Next Article in Journal
Rapid Determination of Soil Horizons and Suborders Based on VIS-NIR-SWIR Spectroscopy and Machine Learning Models
Previous Article in Journal
Magnetic Fault-Tolerant Attitude Control with Dynamic Sensing for Remote Sensing CubeSats
 
 
Article
Peer-Review Record

The Retrieval of Forest and Grass Fractional Vegetation Coverage in Mountain Regions Based on Spatio-Temporal Transfer Learning

Remote Sens. 2023, 15(19), 4857; https://doi.org/10.3390/rs15194857
by Yuxuan Huang 1,2, Xiang Zhou 1, Tingting Lv 1,*, Zui Tao 1, Hongming Zhang 1, Ruoxi Li 1,2, Mingjian Zhai 1,2 and Houyu Liang 1,2
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Reviewer 3: Anonymous
Remote Sens. 2023, 15(19), 4857; https://doi.org/10.3390/rs15194857
Submission received: 14 August 2023 / Revised: 1 October 2023 / Accepted: 2 October 2023 / Published: 7 October 2023

Round 1

Reviewer 1 Report

 

This paper is well written and easy to read. The main question I have is regarding the use of random pixel in training and validation. Why not use a group of objects. In remote sensing, a map producing salt and pepper is not really useful in practice. What I mean is giving the FVC of a pixel being 0.9 next to another pixel of 0.01 is not of any importance.

 

The other comments I have are the following.

 

 

L102:

“The flowchart of the proposed approach is shown in Figure 1: (1)

Established a sample classification system based on the characteristics of mountain areas,

considering the type of surface features and the characteristics of altitude gradients. (2)

Based on the high-resolution remote sensing data, the FVC sample is established by using

the PROSAIL model. (3) Used FVC samples to train 1DCNN and LSTM models. (4) Model

transfer (fine-tuning) and obtained the distribution of FVC according to the model.” Needs to be rephrased. The verbs and style of the four points don’t align. It makes it hard to understand if this is an explanation of the chart or the contribution of the paper. The authors could use a style such as “… (1) we established a sample classification…., (2) Based on high resolution … we established the FVC samples by using….”

 

L114: Km2 should be km2

L123: RS should be spelled out as Remote sensing

L136: “obtained” should be “obtain”

 

L139:

“The DEM was resampled to a resolution consistent with the optical  imagery, and topographic slope and aspect were calculated”. It is not clear to me what you did. You resampled the DEM to 10 m? What about the target data that has a resolution of 16m? What did you do with it?

 

Please revise the citation style to be consistent throughout the text. For instance, on Line 203, there is a need for space before [50]. Same on Line 201.

 

L223: Replace “gat” by “gate”

L230: Rephrase “The LSTM layer consists of 5 layers of LSTM layers” by something like “The LSTM layer consists of 5 layers of LSTM units"

 

Sections 2.5.1 and 2.5.2 please change the reference to figure 4.a and 4.b. You reversed the references.

 

L238: “The high-precision pre-training samples were inputted to the model consist of more” remove “were” to make the sentence read well. As matter of fact, the whole paragraph needs to be rephrased.

 

 

 Throughout the document, R2 needs to be updated to R2

The language is great except a few minor comments

Author Response

Thanks for your suggestions. All revised details in the article have been uploaded, please check them.

Author Response File: Author Response.docx

Reviewer 2 Report

A brief summary

A review of the manuscript entitled: „Retrieval of mountain forest and grass fractional vegetation coverage based on spatio-temporal transfer learningis very interesting. The manuscript is coherent and well written. It contributes to enhance the knowledge on this topic.  Based on this general evaluation and the specific comments, reported below.  I recommend a minor revisions of the manuscript. I have few specific comments, which might improve the manuscript.

 

Specific comment

Abstract

The abstract contains the results of the study, conclusions are missing.

 Introduction

The aim is very general - should be specified.

Author Response

Thanks for your suggestions. All revised details in the article have been uploaded, please check them.

Author Response File: Author Response.docx

Reviewer 3 Report

It is a really good paper, the methodology is good, and the new model will be usefull for future research. There are only some minor issues that have to be checked. For example, Line 13 scale (singular) instead of scales, Line 22-23 you use bigger letters and Line 281 Grassland should not have a capital G. Also, why in Figure 3 you have c prior to b? You have to change it.

 

The only section that has many mistakes is the References. You do not follow the journal's standards (e.g., Italics for the journals' names) and some are bad written, like reference [3] ... No Title (?). You have to double check the whole section

Author Response

Thanks for your suggestions. All revised details in the article have been uploaded, please check them.

Author Response File: Author Response.docx

Back to TopTop