Round 1

Reviewer 1 Report

This study proposes a method to generate real-time constant viewpoint from videos acquired by stationary UAVs. Its main contributions consist in combining short-term trajectory with a long-term registration technique resulting in a registration method which is jitter-free and driftless.

The document is easy to read and follow.

The English is in general clear but should be revised.

The subject of the paper is interesting, with potential of application.

Authors should add images to illustrate de results and not only graphs so that readers can have a clearer understanding of the impact of the proposed method in the final results since you claim jitter-free and driftless.

For the sake of comparison authors should also consider including in their acquisition setups a gimbal to compare the effects of video with and without gimbal.

The conclusions section is very limited. Authors should detail more their analysis.

Although authors refer in the final paragraph of conclusions section that in the future the proposed method should cope with lighting variations during day long videos the problem is that even in short time videos that light variation can occur which is very common. This is one of the drawbacks of the proposed method.

In line 20 Fig.1 is referred but only appears 2 pages after?! Please correct.

The caption of Figure 1 seems too long. The description of the videos at least should be written inside a paragraph.

The first reference to Fig. 2 in the text appears after de image insertion. Please correct.

Table 1, Figure 5, 6 captions should also be reduced and the commenting part relocated to a paragraph.

In line 320 Table 3 is referenced after presenting the table. Please correct.

In line 343 Fig. 8 is referred but appear in the middle of references?! Please correct.

Author Response

Please see the attachment.

It includes a letter common to both reviewers as well as the updated manuscript (starting at page 8). We hope that this way of answering your review is convenient.

Author Response File: Author Response.pdf

Reviewer 2 Report

General considerations

The paper shows a workflow to reduce the jitter and drifting of a video captured by an UAV in a stationary way. The proposed approach using warping transformation based on the estimation of camera motion and a motion error. This work is an extension of the conference paper “Jitter-Free Registration for Unmanned Aerial Vehicle Videos” written by the authors and widely cited in the present document.

Of course, the paper is suitable for this journal. The paper is well organized: the paper is divided in six section and the titles of them are reasonable. The length of some sections is not reasonable, such as the Conclusions, Introduction and the Section 4: proposed approach.

The abstract is readable although the presence of a reference to the paper [1] strongly limits the comprehension of the topic paper and its motivations.

The introduction provides a sufficient description of the topic a deeper state of art should be reported. The section 3 is sufficient clear. The motivations for this study are clear, although in the paper lacks a precise description of some passages of the workflow that were hidden using a citation to the paper [1]. In order to obtain a uniform and accurate explanation of the proposed approach I strongly suggest writing the key passages of the previous work. Further details should be added in the section 5.

The conclusions are appropriated and coherent with the paper but can be extended with further considerations.

The number of references is sufficient.

General evaluations

The contribution is readable. The organization of the paper shall be improved, while the clarity is not always present. In some parts there are insufficient details, for this to be considered an accurate description of experimentation methods especially when the authors explain the implementation.

Of course, the paper is appropriate for this journal, the quality of writing is not always satisfactory: even if in general the paper is readable. In general, the length of the paper is not reasonable. The images are appropriate to the text and its contents but the position inside the document should be improved and some references inside the text should be added. The paper is technically accurate for the description of method, like the mathematical notation. The details in this paper are not so numerous, especially for the experimentation part where the authors talk about the implementation.

Further and detailed reviews are in the table below.

Review Table

Legend:

green = suggested review ; yellow = edit is no mandatory; cyan= please edit or add that is requested

Author Response

Please see the attachment.

It includes a letter common to both reviewers as well as the updated manuscript (starting at page 8). We hope that this way of answering your review is convenient.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Dear authors, thank you for your answers. First of all, I’d like to apologize to you for two mistakes:

• I did not provide the colors on the table;
• I lose during copy&paste operation part of sentence on the second row of the table

By the way, thank you for your answers and your work, the overall paper clarity was improved as well as the discussion of conclusion. I appreciated your modifications.

Comments for author File: Comments.pdf