Estimating Plant Nitrogen by Developing an Accurate Correlation between VNIR-Only Vegetation Indexes and the Normalized Difference Nitrogen Index
Round 1
Reviewer 1 Report
Major Revision:
This study focuses on monitoring vegetation nitrogen concentration, a critical factor in plant physiology and development, as plants require substantial nitrogen during their growth period. Nitrogen plays a vital role in supporting the growth of roots, leaves, stems, branches, shoots, and fruits, as well as promoting flowering. The Normalized Difference Nitrogen Index (NDNI) has been established as one of the best indicators for vegetation nitrogen concentration estimation, utilizing spectral bands at 1510 nm and 1680 nm from the Short-Wave Infrared (SWIR) region of the electromagnetic spectrum. However, many remote sensing sensors, including cameras and satellites, lack SWIR sensors due to their high costs. Overall, the manuscript presents a valuable contribution to the field of nitrogen estimation in vegetation and offers a potential practical solution for real-time monitoring and precision agriculture applications.
The details comments are as below:
Major Comments:
1. The study focuses on a specific region (Harran Plain in Turkey) and uses data from a particular sensor (Hyperion). The findings may not be fully generalizable to other regions with different vegetation types or to data collected from other sensors.
2. The manuscript proposes a novel approach to estimate nitrogen content using VNIR-only band indexes. However, a direct comparison with results obtained using SWIR bands for nitrogen estimation could provide more insights into the effectiveness of the proposed method.
3. The manuscript does not mention external validation of the trained model on data from other sources or sensors. External validation is essential to demonstrate the model's robustness and its applicability to datasets beyond the training data.
4. The manuscript lacks a comparison with existing methods for vegetation nitrogen estimation, such as those using traditional spectral indices or other machine learning techniques. A comparative analysis would help assess the novel approach's superiority over existing methodologies.
5. Incomplete Atmospheric Correction Discussion: While the manuscript argues against atmospheric correction due to potential negative effects, a more thorough discussion of this decision, including the possible impact on the accuracy of the nitrogen estimation, would strengthen the paper's argument.
6. Limited Discussion on Outliers: The presence of outliers in the regression results is mentioned, but the manuscript does not elaborate on their potential implications for the model's performance and accuracy.
Comments for author File: 
Comments.pdf
none
Author Response
Dear Reviewer,
Thank you for your valuable comments. Please find attached my responses to the reviewers' comments and the revised version of my manuscript.
Author Response File: 
Author Response.docx
Reviewer 2 Report
The experimental topic is a current and important area. The importance and usefulness of data collected by remote sensing is constantly increasing.
The study is useful for accurate understanding of the data collected with the help of various remote sensing devices. Nowadays, near-infrared and rededge are a very useful channel for calculating vegetation indices, their further refinement is promising for increasing the correlation of remote sensing data with the crop yield.
The article is well edited, but the experimental methodology, mainly the basic agrotechnical and analysis strategies need to supplement.
The literature chapter details a large number of properly edited literature of adequate quality, but some literature would need to be supplemented or replaced. The quality of the chapter can be improved by collecting new, fresh literature after 2020.
In table 2, it says that it uses a value greater than 0.2 units for its analysis, so that it is definitely data from plants. Basically, your findings are correct, however, for analyzing satellite images, we usually use all values, since the given pixel can clearly contain plant and non-plant parts. In my opinion, it is not advisable to filter out values from the soil or other non-plants with respect to NDVI below a high-precision GSD value of less than 5 cm.
The quality of the article could be improved by editing the figures and tables in a better, more aesthetic way, if possible.
The results are adequate, but the presentation and comparison of the new results with the results of other international researchers requires a more detailed description. Need to suplement with some sentences the results, beacause the described results are more wider and well prepared than the last part of the result.
Author Response
Dear Reviewer,
Thank you for your valuable comments. Please find attached my responses to the reviewers' comments and the revised version of my manuscript.
Author Response File: Author Response.docx
Reviewer 3 Report
- I think Sec. 1 is too long. I recommend the author to divide it into two sections, for example, the Introduction and Related works. Also, I recommend that the contributions of this paper can remain in Introduction.
- Could you provide more details about the samples (ground truth) how you collected?
- Could you provide more details about the architecture of your neural networks?
- Also, Please add some comparisons among different methods, such as DNN, RF, GBDT, etc.? The authors could add a subsection in Sec. 4 Discussion.
- I am also curious about the transferability of this method. The author could add some simple discussion about it, such as TGRS23-Partial domain adaptation for scene classification from remote sensing imagery
- Some references are recommended to cite in this paper:
- ISPRS21-Growing status observation for oil palm trees using Unmanned Aerial Vehicle (UAV) images (plant monitoring from UAV images)
- RS21-Estimation of paddy rice nitrogen content and accumulation both at leaf and plant levels from UAV hyperspectral imagery
Author Response
Dear Reviewer,
Thank you for your valuable comments. Please find attached my responses to the reviewers' comments and the revised version of my manuscript.
Author Response File: Author Response.docx
Round 2
Reviewer 3 Report
The authors have addressed all my issues. I think this paper is ready for RS Journal.
