A Refined Apple Binocular Positioning Method with Segmentation-Based Deep Learning for Robotic Picking

Round 1

Reviewer 1 Report

The authors have made a good contribution to detecting the binocular images of apples. I have only two suggestions to further improve the quality of the paper.

1. Compare the proposed model with other deep learning segmentation models.

2. Show the training and validation accuracy and loss plots to further understand the performance of the proposed method.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

1. The paper is well written. It would enhance the quality if authors present an algorithm to show exactly the steps they followed. Algorithm is one little part explaining things very clearly.

1. English language is good. Try to avoid long sentences to make it easy to read.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

This research paper proposes an apple binocular positioning method based on Mask R-CNN. The manuscript is well-written.

I have some questions about the manuscript. There are more object detection and segmentation algorithms, so it would be better to add some comparison results on the accuracy of the proposed method. The comparison matrix can be checked in the following reference. 

Wang, X., Kang, H., Zhou, H., Au, W. and Chen, C., 2022. Geometry-aware fruit grasping estimation for robotic harvesting in apple orchards. Computers and Electronics in Agriculture, 193, p.106716.

Additionally, it is worth investigating and comparing the accuracy by using different sensors, like RGBD, RGB-LiDAR fusion for objection detection and segmentation in agriculture field. 

 

Please conduct further language editing on the manuscript.  

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 4 Report

The authors proposed binocular position method combined with Mask-RNN for detecting apples. The authors claim that the segmentation accuracy is 99.49%. The topic is interesting for reader of Agronomy, the paper is in well written with good novelty and results. It can be considered to publish in Agronomy. There are some minor concerns is listed as below:

11.     Line 26-28: this statement is vague since there is no specific requirement described.

22.     Keywords:  should not Capital for each keyword.

33.     Line 33-34: is the statement correct? The author should cite a related reference to prove it.

44.     Line 53-54: the sentence is out of place. The authors should move this sentence into the second paragraph or consider to ignore to keep the flow of the text.

55.     Line 66: “some researchers” phase is very vague. The author should be more specific.

66.     Section 3.3: the author should add a table for comparison with results from other groups in the literature.

77.     Future Works section reduces the novelty and feasibility of the proposed technique. The authors should consider ignoring it. Add some same discussion at the end of session 3.

88.     Fig.13: some apples are not detected. The author should add some discussion or improvement to increase the accuracy of the results.

99.     Line 308-310: the authors should specify the criteria of the target of the current study. Also, the requirement is based on which standard?

110.  The session of “author contributions, funding, data availability statement” should be completed.

Author Response

Please see the attachment.

Author Response File: Author Response.docx