Next Article in Journal
Multi-Species Individual Tree Segmentation and Identification Based on Improved Mask R-CNN and UAV Imagery in Mixed Forests
Previous Article in Journal
Complex Ecosystem Impact of Rapid Expansion of Industrial and Mining Land on the Tibetan Plateau
 
 
Article
Peer-Review Record

Stripe Noise Detection of High-Resolution Remote Sensing Images Using Deep Learning Method

Remote Sens. 2022, 14(4), 873; https://doi.org/10.3390/rs14040873
by Binbo Li 1,2, Ying Zhou 3, Donghai Xie 1,2,*, Lijuan Zheng 4, Yu Wu 5,6,7, Jiabao Yue 1,2 and Shaowei Jiang 1,2
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Remote Sens. 2022, 14(4), 873; https://doi.org/10.3390/rs14040873
Submission received: 20 January 2022 / Revised: 6 February 2022 / Accepted: 8 February 2022 / Published: 11 February 2022
(This article belongs to the Section AI Remote Sensing)

Round 1

Reviewer 1 Report

In my opinion, the authors have addressed properly the issues raised in the first draft.

Author Response

Thank you very much for taking the time out of your busy schedule to review our article, we have made the following changes according to your good suggestions, thank you again for your review.

Reviewer 2 Report

See attached file.

Comments for author File: Comments.pdf

Author Response

Thank you for reviewing our article in your busy schedule. We have made revisions to your suggestion, and the contents of each reply are in the attachment. Thanks again for your review.

Author Response File: Author Response.docx

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

This study, authors propose a new algorithm called LSND (linear stripe noise detection) for stripe noise detection. The problem is well introduced and the proposed solution is convincing and compared to the state of art other methods.

Reviewer 2 Report

Authors have proposed a DL method to stripe noise detection in high-resolution satellite imagery.

My main concerns with the paper are because of the method, and the tests are based on simulated stripe noise only. A mixture of data (simulated and real) should be tried at least. Without that researchers on this topic would feel the method is not ready yet for publication.

Reviewer 3 Report

see the attached file

Comments for author File: Comments.pdf

Reviewer 4 Report

The paper presents a deep-learning approach for detection of strips in remote sensing images. Some tips allowed the method have better performance and faster answers with respect to the state of the art.

REMARKS

>  the meaning of "linear" adjective needs to be disambigued. please provide a clear definition at the beginning of the method description

> Section 3: please assign a more explanatory title

> Pag 7, lines 187-191: please assign a different symbol to alpha and beta in the last and previous iterations

> lines 269-273: please use subscripts in the text as in the equations

> please provide a definition of the non-maximum suppression method

> please describe IOU, DIOU, CIOU indicators

> line 461: please check the number of images of the training set

 

Reviewer 5 Report

Summary

Stripe noise is a phenomenon that widely exists in space-borne imaging. Stripe noise seriously degrades image quality and adversely impacts on the subsequent extraction and use of the image information. This study proposed an algorithm called LSND (linear stripe noise detection) for stripe noise detection and train it using several types of high-resolution remote sensing images. The structure is logical, the figures are of good quality and the historical background has given credit as is appropriate. The references are adequate. However, the paper only applied the existing deep-learning methods and presented the result by these open-source DP codes already in the web. The innovation is much insufficient. Too many equations which are already in the public, too many repetitive introductions and results, this article should have a high repetition rate. Based on this, I think this manuscript should be rejected address the comments below.

Specific/Detailed Comments

  1. The Section 2 Related Work should be moved to appendix section.
  2. Simulated images with stripe noise should consider the different instruments’ characteristics. Hence the huge, simulated images for training should consider clarifying to some categories taken account for different instruments’ characteristics.
  3. The results only present the existing available public deep learning methods, not focused on the topic of this study on stripe noise. More works should focus on the specialties and difficulties when using these methods on stripe noise images processing.
  4. Too many equations which are already in the public, too many repetitive introductions, this article should have a high repetition rate.
  5. This study indeed repeats the past works, the only difference is study object became the stripe noise, hence the innovation is much insufficient, and the results are only repetitions from the existing public deep learning codes.
Back to TopTop