00:00.00 archpodnet Welcome to the show. Everyone Paul how you doing. 00:04.68 Paul I'm doing pretty good I was just out in the yard playing with a a sifting screen that I made that I haven't had the chance to use in the field but my wife needed some compost sifted so I got to play archaeologists for a little bit in the yard and that was a lot of fun how you doing. 00:32.80 archpodnet Nice I'm I'm doing good I just gotta comment on that real quick because that is a great use of a screen right? And as you're gonna hear on this podcast because I think the ad is still running. We have a contest that we're running for anybody who. You know mentions us on social social media you have to at us on social media or we don't see it. You know so at arcpodnet and or anybody who's currently a member or refers a member or becomes a new member in this by the end of the month gets to choose any screen they want from aeo screen. It's a website. In fact I think they're if I had to get. Think they based out of new england somewhere if I remember right? I might be thinking of somebody else. But either way they've said that you can pick anything off of their website and you might be thinking. Well I'm not an archaeologist what I would need a screen for and thank you Paul for illustrating an exact use of a screen you may not need as an archaeologist. Oh. 02:09.40 Paul Yeah, no I mean oh those are beautiful looking screens. They've got on their site. Ah yeah, no, no, they're really handy um in the yard for a lot of things. Especially if you're an avid gardener like my wife is um, not that she uses the screen but I guess I do now. 02:25.82 archpodnet Yeah, yeah, so and database button. 02:45.48 archpodnet Yeah, there you go so all right? Well I hadn't intended to talk about that. But you let me right into it so that was perfect. Um, we're doing great. We're out in seaside organ. It's ah it's a kind of a neat little coastal town that was. Actually established in eighteen Ninety Nine I think and they've got this really cool. You know promenade it's just south of probably like 20 minutes south of anacort sorry Astoria organ and Astoria a lot of people know maybe not by name but that's where the movie goonies was filmed so that's how. 03:46.92 Paul Ah. 03:57.10 archpodnet But usually no westoria. Yeah, but you can actually go to the house they have signs leading up to it that says you know hey people live here so be respectful to like traapes around in the yard. But you know either way you can see it from the road but all right? Yeah well. 04:22.70 Paul No, that's cool. 04:36.98 archpodnet Speaking of seeing stuff from the road except not really, it's going to be a drone heavy drone heavy episode this time around with the article we have in the and the topic we're going to talk about at least the None part iss gonna be drone heavy. So hope you got your hope you got your drinks ready paul what are we talking about today. 05:03.62 Paul Okay, well um I think probably when I was away in Iraq you were asking me for some topics. Um, and there was an ah, an addition of the advances in archaeological practice that came out in November and it was chock full of really interesting articles to me and so. 05:22.14 archpodnet Yeah. 05:41.68 Paul You know as you recall last ah, last episode we recorded about the hedgehogs and marvelous minds which came out of that same edition and um the same issue and None of the other ones I pulled aside that I thought might be good to talk about was ah, an article called a method of. Methodological framework for free and open source uav-based archaeological research and you know of course that free and open source that you know ding. That's one check for me. Ah u a v-based well, there's another check for me and ah and it's ah the the co-authors are. Kelsey M Rees and Sean field and Sean field was another ding for me because we actually interviewed him here on the architect back in episode None in November of 2019 when he discussed. 07:24.26 archpodnet Nice. 07:24.20 Paul Work He was doing on a chaan road network and for my money. It was a very interesting discussion. It was a very interesting article that that came out of too so I wanted to see what what they had to say about ah about this framework. Um, and as I read it as I started reading the article. Actually started going in a slightly different Direction. So at None glance I thought it's going to be a lot like that hedgehogs Article. Ah,, but it turned out to be something really for me quite different was the previous one that hedgehog article was more of a a show and tell this one. Was more of a justification for a particular tool set and a particular workflow along with a link to instructions to reproduce that toolset and workflow and that got me thinking about a whole bunch of other things that weren't directly in the article but were sperm you know ideas that that popped up because of this article. 08:53.78 archpodnet Right. 09:10.70 Paul And as I've said a None times that's something I like when something you know tickles my fancy and starts making me think about other things that may or may not be related and how they can interrelate and you know how they can make the practice of what we do better. Ah, it's it's. You know at the very least. It's some fun thought exercises but at the very best It's ah it's practical things that I can incorporate into my own work. 10:00.30 archpodnet Yeah, and that's what I love about some of these articles too right? is it's taking things that I always I always love taking things that you may not think have ah a practical use for the you know the science of archeology. But then again it does you know and sometimes it takes a suite of tools like. These guys approach and you know if if you want to do it free and affordably because what they mentioned is in the in the beginning of the article is using drones to augment and maybe in the future replace sorry I got to clear my throat maybe in the future replace archeological survey like pedestrian survey. Yeah, in some cases maybe like the full scale survey and I've actually talked about this I was pretty excited when I when I at least started with the abstract I was like wait what a minute they're they're they're actually talking about something that that I've been talking about for years but you know I've never done a study on it or actually completed it. Ah, just been an idea since doing survey in China Lake down in California but anyway I love that they're they're taking this approach because like they say the barrier to entry is often cost. You know in photoscan. What do they call it now a addressoft photo scan I still call it that but it's called mesh meta shape. That's right. 12:15.16 Paul Meta shape. 12:25.20 archpodnet And it's really expensive to use that and it's not. You know it's not necessarily user friendly unless you're used to those types of softwares. But it's expensive and you know it's kind of the cadillac of of doing this kind of stuff but as they show there's other ways to do it. It might take a suite of tools that metashape just does you know because they've got it all built in. 12:54.22 Paul M. 13:04.50 archpodnet Um, but that's what we're talking about So ah, excellent, good on them for for bringing this to light. Okay. 13:06.46 Paul Yeah, and what they do is they compare Meta shape against open drone map which is another photogrammetry tool specifically catered towards ah towards drone mapping and these are both excellent. It's not. It's not a one is better than the other. 13:37.80 archpodnet Right. 13:43.16 Paul Ah, 1 is definitely less expensive than the other ah but they want to have ah a system of testing a way of doing their work that they could then document so that it could be replicable which is really what what got me interested in this. It's that. 13:51.20 archpodnet So. 14:18.60 Paul Idea of documenting things in such a way that they are replicable because we like to pretend in a lot of ways in a lot of ways we are in archaeology scientists and one of the key components of science versus other human endeavors is replicability of of tests and you know so it's often been said that you know you can't re-excavate a site. Ah, so you have to document everything so we don't we you know when it comes to Excavation. You don't have that replicability that you might have with running another you know chemistry test For example, but that's not the only thing that we do. We do all sorts of other kinds of work and. 15:21.00 archpodnet Right? Brett. 15:36.40 Paul The idea of documenting how you do it so that you don't have to think through every last step and so that you know if I do it and I've got it documented and then the next person that takes over the task or they read what I've done and they apply it to their site. They don't have to reinvent the wheel that is a different kind of replicability. But it is one that I think really benefits our field you know and this is nothing new, but here they are talking about it in an open access Journal article good for them talking about using it with free and open source software good for them that lowers the barriers for entry ah and with links then. 16:18.64 archpodnet And. 16:48.92 Paul Go from this article back to the protocol. Um, and if I haven't used that word yet protocol the way that I'm going to be using it whenever I mention it today is this notion of it's basically and a set of instructions and so in. Lots of different scientific endeavors definitely in medicine you have protocols you do things Xy and Z you do step None step 2 step 3 maybe it's a decision tree but you follow along so that you can always get things in the same way. Um, so they have a link then from their article. 17:44.44 archpodnet Right? so. 17:51.40 Paul Back out to the protocol that they've got hosted on github which means that that protocol is also open and that's where I started thinking about other things because I've been receiving all sorts of I guess the algorithm is throwing them at me sometimes it shows up in Twitter sometimes it shows up on my newsfeed but different open source. Protocols not necessarily using open source software but the protocols themselves are open source for different kinds of archeological work and and that's where this really started to you know to make me think a little broader not the specifics of what they were doing ah though I should probably talk about that a bit but. How what they're doing also serves as a model for what I think is a good way for us all to be working so I'm going to actually reel us back here a little bit for what what they do. Ah I don't think they get really to the point of identifying archeological sites but they they choose an area. 19:21.16 archpodnet Right? right. 19:44.56 Paul Called the ah the mesa verde north escarpment which is slope terrain that's adjacent just to the north of Mesa V or national park. Um, and across that area. They chop it up into different squares and they categorize those squares by the by the density of the ground cover. And then they choose 4 different squares that they determine our so of a size that they can fly them in a certain number of batteries and then those different squares those four squares are a. 4 different densities of of ground cover low medium medium high and and extreme extreme ground cover. Ah and ah and then you know they they. 21:09.00 archpodnet Ah, right. 21:27.46 Paul Fly their drones they get there and most of the protocol actually has to do with the what you do with the um with those photographs and how you process them but they also do a side- by-side comparison of the results they got with the settings that they used with O been drone map versus those that they used with ah with Meta shape. Ah, to show that open Dronemap is entirely a good replacement for their case. Um, and then again that the the real meat of it for me is that if you go to that Github page. Ah, you can then see Step-by-step how to install the exact same software. 22:16.48 archpodnet Yeah, yeah. 22:40.24 Paul You can download their sample images. Um and then test it yourself on your own hardware see how long it takes see if this is going to be a workable workflow. Maybe adjust it a little bit for yourself. But um, that that is a huge value for people so we don't all have to reinvent the wheel. We don't all have to start at ground None and try to figure out each bit of not just how to use a software. But if you're using open source software a lot of it getting it onto your computer as a hassle and so they tell you how it's a very easy way to do it if you're running a mac. 23:45.62 archpodnet Nice, nice. Yeah, it was well a well done like how to guide and in fact I think it was in if I remember pulling up the article here again. It's it's in their how to series anyway. So they they had a really good like you know, step by step and a couple things I want to point out. And None of their figures figure 3 they actually have side by side openrome map and then it says ajasoft photo scan like they changed the name of that what like four five years ago I thought but everybody still calls it ajasoft photos scan which I think is hilarious that they're never gonna get rid of that name from all the from all the the. 24:46.68 Paul Yeah now no. 25:00.22 archpodnet Professional users but and and you mentioned that they didn't really do artifact analysis or feature analysis I should say on those images but they do have they did do a ground survey like a Pedestrian ground survey after the fact after they you know flew their drones. 25:08.36 Paul O. 25:33.52 archpodnet And the ground survey identified some features and I thought it was really interesting in figure 4 the article seeing those features drawn on one of the maps versus you know the side by side where it's not drawn and the side where it is drawn and I'll tell you what some of them are relatively obvious when you see them and others are. Not so obvious when you see them and I'm not sure that if I were looking at this I could pick it out but I'm not the one that has to be convinced right? because ultimately the the end goal of this kind of thing once we kind of nail down drone survey and we get the resolutions down. We get the the flying heights down and we figure out. Okay, this is the best way to do this for the imagery. 26:38.88 Paul Um, yeah. 26:47.26 archpodnet Now we got to do something with that imagery and you know having a person look at it and identify features is 1 thing but teaching a computer in an ai to do it would be a totally other thing and that would be kind of the way to go and then ground truth it by archaeologists. That's what I've been saying for years and I think we're finally starting to get to the point where this is. 27:11.16 Paul Um, yeah yeah I think it is and um I don't think I'm giving you up any trade secrets but the project I'm going to be working on for the next couple months in saudi that's one component of it. There are a number of known archeological sites that we're going to visit but there are. 27:24.22 archpodnet Possible. 27:41.86 archpodnet So. 27:50.48 Paul We're expecting to find other sites and part of that finding any other sites is to work on building a um, a machine learning environment that can identify additional sites. You know so it'll be a combination of none person infield on foot data collection that we feedback to the Gis team. Tries to generate models that they'll send back to us to test right? and so it'll be a ah process of refinement. Um, and that's what they're doing in this article too that just that's not the point of this article. Ah but it's clearly gearing up to something like that. 29:00.96 archpodnet Mm. 29:02.14 Paul And that's what they used for part of the test I thought it was interesting that those side by sides. Ah for my eyes. The open drone mapped ones actually looked better than the photo Scan than the photoscan ones. Ah, and ah I. Think the reason is is that the open drone wrap ones were a little smoother and I think they're a little smoother because their resolution isn't quite as high which actually is a good thing for this particular use case. Um because what we're talking about here Remember this is different ground Covers. We're looking at digital. 29:48.32 archpodnet Yeah, they are smoother. 30:06.64 archpodnet M. 30:16.46 Paul Digital terrain models not digital elevation models and so that becomes a software thing. Oh. It's all a software thing but right so you stitch together your your your structure for motion your your photogrametric images and you get your. Digital elevation model and it shows you the elevations at the top of all the Bushes. For example, that's usually not what you really care about you. Want the bottom of the Bushes. So the digital terrain model strips out those anomalies and you know there are a bunch of different ways for it to do that and some may be more successful than others. 31:00.10 archpodnet Right? right. 31:27.54 archpodnet So. 31:26.76 Paul So maybe overly aggressive but in those side by sides I always saw the features that they were highlighting that they traced I saw them better in the openrome map. Once. 31:48.86 archpodnet This right? Okay, well I personally I personally like the crispr images myself the smoother ones that they are. They do have a certain. I don't know appealing look to them because they are smoothed out and then the the photo scan ones again like you said were a little more crisp and and they're a little more abrasive from that standpoint but I felt it personally easier to see stuff. Especially if you I mean I zoomed in on the pdf I don't have the original images but I'm assuming if you zoom in on the original. Images especially inside the software you're going to see some incredible detail on both of those and from the looks of just these images in the article. Maybe even more detail in the photo scan one which from an ai perspective I don't know if that's better or worse for it to be doing shape recognition you know is it better to have. Ah, higher definition in detail so we can really pick them out or does that add more artifacts so to speak like so like digital artifacts to the image where it can't really pick them out. But if it's smoother it can it can pick up these you know identifiable shapes in this particular context a little better who knows. Anyway, let's go ahead and take a break and continue discussing this article on the other side if you like here and stuff like this. Don't forget to go to http://arpodnett.com/membersto help support us for just seven ninety nine a day or less if you buy it annually and it really helps and and keeps all this going and you know keeps Paul going to Saudi Arabia now it doesn't pay for that. But I'd like to think that it would send us to other places at some point in the future so we can so we can really get some boots on the ground podcasting all right with that. We'll see you guys into the side of the break back in a minute.