I was recently invited to visit Ewha Womens University in Seoul, South Korea, to attend an acoustic symposium they had put together. I arrived on the 3rdof October and on the 4thI gave a seminar presentation to one of their courses on the ecology and acoustics of the amphibian community on Mount Kinabalu. The afternoon was spent with Proj. Yikweon Jang’s research group talking about acoustics and going over some analysis methods of acoustic data. Over the weekend I spent some time hiking up one of the mountains in Seoul with some of the lab members, including being shown the river which runs through the city. It used to be an underground river with a road over it and nothing thriving in it, but a few years ago they decided to remove the road and have now worked on making the river alive again and the pathways seem to be appreciated by the Seoul citizens to walk along. The symposium started on the 7thstarted, I was very excited as Prof. Jerome Seuer from the Natural History Museum in Paris was coming, he has written or been involved with the majority of acoustic packages in R so it was great to finally meet a person that is so influential in your everyday work life. Jerome presented on Ecoacoustics, I presented on our acoustic system and the information we can get out of our monitoring data, Dr Tzu-Hao Lin presented on clustering and marine acoustics, Dr Kyong Seok Ki presented on some of the work done in South Korea on impact of airports on acoustic landscape amongst others, Dr Ming-Feng Chuang presented on acoustic components of an amphibian species and Dr Hortense Serret talked about citizen science and acoustic data collection. It was a great symposium as everyone was working on very different things and using the data in different ways so it truly demonstrated the wide range of applications of soundscape data. I also attended a workshop the day after by Dr Tzu-Hao Lin on analyzing long-term spectrograms. It was a fantastic trip and I learnt a lot as well as having a great time finding out about the research conducted in Prof Jang’s lab, so I am very grateful to him for inviting me. The university itself have beautiful grounds which are surprisingly calm considering the amount of students and how busy Seoul is! Another thing that was great was that I got to meet up with an old friend, I met Dr Ming-Feng Chuang in Taiwan in 2014 when I visited his lab– since then we have met once or twice at a conference but it was great to see that he is doing so well in South Korea.
In September I attended the Association for Tropical Biology and Conservation – Asia Pacific chapter meeting in Sri Lanka and presented on acoustic monitoring using open-source hardware. I met quite a few new people that I have wanted to get in touch with before, like Eben Goodall that works on acoustics in birds. I also made some new connections and saw quite a few interesting talks on elevation studies as well as co-occurrence networks. There were also quite a few people there that I haven’t seen in a while and it was nice to catch up and see where they are with their research. One of the evenings there was a tribute night to Anslem de Silva who have contributed greatly to herpetology in Sri Lanka, he used to be a magician and even showed us some magic tricks. It was a nice meeting and it was good to see such strong attendance from Sri Lanka and India, especially by young students. I also went to the business meeting (a good idea if you want to ensure you know what’s going on with a society you are a member of) and amongst other things witnessed the new Asia-Pacific board get elected in.
I somehow managed to also purchase and wear a saree for the first time for the opening ceremony as the Sri Lankan president (who is also the Environmental minister) came to give an opening speech and to plant a tree.
In June I attended the Young Conservation Scientist Conference in Thailand. I presented on acoustic monitoring using open source hardware as a representative from the Lndscape ecology group at Xishuangbanna Tropical Botanical Garden (we had four people in attendance, one professor, me as a post-doc and two master’s students (now PhD)). This was a great conference for young aspiring conservationists, the atmosphere was friendly and welcoming, including certain talks (such as the speed talks) being on purpose designed to have a committee that could give feedback to the presenters. There were a range of interesting talks on everything from snake tracking to using paleo data for predictions. For many I think this was the first time to present and many presented in their second language, so it was great to see that both the organizing committee and the attendees ensured that it was a nice and encouraging environment to have their first experience of presenting in.
Recently I have had to submit both contract and grant applications, which often requires you to submit information related to the measurement of success and progress in a project. Gantt charts is a great way to get a good overview of major milestones in a project so is a nice thing to submit with applications to ensure that whoever reviews the application can get a good understanding of the different tasks in the project and when they will occur. It also allows a good overview for the project manager if the project gets approved so that they can ensure that the project meets its milestones on time and can help team members to ensure that their specific tasks are delivered on time. So here is some R code using ggplot allowing you to produce a neat Gantt chart which organize the different tasks into groups as well.
Recently a friend of mine posted a question on Facebook regarding alternative solutions to camera trapping rather than the off-the shelf trail camera providers. The main reason being that the quality of the photo’s are not good enough, they do not shoot in RAW and the resolution is often just 5-8 megapixels – way too low if the photos are needed for promoting the research of a specific conservation program even if it is functional for species identification.
Now, whereas the person in question know how to build their own using a DSLR they needed something less complicated for their parabiologists to use that still produces photos of a high enough quality. I suggested two solutions and though that I would write them up and post them here. The first one is a lot simpler and suitable for parabiologists, whereas the second one is cheaper and will hopefully improve the field significantly in the future.
1: Hacking a Canon camera for use as a camera trap
Pregnant deer caught by a camera trap. Image is from MammalWeb.
That’s right! Because there is something called the Canon Hacker Development Kit (CHDK). These hacks only work on the point and shoot range but can for example extend the settings to incorporate the ones available on the DSLR range. Multiple point and shoot Canons can do RAW formats and you will certainly be able to find one that shoots higher than 8 megapixels. Now you need to ensure that you know the correct firmware on the camera, as the type of CHDK you need to download will differ between the different ranges of camera. There are two different main scripts:
There is also a low light solution which can be read about on this forum thread.
To install the script, you download the script and then load it into the CHDK scripts folder. Now you can load the script, set the parameters and finally use the script by following the steps on the CHDK page (they are relatively simple). For some newer Canon versions ACID is needed instead.
If this was so great, would it not have been used by some scientists already? Well, it has! So you can read either the scientific paper or the blog.
Racoon caught by a camera trap. Image is from MammalWeb.
The great thing about CHDK is that users can develop their own script for the Canon cameras–allowing for new innovations. The camera can also be used for multiple things, not just as a camera trap.
2: Setting up a Raspberry Pi as a camera trap
Saad Chinoy’s write up of using his R-Pi as a camera trap during the Dinacon conference 2018. Copyright Saad Chinoy (to see more updates from Saad follow @saadcaffeine on instagram).
If you haven’t heard of microprocessors in the open-sources hardware community before, then here’s your first introduction! There are multiple models such as Arduino, Micro Bit and of course the Raspberry Pi. They are a very tiny computers at a very small price, with an excellent open hacking and code developing community online. Part of the movement is to make computers available to everyone but also teach coding. In short, they can be used for a lot of things! Let’s focus on the camera trap. First you need a Raspberry Pi, a Pi Zero is probably adequate and they are only around 5 USD. Secondly you need a Raspberry Pi camera (sorry! sadly the USB cameras won’t work) which is around 30 USD. If an IR sensor is needed, they are around 6.50 USD. Unfortunately, these cameras are only 8 megapixels, but hopefully in the future we’ll see some better ones coming out, or someone providing a hacking solution for an easy to use higher resolution one so considering the low cost these are worth putting out there as a potential solution.
The camera trap itself, the R-Pi board is behind the description but the camera itself is hanging over the front. From the Dinacon conference 2018. Copyright Saad Chinoy.
I’m going to give links to a script that I’ve seen someone use that seemed to be working fine, however, this is a lot more complicated to install and set up than the Canon! The benefit is that you have a much wider range of how to set up the script (full list on the wiki page). You can choose between video or images, single or multiple images . The camera settings can still be controlled and if needed you can schedule when things are going to be captured. An additional benefit? You control it from a computer after the initial installation and can see photos in real time or just get them uploaded. Great if you have WiFi nearby as they can auto upload, less good if you are in the middle nowhere.
For an alternative with a suggestion with a housing solution see this post. If you are just interested in building a camera scratch see this Adafruit post.
Whereas I do not think this is the best solution for parabiologist at the present time, I think we will see some major advances in this arena soon so hopefully we will see some new and cheaper camera traps with better components getting developed and tested. The major benefit of using a Raspberry Pi? Well, once it is not needed as a camera trap anymore you can just put in an SD card loaded with something else (or wipe the one used for the camera trapping) and then the sky is the limit as to what you do with your computer. The Raspberry Pi community is a huge open source movement with open-source hardware and scripts constantly being developed. If nothing else, giving one to a child and getting them to follow some online tutorials or getting them a basic starter kit of sensors and cables is a great way to get them knowledgeable in coding.
3: Get super swanky and make your own 360 camera trap
While these inventions do exist, they are pricey! Luckily, once again, you can potentially hack a point and shoot camera, if it is capable of taking 360 degree photos. If you want to know more head over this instructables page created by the amazing Andrew Quitmeyer on how to put one together using a 360 camera, an Arduino and some motion detectors (or just watch the movie below).
____________________________________________
Thank you to Saad Chinoy for input, Scott Trageser for the initial question and to everyone that participated in the discussion online!
Monitoring of amphibians in Southeast Asia is scarce, however the rates of new species descriptions are substantial. This biodiverse region is undergoing high rates of land clearing, which coupled with a changing climate, is expected to have a major impact on amphibians in the region. Over 40% of amphibian species in Borneo alone are listed in one of the IUCN threatened categories, and for the majority there is limited information available on ecology, population numbers, and life history. This thesis investigates; shifts in distribution, ecology, community composition of the amphibian community on Mount Kinabalu in Borneo, Malaysia, as well as the use of automated acoustic techniques for monitoring these communities.
Andrew Quitmeyer’s communication students produced an awesome video about my research! It was great to have them out in the field for a day and thank you to NParks for giving us the permission to film.
This is probably one of the most amazing experiences I have had in my life. Mainly thanks to the people that were present, and the format of the conference itself. This was no ordinary conference, where you spend hours for just one week listening to updates about research and networking with as many people as possible. Dinacon is actually running over six weeks, and there were (are – it is still running until July!) not just academics and scientists in presence. Here you instead had an amazing collection of hackers, engineers, scientists, artists, educators and a lot of other professions. The point was to make something, to interact with others, learn from them, get help from them and support other people in their projects.
Making a rainfall water collection system.
My goal was to try and get a LoRa system working (if you do not know what it is, check out this link) for my acoustic and environmental monitoring stations. This would enable us to transfer some of our data back to a server, at a low cost and low power consumption. Whereas it can’t transfer our audio files (they are too large) we could get summary data that ensures we know the stations are still running and get some updates about what is going on. I arrived to Koh Lon being super ill, I had a fever for several days that had just passed and still had a bacteria infection in my throat – but work goes on as they say.
My equipment waiting on the rainy pier
First of all, I was pretty lost before I even arrived. I had spent a lot of time reading up on LoRa and LoRaWan but still did not have a clear idea of how to get this system running. I ordered some components, not even knowing if these were the correct components. First of all, you need a server (or gateway – that receives data from multiple clients), and you need at least one client (node – the one that collects and then send data on). I arrived with some adafruit RFM9X components with antennas for the clients, and I had a Dragino LoRa/GPS hat and an Adafruit LoRa/GPS hat to try to use as servers. I spent a lot of time reading up on how to get this working, but the majority if instructions were only for using a Raspberry Pi as the server, not as both a server and a client and that is the set up we needed! In addition, it turns out that the code to get these to run is written in C++, and I have never coded in C++ and at the time did not even know what a make file was or what “making” the files meant. I has some excellent help from Saad Chinoy in the start who helped with explaining some of these things, without making me feel stupid or as if I should know this already. We did get the system to start running after a day or so, but it kept on crashing, not even a little crash. Both the server and the client would freeze up after a message or two and you couldn’t stop the running or it, or get it to respond even through SHH. I spent days staring at code, and with help form others trying solutions to no avail. I spent a day or two having a break from it as well to try and clear my brain and figure out the issue. Normally whilst engaging in a side project or helping others with theirs. In the end, a man called Harold Tay spent several hours just bouncing ideas and trialing code. I had googled the same issue so many times, without any returns online – how was it possible that such a major issue had never happened to anyone else? Well, as I was searching for the mysterious answer, a line describing what the different GPIO pins on the RFM9X does flashed past and something clicked – “the IRQ is not enabled on the Raspberry PI”. We had already tried something with this several days earlier, to no avail, but at this stage I had learned what an h file and we had an h file that defined what some of these pins were doing. So instead of going into the make file, I accessed the h file and disabled this pin. All of a sudden, the client was running without issue! The server was more of a problem and took a couple of more days to solve. The last day on the conference, whilst coding aboard Diva the ship, with some help from Brian Huang and Alex Rogers we finally figured out how to enable the server to also run without the IRQ defined, whilst sending the correct type of data. I would probably have figured this out in the end, but it would have taken several weeks or even months. So I would like to thank everyone at the conference for the input, help, support and for making it ok to ask for help.
Brian Huang and Alex Rogers on board the Diva, everyone working and we managed to solve the final coding issue with the server.My grant agency that have supported the LoRa work.
I also got up to other things than coding, and saw so many inspirational things! I slept in a tarp for the whole nine days because I concluded that me and the hammock were not going to get on so the ground was better (actually very comfortable). The first morning I woke up with loads of hornbills flying around. The second night I got attacked by a lot of frogs, but they calmed down on their “love” for the frog researcher after that. I woke up between 6-8am every morning, by myself, and excited about getting to work. I crocheted for the first time in 20 years and made a dragon’s tail for a robot, as well as an ant. The bikini pouch saw the light of day, finally, with the use of some second hand ties and the Dinacon sewing machine. I made a small little bag out of the tie remains. I learnt how to make coconut rope and now know how to open a coconut with only a stone. I helped create a rainwater collection system. I helped someone that wanted to learn how to swim and saw them go from being uncomfortable to not being scared of the water – a remarkable thing to witness. I have been to workshops on geobacteria batteries, seen presentations from artists and their artwork and met educators that do their best to inspire the future generations. I have seen people that never stop creating and seem to use every minute of every day to make things or do things. People that appear to put in 48hrs in a day where I can only squeeze out 24. I saw at least three different types of camera traps created by different people, a pirate flag with a light up eye got soldered together, saw a plant sensory system set up and collecting data and a robot arm making intricate drawings. I once had this discussion with Mo Donnelly, that researchers and artists are probably two of the few careers where you have the freedom to constantly create and come up with ideas – the two fields are more similar than some people realize, without creativity us academics will end up running out of things to explore. Allow people to come up with those crazy ideas, they might fail, but sooner or later they will figure out a way around the issue – or straight through it, and it is what drives our society forward.
Forest hike.
Most of all I met people from all around the world, in a loving, caring and non-judgmental environment. Andrew and Tasneem have I believe, pulled of a conference like no other.
Saad is soldering together pirate eyes in the maker space on the porch.Brian Huang helped me do a test of how far the LoRa could send messages, we made it all the way to the mysterious bear cage.Wire mess for one of our LoRa clients! They record environmental data as well and we use both the SPI (for LoRa) and I2C (for the sensors).LoRa server with a Dragino LoRa/GPS hatFirst time the client managed to send a message and it was received!Finally both the client and the server are working without crashing!Squinty the AwesomeOne of my diary pages depicting the autonomous Sparkfun robot and the dragon’s tail I crocheted for it.A crochet antThe pirate flag on board the DivaThe DivaStig setting up and running the plant sensor networkTarp tent modification to avoid frog attacks.What my tarp tent looked like the first two nights.A cute little Microhylid was hanging out in my bed when I got back.Morning view the first day, sunshine and horn bills around.Forest adventure!Client transceivers being soldered. Getting the antenna attachment soldered was super fiddly, only broke one luckily.
Our short communication on the reproduction and contact call of Kalophrynus baluensis have been published in Herpetological Review. This is the first formal description of eggs and confirmation of tree hole breeding in this species. We also describe the contact call of this species that the male use to to encourage the female into amplexus which is different from the previously described advertisement call. Our section is on pages 96-97:
On the 14th of September I followed Merike, a Master’s student from John Measey’s lab from Stellenbosch university into the field. The reason was to have a look at the aSCR technique, an acoustic monitoring array that can be used to calculate population densities (amongst other things). I saw John Measey present at the World Herpetology Congress on this technique and asked him about it, he kindly invited me to come and have a look at it in action whilst I was visiting South Africa. It was a very interesting and exciting day, apart from the seeing the ASCR field set up we also got to visit the Kenilworth Racecourse Conservation Area. Amazingly one of the most important Cape Flats Sand Fynbos remaining is preserved in the middle of this horse race track.
