Review Reports

Round 1

Reviewer 1 Report

The paper is well-crafted and presents promising results. With minor revisions, it is suitable for acceptance.

Please incorporate longitude and latitude information in the maps.

Regarding Table 1, could you specify the flight height of the UAV and the field of view (FOV) of the sensor for assessing spatial resolution?

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Mapping Fractional Vegetation Cover Using UAV Imagery to Guide Conservation of a Rare Riparian Shrub Ecosystem in Southern California

This is an interesting study which objective is to map fractional cover of riparian shrubs using multispectral UAV imagery. The authors tested different classification algorithms and spatial resolutions. The novelty of the work is not in the techniques used but in the application of these techniques to fractional cover estimate.

In general the manuscript is well documented and structured, methods are well depicted and appropriate. Thus my decision for publication after minor revision. I include some minor comments in the following lines:

I suggest to use the term unmanned instead of unpiloted in the term UAV, as I imagine the UAV is piloted in the distance.

I also suggest to use in the abstract the overall accuracy to compare the classification performance. Although kappa has been used traditional and it is worth to compute it for comparison purposes, there exist some concerns on its usefulness. Please, revise https://doi.org/10.1016/j.rse.2019.111630

Concerning the results section, I suggest to improve section 3.1. The results should be not only presented in figures but assessed in the test. Please, include some comments on the maps. I also suggest to change the order of appearance on points 3.2 and 3.3. It would be more self-understanding to know which classification and spatial resolution perform better before comparing the area estimates.

Finally, I would be desirable that the author’s state which resolution would select finally, 25 o 50 as the differences in results from the statistical point of view are low, for the final objective one resolution may be better than the other.

Lines 159-163 revise formatting.

Kind regards.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

 

The authors propose a drone-based habitat monitoring and classification method, which has some scientific significance for ecological conservation, but there are still some issues to explore.

 

1、In the third paragraph of the introduction, SVM, RF are already very mature algorithms in the field of remote sensing, and the author must reduce the description of such classification methods in the manuscript

2、The article discusses the classification problem at the medium resolution scale, UAV can achieve more accurate high-resolution monitoring, it is recommended that experiments can be carried out in higher resolution scenarios, we are very curious about the monitoring results at these high-resolution scales.

3、In terms of adding features, why was only the NDVI index considered in the study? Other index EVI, DVI, SAVI, etc. or texture features can also assist in classification, and a large number of studies have shown that the addition of texture features can significantly improve the classification accuracy of crops such as trees!

4、We recommend adding more work from the last two years to the cited literature to prove the timeliness of the work.

Author Response

Please see the attachment

Author Response File: Author Response.pdf