A Machine Vision Method for Detecting Pineapple Fruit Mechanical Damage

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The study developed a Machine Vision Method for Detecting Pineapple Fruit Mechanical. The manuscript seems interesting for pineapple growers. Below are the comments.

1. What is the accuracy of the damaged vs non-damaged identification? This information is not mentioned anywhere in the manuscript. First step should have been identifying pineapple from the plant, then finding damaged ones and last step to be the area of the damage.
2. Put more quantitative values in abstract and conclusion sections to make it more appropriate.
3. How this study helps in automated pineapple harvesting process. Please discuss this in the discussion section.
4. The results and discussion section seems weak with information. Please discuss more on the finding of the study here and discuss its implications and comparison with the similar recent studies.
5. Image quality for image no. 12, 13 and 14 needs to be improved.
6. Discuss the limitations and future studies required in this area to make a perfect pineapple harvester.
7. Now a days lots of AI tools are available such as ANN, AlexNet, GoogleNet, Yolo etc. Did authors try to use those tools to get more identification accuracy?
8. Can the findings of this study be useful for other fruit mechanical damage identification?
9. Several other comments are given in the attached pdf.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This paper presents a well-organized and practical computer vision system for mechanical damage detection in pineapple using RealSense D435i. The approach is clearly described, and the system performs well indeed. However, there are improvements to implement, namely:
1 - Authors must present information on the generalization of the method for different lighting conditions, pineapple varieties and types of damage.
2 - Include a brief discussion or comparison with learning-based methods (e.g. CNN or deep segmentation models), even if not implemented.
3 - Consider commenting on computational time or feasibility for real-time applications, as this is important for industrial deployment.
4 - Although English is acceptable, some sentences are long and overly descriptive. An English review is required.

Comments on the Quality of English Language

The English could be improved to more clearly express the research.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have done good job with the revision. 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf