcps) for accurate georeferencing of the UAV hyperspectral orthomosaic. The workflow needs to be explained more effectively, especially the "parameter optimization" mentioned in Figure 4. While this step is crucial, there is no subsequent reference or explanation in the text.

While the results are good, an insufficiently in-depth discussion undermined them. A comprehensive discussion that contextualizes the results within the broader scientific literature would substantially enhance the document. Currently, the manuscript seems disconnected and doesn't discuss how its results relate to other papers. Therefore, the discussion should be enhanced by comparing this study with others that have used UAV and imaging sensors to detect diseases in forests/trees/woody crops. This would help clarify the novelty and contribution of their research. Moreover, the authors should indicate what work lines are interesting for future works.

In general, the quality of the figures should be improved. For instance, the numbers in Figure 7 are challenging to read, which hampers comprehension of the content. In addition, some figures are not referenced in the text.

Specific comments

Line 56

“UAVs equipped with hyperspectral”

The authors should explain the reasoning behind selecting hyperspectral over other sources such as multispectral, lidar, or thermal options.

Line 109

“a total of six sample plots (25 m × 25 m) at the study site”

Are four of these sample plots located outside the ROI of the hyperspectral flight? Why? Please clarify.

Line 117

“DJI m600 pro UAV equipped with a Nano-Hyperspec”

Additional details about the manufacturers of each device/software should be provided thoughout the manuscript.

Line 119-121

“The flight parameters of the DJI m600 pro UAV are shown in Table 1 and the main parameters of the hyperspectral sensor are shown in Table 2.”

This sentence structure could be perceived as informal. Please consider restructuring and paraphrasing for a more formal tone.

Line 126

For completeness, Table 2 should include detailed sensor information such as Field of View and Focal Length.

Line 132

“?????? is the corresponding image element dark current data recorded by the sensor”

Could you provide more information on which element this refers to and the method used for recording this data?

Line 138

“less affected”

This statement is unclear without a reference point. Please specify what it is “less affected” compared to.

Line 142

“DGPS device”

Additional details about the DGPS device, including the manufacturer, would be helpful.

Line 153

“combined the ground survey and the spectral”

Could you please elaborate on how the ground survey and spectral data were integrated, given that most of the sample plots fell outside the ROI UAV flight area?

Line 350

The units for the “output size” should be specified.

Line 351

“eCognition”

Please provide information about the developer of eCognition.

Line 360

“using Python programming,”

Could you also mention any other software that was utilized during this research?

Line 414

Figure 10 is not referenced in the text.

Line 450

“Table 10. Overall accuracy evaluation table.”

Consider delete “table” and revising the caption to "Table 10. Overall Accuracy Evaluation" to avoid redundancy.

The authors should ensure consistency in their use of first-person and third-person perspectives throughout the text. Additionally, the manuscript would benefit from a language review to improve fluency and correct minor errors.

Author Response

Dear Reviewer, Thanks very much for taking your time to review this manuscript. We really appreciate all your comments and suggestions! 

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

(1)The issues raised in the introduction should be addressed in this manuscript, and those that cannot be addressed do not need to be raised.

(2)Clarify the logical relationship between methods.

(3)The conclusion section should be more clear, concise, and informative.

Author Response

Dear Reviewer, Thanks very much for taking your time to review this manuscript. We really appreciate all your comments and suggestions! 

Author Response File: Author Response.docx

Reviewer 2 Report

The authors have successfully incorporated my comments and suggestions, leading to a significant enhancement in the manuscript's quality. The paper's contribution towards disease screening is now more distinct and comprehensible. I believe the manuscript is ready for publication.

The manuscript is well written. However, I recommend a final revision in order to correct minor language issues. For example:
- In line 11, instead of "major invasive alien species," consider using "major invasive species." The term "alien" can be redundant and rare in this context.

- Line 525: The term "researches" should be "research".

- Line 534: Consider rewording "PWD early infected trees recognition accuracy" to "recognition accuracy of early PWD-infected trees". And do the same throughout the document.

Author Response

Dear Reviewer, Thanks very much for taking your time to review this manuscript. We really appreciate all your comments and suggestions! 

Author Response File: Author Response.docx