A Review of Artificial Intelligence and Remote Sensing for Archaeological Research

Round 1

Reviewer 1 Report

between two numbers use n dash no minus.

expand acronymous the first time they appear please

in the figures be consistent with fonts and font sizes, some are too small and are impossible to read.

In 3.2 line 212 you say that spatial resolution increases with the height of the platform. I think is all the other way around. In 3.2.2. line 287 you say that spatial resolution decreases with the height of the platform, please be consistent.

quantify resolutions for aircraft and drones, check the bibliography and quantify it, even in a general and orientative manner to be able to compare it with satellite images.

line 322 give some examples and comment the variety of research, at the very least some citations will be needed.

line 347, as this is illustrated in literature, please give examples and comment on them. also with citations.

You need to expand the remote sensing part, techniques, examples, different indexes...

you need to expand the AI part, is just a brief abstract of a couple of study cases.

The discussion is meaningless. It may be part of the introduction but as discussion doesn't add anything to the rest of the article. coment on current trends, future possibilities, actual uses etc. I like the the last sentence about the need to clarify the basis in wich archaeologists locate and classify archaeological monuments and remains, may be work a bit more on that concept for the discussion.

 

 

Author Response

"Please see the attachment."

Author Response File: Author Response.docx

Reviewer 2 Report

The manuscript on the review of artificial intelligence and remote sensing for archaeological research has been reviewed. The article is very well written.

In Figure 1, the number of articles in 2012 is also higher than that in 2011 and 2013, but the authors have not explained it.

172 lines of literature references need to be enclosed with left brackets.

In Figure 3, can different picture styles be unified? Replace the two pictures on the right. Moreover, the picture content does not correspond to "i", "ii," and "iii" in 3.2.2. At the same time, the passage in this part lacks strong literature argumentation, so it is recommended to add literature citations.

The authors might be interested in reading https://doi.org/10.3390/ma15082910 . This article discusses artificial intelligence may be helpful for the authors.

 

Some pictures are blurring. It is recommended to increase their definition.

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 3 Report

This paper provides a review of automated approaches to remote sensing analysis in archaeology and cultural heritage research, looking specifically at AI and machine learning applications. The article is one of several recent papers that reviews the use of automated methods for archaeological remote sensing (see for example, Câmara et al. 2022, Davis 2021; Fiorucci et al. 2020), and will be of interest to readers of Remote Sensing, and the special issue more specifically. Still, given the recent surge in review articles focused on AI in archaeology, the authors should be careful to detail how this contribution is distinct from these other recent reviews (especially that of Câmara et al. 2022 which was published only a month or so ago).

Additionally, I do have several comments and concerns that need to be rectified before I can recommend publication. I go into more depth in the attached annotated copy of the manuscript, but give some general comments below:

1. The exact scope of the paper is a bit unclear between the abstract and main text. The abstract makes a specific reference to spaceborne/satellite remote sensing instruments, but the focus of the actual manuscript contains aerial, spaceborne, and ground-based sensors. This should be rectified to better encapsulate the actual content of the article. Additionally, the authors provide at least one example of AI applications in archaeology that do not involve any of these kinds of sensors, but focus on material analysis (ceramics, ref 105). There is a whole paragraph devoted to this between lines 439-459 which seems out of place with the rest of the focus and scope of the article. As such, I'd recommend either removing this or redefining the scope of the review to include such applications (but doing so will likely require adding additional background and literature review on the ArchAIDE project (from where ref 105 comes) along with numerous other examples of how AI has been used to improve ceramic, lithic, and other artifact identification as well as radiocarbon dating estimates (e.g., Cole et al. 2022; Garcia‐Molsosa et al. 2021; Reese 2021, etc.).

2. The authors present two bibliometric data searches from SCOPUS to support the notion that remote sensing archaeology has increased rapidly in recent years and that AI applications have also risen but at a much slower rate. I see numerous potential issues with these analyses, specifically that only one search term was used (which cannot encapsulate the entire range of applications that might fall under this broad category) and is only conducted in the SCOPUS database, which is not an exhaustive compendium of published literature. Other search terms need to be added as well as other databases to ensure that the results as they currently stand are accurate, and not a product of the database chosen and the very limited number of search terms used in the search query. (You can see Davis 2020 [cited as ref 15 in the manuscript] for some ideas of how to broaden search terms and bolster the analysis). I would also recommend comparing SCOPUS with at least Web of Science and/or other databases like Dimensions (https://www.dimensions.ai/).

3. You state in the introduction (lines 70-71) that a main question that remains open has to do with "the future and the contribution of Remote Sensing and Artificial Intelligence techniques in archaeological research". However, you never really return to this point to lay out what you see as the future trajectory of AI in Remote Sensing Archaeology, current limitations, future ways to improve, etc. This would make your contribution far more significant and I would suggest adding to your discussion with some of these points.

4. Please check for grammatical/typographic errors, as there are a few instances that I noticed while reading the manuscript (some of them are indicated on the attached annotated document).

References Mentioned Above

Câmara, A., de Almeida, A., Caçador, D., & Oliveira, J. (2022). Automated methods for image detection of cultural heritage: Overviews and perspectives. Archaeological Prospection.

Cole, K. E., Yaworsky, P. M., & Hart, I. A. (2022). Evaluating statistical models for establishing morphometric taxonomic identifications and a new approach using Random Forest. Journal of Archaeological Science, 143, 105610. https://doi.org/10.1016/j.jas.2022.105610

Davis, D. S. (2021). Theoretical repositioning of automated remote sensing archaeology: Shifting from features to ephemeral landscapes. Journal of Computer Applications in Archaeology, 4(1), 94–109. https://doi.org/10.5334/jcaa.72

Fiorucci, M., Khoroshiltseva, M., Pontil, M., Traviglia, A., Del Bue, A., & James, S. (2020). Machine Learning for Cultural Heritage: A Survey. Pattern Recognition Letters, 133, 102–108. https://doi.org/10.1016/j.patrec.2020.02.017 

Garcia‐Molsosa, A., Orengo, H. A., Lawrence, D., Philip, G., Hopper, K., & Petrie, C. A. (2021). Potential of deep learning segmentation for the extraction of archaeological features from historical map series. Archaeological Prospection, 28(2), 187–199. https://doi.org/10.1002/arp.1807

Reese, K. M. (2021). Deep learning artificial neural networks for non-destructive archaeological site dating. Journal of Archaeological Science, 132, 105413. https://doi.org/10.1016/j.jas.2021.105413 

Comments for author File: Comments.pdf

Author Response

Please see the attachment

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

the article is supposed to be a review of AI and RS on Archaeology. About RS you only mention that is a no contact non destructive technique to get information and then you explain the different platforms used to get the data from.

I think you will need to explain how does RS works, what kinds of data it works with, the visible spectrum, the hyperspectral images, multispectral images and the indexes we get from there, how are those indexes calculated, what are they used for, and what data do they gave us, explain compare and illustrate different techniques, what are the best for what purposes. Lidar, visible indexes and Osabi, Hyperspectral as NIR, thermal images, UV. Multispectral as NDVI, narrowband as red edge. describe them and put some examples of each with figures and compare them to see what are the strong points and what can we use each one for. As a review of RS techniques you will need to explain and compare those techniques, their history and their development, it is not enough to say that they are nondestructive and are taken from the air, an explanation, examples and discussion about them is needed.

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 3 Report

I thank the authors for addressing my previous concerns regarding their manuscript. For the most part, I think these issues have been resolved, but a few things still remain that need to be clarified before the paper is publishable. Specifically, please see the major point, below, which the authors began to address, but I believe requires further elaboration.

Original Comment - Point 3: You state in the introduction (lines 70-71) that a main question that remains open has to do with "the future and the contribution of Remote Sensing and Artificial Intelligence techniques in archaeological research". However, you never really return to this point to lay out what you see as the future trajectory of AI in Remote Sensing Archaeology, current limitations, future ways to improve, etc. This would make your contribution far more significant, and I would suggest adding to your discussion with some of these points.

Author Response: The discussion has been change adding future changes and possible implementations using RS and AI in archaeology.

‘Ultimately, many will be wondering how the future of AI and other automation tech-niques within RS archaeology will be and if is there anything more that can offer be-yond data collection. We can envision soon the use of RS platforms, with more sophisticated sensors, which will be able to detect archaeological remains with higher accuracy within less time. Archaeologists might be able to use more intelligent automations to collect samples without disturb the condition of the archaeological remains.

Also, climate changes like temperature increases as well as changes in moisture cycles, continue increases dramatically, and inevitably will be expected an increase impact on the archaeological heritage (both to known or still buried) [109]. Perhaps we should expect in the future a higher contribution of AI to climate change risk assessment at least in studies that have been done for specific cultural heritage monuments. Therefore, with continued technological progress of RS and AI, it seems potential that archeology soon will become an advanced technological science.’’

Comment on Author Reply:

I still think that the original contribution of this article needs to be clarified. Specifically, think of how to answer the following:

1. Can you speak to the current limitations of AI approaches for remote sensing archaeology that need to be addressed in future studies? I.e., methodological limitations/accuracy/etc.? Lack of research on AI for specific kinds of cultural heritage management, for example semi-permanent occupations? Hyperfocus on the Global North, with limited applications to the Global South? Etc.

2. With AI at the state where it is, what sorts of questions can begin to be addressed, and are there any current examples of such studies that are making active use of automatically generated datasets that relied on such methods? Some examples that come to mind are Cerrillo-Cuenca & Bueno-Ramirez (2019), Parton & Clark (2022), and others that seek to make use of automatically generated datasets to address broader questions relating to settlement patterns, population dynamics, and other lines of inquiry.

3. You mention using AI to help mitigate climate change issues, can you provide more specifics? Any studies that have done similar things to date?

References:

Cerrillo‐Cuenca, E., & Bueno‐Ramírez, P. (2019). Counting with the invisible record? The role of LiDAR in the interpretation of megalithic landscapes in south‐western Iberia (Extremadura, Alentejo and Beira Baixa). Archaeological Prospection, 26(3), 251–264. https://doi.org/10.1002/arp.1738 Parton, P., &

Clark, G. (2022). Using lidar and Bayesian inference to reconstruct archaeological populations in the Kingdom of Tonga. Journal of Archaeological Science: Reports, 45, 103610. https://doi.org/10.1016/j.jasrep.2022.103610  

Other specific comments: 

With respect to your bibliometric searches, particularly the second one focused on AI and Archaeology, I would argue that your search still needs to be bolstered with additional search terms. Not everything under "AI/artificial intelligence" is going to capture more specific applications like convolutional neural networks, deep learning, or machine learning, specifically. While it may not impact the argument you're making by much (which demonstrates that these trends are growing), perhaps at least cite your analysis alongside other larger ones like your ref 17 [Davis 2020], or McCoy (2021) (for a general analysis of trends in digital methods like remote sensing and GIS in archaeology).  

McCoy, M. D. (2021). Defining the geospatial revolution in archaeology. Journal of Archaeological Science: Reports, 37, 102988. https://doi.org/10.1016/j.jasrep.2021.102988      

Author Response

Please see the attachment

Author Response File: Author Response.docx