Detection of Standing Dead Trees after Pine Wilt Disease Outbreak with Airborne Remote Sensing Imagery by Multi-Scale Spatial Attention Deep Learning and Gaussian Kernel Approach

Round 1

Reviewer 1 Report

In the manuscript titled " Detection of standing dead trees after Pine Wilt Disease outbreaks with airborne remote sensing imagery by multi-scale spatial attention deep learning and Gaussian kernel approach ", the authors have proposed a new detection method called multi-scale spatial supervision convolutional network to identify standing dead trees in a wide range of complex scenes based on airborne remote sensing imagery. The study justifies the utility of the state-of-art earth observation and deep learning models, particularly for ecological applications in the field of phytopathology.

Merits:
The authors proposed a new multi-scale spatial supervision convolutional network technique to identify standing dead trees in a wide range of complex scenes based on airborne remote sensing imagery.
The authors tested an oversampling method for the augmentation of undersized samples in CNN training to solve the imbalance and small PWD samples in their study area.
In comparison with traditional machine learning algorithms and the decision tree algorithm, this study used a deep learning network to establish the pine wilt disease dead tree detection model, made some improvements based on the original network and achieved an accurate model and better detection effect.

Weakness:
Authors claimed that future research could investigate the characteristics of the canopy at different diseased stages. Pl. elaborate. The authors also mentioned that the severity of disease outbreaks in the studied region had an effect on detection accuracy, which can be minimized to get better accuracy. The Authors should try to maximize accuracy.
Authors utilised the two-dimensional spatial Gaussian kernel. The kernels allow very flexible hypotheses, therefore one needs to choose kernel parameters.
  English language:
Line 74-75: Please recheck and verify once.
Line 83: Please add spatial resolution for Worldview imagery. Also, mention whether it is Worldview-1 or 2 or 3?
The manuscript must be checked for grammatical and other English-related errors.
The references should be checked according to the journal standards.
There are grammatical, uppercase/lowercase, punctuation, and sentence structure errors. The authors are suggested to give a thorough check for English writing.

Author Response

Dear Reviewer,

    Thank you for the positive evaluation of this paper. We have carefully checked the manuscript and tried our best to revise the grammar errors. We agreed on the weaknees comments on our manuscript. I think It is our future work to maximize accurcy by considering the forest type and Gaussian kernel parameter. The future research will evolve with this research.  Thanks again for your comments.

Best wishes,
Dian Yuanyong

Reviewer 2 Report

The manuscript has proposed a novel method to detect the standing dead trees caused by PWD which is the most destructive diseases around the world. Then, it discussed the precision, recall, and F1-score of their proposed method compared with the existing methods such as FCN8 and U-Net. As the manuscript has the originality and the novelty, the significance of content, the high quality of presentation as well as scientific soundness, it is recommended to be published after minor revisions with comments below:

Table 1: "hm2" should be "ha".

L245: "Figure 5" should be "Figure 4"?

L298: "Figure 7" should be "Figure 5".

L320: "Figure 11" should be "Figure 7".

Author Response

Dear Reviewer,

    Thank you for the positive evaluation of this paper. We have carefully checked the manuscript and tried our best to revise the grammar errors. The response information was listed in attachment.

Best wishes,
Dian Yuanyong

Author Response File: Author Response.docx

Reviewer 3 Report

Good for the publication in the present form

Author Response

Dear Reviewer,

Thank you for the positive evaluation of this paper. We have carefully checked the manuscript and tried our best to revise the grammar errors. 

Best wishes,
Dian Yuanyong