Enhancement of Boring Vibrations Based on Cascaded Dual-Domain Features Extraction for Insect Pest Agrilus planipennis Monitoring

Round 1

Reviewer 1 Report

Major comments

Wood-boring beetles are a major threat to forests, with larvae of some species being difficult to detect as they live within trunks and are covered by bark. To address this, a promising surveillance method involves inspecting the vibrations induced by larval activity in the trunk. However, environmental noise can disrupt this identification process and reduce accuracy. Previous studies have shown the necessity of adding an enhancement procedure before identification. In this paper, the authors propose a small but powerful vibration enhancement network based on deep learning that combines frequency and time domain enhancement in a stacked network. The dataset used includes boring vibrations of Agrilus planipennis larvae and various environmental noises. The authors report an increase in signal-to-noise ratio (SNR) by 18.73 dB after applying their enhancement method, with a processing time of 0.46s for a 5s segment on a laptop CPU. Several classification models showed a substantial increase in accuracy using enhanced clips. The experimental results suggest the authors' approach contributes to the early detection of wood-boring beetle larvae.

The paper is clearly written and relatively easy to read. The current form of the paper has no substantial problems, with some particularities that could be improved.

Introduction

The introduction is very comprehensive and brings the reader into the subject of the paper. It is very suitable even for a reader who is not familiar with environmental issues.

However, I think there are some phrases that are not quoted and that should be quoted accordingly. E.g line 32-33.

Methods

Regarding the Materials and Methods section, it can be stated that the authors have provided a meticulous description of the experimental protocol down to the last detail. The methods employed were thoroughly validated, and all possible sources of interference were eliminated to ensure the integrity of the experimental results. Overall, the authors have ensured a high level of scientific rigor in their approach.

Results + Discussion

In the Results section, it can be noted that the authors have provided clear and concise descriptions of the obtained results, ensuring their accessibility to the reader. Additionally, in the subsequent Discussion section, the authors have meticulously compared and contrasted their results with the existing scientific literature, thereby demonstrating a high level of scientific rigor and scholarship.

However, I suggest that a conclusion section should be added, highlighting the most important points of the article, to ensure that readers are left with a clear understanding of the key takeaways. The article is quite complex, and a conclusion section would help to summarize and emphasize the most important findings.

Author Response

Thank you for your appreciation of our work. I have summarized the main contribution of our work and added the conclusion section to the article. The missing references in line 32-34 are added too.

Reviewer 2 Report

This is an interesting research topic. The style, structure and writing of the article, in my opinion, look great.

The topic of the manuscript related with managerial efforts to regulate the emerald ash borer (EAB) Agrilus planipennis populations. The main aim of this research is comparison of algorithmic solutions for improving EAB-damaged trees. The manuscript is interesting mainly for specialists in signal processing or bioacustics, but also for managers and field workers who deal with EAB. 

This is not first article of this kind, but technical and algorithmic solutions for insect pests detection inside the material are not well-investigated and, even more so, widely used. Hence, the topic is state-of-the-art. The results of this manuscripts have added to insect detection inside substrates the way to more effective signal processing and to development of practice-oriented methods of damaged objects detection.  The paper is well written and easy to read. The conclusions are consistent with the evidence and arguments presented, and the authors addressed the main question posed.

I believe this study provides important estimates that are useful for insect pest monitoring, as it can be integrated into larval surveillance programs via mobile deployment. It can also be used to develop prototypes of early warning systems for borer beetles in forests and cities.

I have only minor comments:

- L. 178, “Fraxinus” should be italicized (generic epithet), and "sp." (species) should be replaced with "ssp." (species).

- There is no "Conclusion" section in the article.

Author Response

It is a pleasure to hear that you appreciate our work. I have set the "Fraxinus" to italics and "sp." is changed to "spp.". Also, I added a conclusion section summarizing and emphasizing the most important points of our article.