A Study on the Rapid Detection of Steering Markers in Orchard Management Robots Based on Improved YOLOv7

Round 1

Reviewer 1 Report

The manuscript is well-written and demonstrates a strong grasp of the subject matter. It effectively communicates the methodology, results, and implications. The organization and flow of content are logical.

 

Suggestion: To enhance readability, you can break up some longer paragraphs into smaller ones, especially in sections with technical details. This will make it easier for readers to follow complex concepts.

 

Overall, your manuscript is well-structured, clearly written, and provides valuable insights into the proposed method for autonomous steering in orchard management. By incorporating the suggestions mentioned above, you can further enhance the clarity, impact, and overall quality of your paper.

Note:  review the lines 198 to 201 and match the variable definitions with the equation (1)

The manuscript english is good and it is not difficult to read and understand.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors propose a “A study on rapid detection of steering markers in orchard  management robots based on improved YOLOv7”, which attempts to address the problem of steering markers detection and localization on orchard planting.

The authors correctly introduce the problem and lacks good literature review of today’s algorithms applied to the topic.

After reading the paper I would like the authors comments related to:

-The authors use a binocular camera and what’s the reason? Wats the difference using a single camera system to feed the deep network if the authors don’t use 3d image or depth information?

-Observing the Figure 4 (b) the dataset is not well balanced, and the author implement changes in the deep network instead on the dataset, why?

-Only 6 images for the test set? It’s not statistical relevant …

-The present research work is interesting in the computer vision field, but from the paper itself it not clear where are the depth images used and to conclude what authors conclude.

As finally, can your proposed method be generalized to different image conditions with no lack of performance?

As a conclusion of the review, the authors are invited to address my questions preferably. The paper is well written and for this reason my recommendation is that it may be accepted.

Recommend a new paper reading for content improvement.

Author Response

Please see the attachment

Author Response File: Author Response.pdf