Deep Homography Estimation and Its Application to Wall Maps of Wall-Climbing Robots

Round 1

Reviewer 1 Report

- the tile should be revised. As is, it sounds like an article related to unmanned climbing machines.

- the structure of the network (Figure 3) is confusing! further elaboration would be useful.

- some further information on the testing environment that was used to obtain results shown in Figure 4 would be useful.

- including the original code and perhaps some video links would add to the credibility of the proposed algorithm.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Line 22: Another useful application could be on ‘open-pit mine walls’. Open-pit mine walls are usually steep or nearly vertical and, therefore, these methods can be highly useful for the mining industry. Please add at the end of line 22 along with this reference:

Sayab, M., Aerden, D., Paananen, M., Saarela, P., 2018. Virtual Structural Analysis of Jokisivu Open Pit Using ‘Structure-from-Motion’ Unmanned Aerial Vehicles (UAV) Photogrammetry: Implications for Structurally-Controlled Gold Deposits in Southwest Finland. Remote Sensing, 10(8), 1296.

Line 33: Elaborate figure 1 caption. What ‘H’ stands for? “Wall map” is not enough. Describe two images on left and right.

Lines 44-45: Spell out without abbreviations ‘SIFT’ and ‘ORB’.

Line 72: Refer to the figure 2 at the end of this line.

Lines 60 to 80: These paras are more like a discussion. I suggest moving these paras to the end for the discussion purposes. Perhaps, add another section ‘Discussion’ just before the ‘Conclusions’. Otherwise, add a new sub-section here: ‘Previous work’.  

Lines 88-92. No need for this para, one can see how the paper is structured.

Line 95: This statement does not fit here “our proposed HomographyFpnNet has a smaller estimation error than previously presented estimation errors”. This should come at the end of this section.   

Line 100: Briefly describe ‘VGG architecture’ here.

Figure 4. Instead of using ‘our’ use ‘this study’ in the graph.

Line 208: This subsection can be the part of the ‘Discussion’ section, just before the ‘Conclusions’.  

Line 255: I can’t see any green box in figure 10. Please highlight or enlarge the labels/photos.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf