Non-Destructive Identification and Estimation of Granulation in Honey Pomelo Using Visible and Near-Infrared Transmittance Spectroscopy Combined with Machine Vision Technology
Round 1
Reviewer 1 Report
English grammar needs to be improved. There are inconsistencies in verb tenses and in punctuation
This paper reports on a new technique to assess granulation in grapefruit. Although the presentation of the results is good, the paper needs to have extensive english editing. The writing is poor, so some parts of the manuscript are hard to understand, as inconsistencies abound, for instance lines 62 to 64. That sentence does not make any sense.
Also, the authors need to make a stronger case on how this is novel enough to warrant publication in this journal. The way this is described implies that there is a new application of an already known technique to another different cultivar. Please state, why this is novel and how potentially would translate to the industry. This is necessary to improve the significance of the paper.
Author Response
Dear Reviewer,
the manuscript had been revised, please see the attachment.
Author Response File: 
Author Response.doc
Reviewer 2 Report
This paper aims to present a correlation between honey pomelo granulation levels and exterior indicators – e.g. area, volume, shape. Moreover, it also intends to support this analysis in spectral imaging, namely visible and NIR ranges. Non-destructive methods to evaluate crops and yields in different aspects are becoming increasingly possible due to spectral imaging technology – indeed, it is becoming more affordable and compact – and to the help of artificial intelligence. Techniques that allow the quick, accurate and reliable identification of issues within precision agriculture can potentially represent an enormous added value to both producers and consumers alike. As such, I consider that quality works in these areas are very welcome and can represent an effective contribution to science, with the possibility of eventually being applied in the real world.
As a general appreciation, the paper needs a detailed and exhaustive language revision. Indeed, the way the paper is written hinders the perception that readers gain on it: it has few typos, but many gender and number issues, badly written sentences, paragraphs that make absolutely no sense at all and punctuation errors. I had to guess some of the authors’ intentions when reading it. This must be addressed.
Regarding the introduction, it states what are the objectives and issues to resolve in the proposed work. However, the authors provide only one reference to a subject like machine vision in precision agriculture. More precisely, fruit. The way the sentence is written let’s the readers deduct that [15] is a review work. It is not. There are many works detecting and estimating fruits’ characteristics using machine vision: they are not remotely referred. It is unacceptable in an area where research has being done at an excellent pace and with quality works published. In line 50, the authors use the word “complicated”. This is a scientific work. Please restrain of using adjectives or when a subject is “complicated” please explain it on a manner that a reader can grasp. It is my understanding also that the authors write way to much information about machine vision architectures. If, as a reader, I intend to know more about a given topic that (granted) has a contribution to the work but it is not central, I would look it up.
The first sentence in results & discussion section is “Mass and volume were reduced with increasing granulation levels.”. I assume that the authors intended to write that there is a noticeable consequence in both mass and volume of a pomelo, when granulation levels increase. However, that is not what the authors’ phrase states. Generally, I like this section and the way that it is discussed but more clearer results in what regards the relation between estimated data and real data would be welcome.
Lastly, conclusions summarize the work but do not draw any significant information. They need to be rewritten to clearly state the paper’s contribution.
One of my main concerns regarding this paper is the feasibility to apply learned knowledge in real life. I only brought up this because the authors state it in the abstract. Moreover, they state also “In the future, it also can be utilized to pathological research of thick-skinned fruit or other citrus fruit.”. This is not proven. Please refrain to write sentences that cannot be substantiated by the work being submitted.
Author Response
Dear Reviewer,
the manuscript had been revised, please see the attachment.
Author Response File: Author Response.doc
Reviewer 3 Report
I enjoyed reading "Non-Destructive Identification and Estimation of Granulation in Honey Pomelo Using Visible and Near-Infrared Transmittance Spectroscopy Combined with Machine Vision Technology."
I think the paper makes a convincing argument that multi-source data fusion was effective and feasible when combining spectroscopy with machine vision to accurately detect granulation. It is a useful application of optics applied to a real-world problem. The presentation is clear with enough detail to make the paper self-contained.
Some observations:
Line 404: "classier" should be "classifier"
Line 429: incomplete spelling of "levels"
My only real criticism is the lack of rigor in defining abbreviations when they are first used in the paper. Not all, but some. For example the full names of the chemometric approaches are not spelled out until line 187. These faults should be corrected in the final paper.
There are a large number of specialized abbreviations in the text and even when introduced correctly, it's challenging to keep them all straight. Although not a requirement here, it would be helpful to have a table for a reference for the abbreviations when confronted with nearly 20 abbreviations as was true in this paper.
Author Response
Please see the attachment.
Author Response File: Author Response.doc
Round 2
Reviewer 2 Report
I would like to begin by expressing my appreciation by all the work that the authors did on further improving this manuscript.
All my suggestions and comments were duly addressed. I would like to have more reference works regarding fruit detection and monitoring using machine learning - available, for example, in MDPI's Remote Sensing - but there is a minimum referred as is.
Author Response
Please see the attachment.
Author Response File: Author Response.doc
