I had a geo-spatial weekend!

Many days in the life of a researcher are spent going to conferences, networking and getting the pulse for the latest updates and advances in science and technology. I am no different species and I enjoy these outings very much, as they are disruptive to my everyday job.

This weekend I have attended the geo-spatial.org seminaries (mmm..well…for the 6th time in the last 3 years) in Timișoara, Romania. For those who are not familiar with this event, I have a brief confession to make: these guys are awesome! I am not saying it just for being friends, but because they are some of the best geospatial professionals in Romania and organize these events in order to share the knowledge and educate present and future generations in the spirit of open geoinformatics and open data.

 

24135549_10211349417435329_34425732_n
Timisoara is a pretty town.

 

The seminaries are coagulated across the idea of open source technologies, applications, and utility, and while I am not the most ardent supporter of this ideology (I promised myself I’ll develop on this idea on a future post), the thing I truly like and enjoy at these gatherings is the “community feeling”. These are not the typical conferences, where you wear formal or semi-formal attire, discuss for a brief period on your topics and interests while you chew some tapas and drink local wine. Nope, these are those types of a semi-informal nature, where your butt won’t sore, you’ll hear plenty of jokes during the sessions, people come because they are curious and not because they are supposed to, and still get to grasp the latest news in the field and feel the pulse of the only Romanian geospatial community (did you notice? It is the second time I call it like this).

The most valuable thing is they are open to students who seek to be active and build a career in the field. There is a mix of academia, professionals, important business players (that unlikely to other events, are here to mingle, not to sell) and young and enthusiastic blood. The seminaries are being held three times a year, each time at a different place ( the University of Bucharest, West University of Timișoara and the Babeș-Bolyai University in Cluj) and every time is refreshing to see old but mostly new faces. There is a lack of intimidation, an openness that students should really benefit from and grasp this opportunity to get noticed, share their work and communicate with people with more experience behind.

Romania does not a have a vibrant geospatial scene. Actually, things move slowly and there is barely something glued here in terms of professional community. There is the academia, with their secluded environment, which sometimes collaborates with the businesses, both sides concluding there is plenty of room for more. Also, businesses are scarce, divided by competition and brought together by the same competition. Here is where the most of the professionals are polarized. Apart from this, you may find some scattered material in the plaiade of state research institutes and agencies. Each sector has its own events, most being that type- with formal attire and pretzels and…Don’t get me wrong, I am not saying these are not interesting, helpful, worthy, or anything. They are. But are also intimidating to students, are not always meant for them (usually that is why student symposiums are held), have a certain approach (more business, theoretical, a different rhetoric) and participation and inclusion is not always high. This might lead to a sentiment of inaccessibility or intimidation. I know, here I can develop a lot and you can disagree or not, also, a lot. But this is not the point of this article, nor is this long paragraph.

Getting back to the subject, well, let me tell you a little more about the event itself: the geo-spatial.org seminaries.

The architecture: 2-day event structured in 1 day of oral presentations and workshops, and 1 day of two 3 hour workshop parallel sessions.

Total presentations: 17 spanning on topics from CAD to Earth Observation, Deep learning, Land administration, new products, applications and career advice.

Total workshops: 9 spanning from Earth Observation to cartography, programming and database intros, data access, product tutorials, and analysis.

 

GeO-SPatial Statistics
Data used here is estimated and might not reaveal the reality, but something very close to it.

 

On the first day, I’ve attended the presentations session, and although all of them were super thought-provoking (congrats to all the authors!), some of them caught my eye as they were closer to my interests, others just because of their nature. Let’s dissect:

  1. A radiograph of the geospatial open source ecosystem (Codrina Ilie) – I liked the topic, their endeavor, and the results. As the author said, I had no idea there is such a great pool of resources and how many relationships are between them. The study was carried by a consortium of companies for understanding the present open source environment and its benefits for business development. The present results can also be consulted as a cute graph here.
  2. Earth Observation Data v1 and v2 (Marius Budileanu, Ionut Șerban, Sorin Constantin and Cristian Flueraru) – a series of presentations that mapped the current possibilities in freely accessing remote sensing data on both states: lower and higher resolution. Did you know Planet offers up to 10.000 km2 of high-res data, based on a brief scientific proposal? No? Me neither.
  3. Using agricultural production data to validate NDVI (Andi Lazăr)- while I am not so prone to agriculture, this presentation was quite interesting in means of exploiting in-situ collected data from production machine sensors, to validate the results you get from satellite data. Recycling old data to peak into new utility.
  4. Some considerations on predicting agricultural production (Ștefan Manolescu) – I know, I know, agriculture, but this presentation had me from the first slides. Some interesting info on traditional agriculture and the comparison with modern models. I liked it for being honest about how prediction models are not perfect, how hard is to acquire all the data and actually getting the expected results.
  5. Object-oriented analysis of geospatial data (Lucian Drăguț) – a very well structured and paced explanation of the whole process, a refreshing feeling to see that this kind of studies is being tackled by the geographic community in Romania and some interesting results. 
  6. Deep learning for Earth Observation (Teodora Selea, Marian Neagul) – again, an introductory 10 minutes on neural networks and preliminary results. I have to agree I have a special interest in these topics (both this one and the previous one) because of my thesis, but I could not help but feel a little hopeful that the future will bring more similar studies in the autochthonous geospatial environment.
  7. Aspects of building a GIS career (Tudor Bărăscu) – I had to forgive Tudor for stealing the majority of the time for my presentation, he is always funny, and he builds his own way towards the students, reinstating many of the true values that should be adopted while building a career, of any type, not only in GIS.

The full programme and presentations can be found here. Notice that some of them are still in Romanian.

For the second part of the day, I’ve attended the Introduction to GDAL workshop. I had some previous attempts at working with GDAL, but some were successful and some failed. Both of the instructors (Ionuț and Sorin) are proficient and thought some external advice would be beneficial. I did not win the final prize, but I was mere seconds to do it. (jk).

Anyhow, it was difficult to choose which workshop to attend, as all three were interesting to me – the nice and useful introduction in PostGIS and the second part of the package about how access EO data (PostGIS – GEOintroduction in databases, Earth Observation data access).


The second day was shorter for me, by choise. I was an instructor for the RADAR 1.0.1- introduction to mosaicing workshop and spent my morning enchanting people with notions about polarization, amplitude, phase and other charming concepts, debugged SNAP, hoped for a miracle in the processing steps and ended explaining most of the things theoretically. People want to drink coffee at 9 am, not be hit by abstract, complicated things, keep that in mind! Feedback was good though and I am more than happy with that. If you’re interested, take a look at my tutorial on the subject.

 

mosaic100_RGB
In case you’ve missed it, here’s Romania under radar.

 

In parallel, there was a line of other super interesting workshops, like the one about Infographic design using QGIS and Inkscape and Tree inventory in the citrus plantations. Also, the afternoon session, seemed very appalling, covering topics like QGIS 3 , editing in Here Map Creator (Here’s own online tool for crowdsourced geospatial data) and analyzing and processing data in gvSIG (already a staple open source GIS software).

All in all, we had a full four days of geospatial knowledge, good humor, friendly atmosphere and community feeling.  If you’re eager on experiencing this, you’re more than welcome to the next session which will take place next spring in Cluj.

See you there,

Cristina


Disclaimer: The choices from this article are simply my own. I do not favor anybody specifically, the selected presentations were merely the ones that I felt more interested in. All the others were equally fine and explanatory and had a lot of useful information. The whole depiction of the seminaries is my own and can be somehow biased by its subjective nature. You are welcome to discuss in the comments section, below.
Advertisements

Bucharest under the radar

It has been a while since I’ve moved to Bucharest to study territorial planning and although it turned out not to be my lifetime career, I still peek into some of the topics, just for fun. This time, I’ve taken a turn from the usual types of analysis and played a bit with some remotely sensed data.

These days I’ve managed to finish ESA’s EO College fantastic radar course, Echoes in Space and since I am taking baby steps on learning more about SAR, I was thinking that replicating one of their tutorials with a more autochthone subject would be interesting and add more to my former urban planning and my current remote sensing skills.

The tutorial concentrates on deriving an urban footprint for Bucharest, the capital, the largest city and by far the most important site of Romania. Bucharest has a long history of chaotic development and has grown more and more horizontally, spreading its tentacles over the nearby area. One way of understanding the mechanisms behind its urban sprawl is to perform some change detection analysis. In order to do so, a clear delimitation of the constructed environment is essential, and one way of achieving that is to dig into the information radar sensors are capable of delivering. One good example is the mesmerizing map of the continent’s urban agglomerations, released by ESA last year and obtained from some extensive processing of TerraSAR-X/TanDEM-X SAR data:

Urban_Footprint.jpg
Urban footprint in Europe. Courtesy of @European Space Agency.

On the map, Bucharest strikes as a single black spot on the face of Romania, but I’ve wondered how it looks from a more zoomed in perspective. In order to find out, I’ve selected two S1A scenes: one from the 10th of July 2016 and the second one from the 18th of May 2017.

The methodology used to get the urban footprint from radar data can be found here. If you are not a fan of long step by step written tutorials, this video is basically the interactive alternative (which I also like better):

After some initial search of compatible overlayed scenes, many attempts to dismiss the whole thing and some long nights spent processing and reprocessing some of the data, the final RGB result was something of a very colored picture, ranging from violet to yellow, with the latter striking as high backscatter and high coherence levels over urban or constructed areas.

I’ve been positively surprised to notice how well defined is Bucharest in this image and how nicely you can trace main traffic arteries by just looking at the continuous stack of villages that accompany them.

 

S1A_IW_SLC_1SSV_split_Orb_Stack_coh_deb_ML_TC_Stack_RGB
Bucharest under the radar.

 

The fun part was to mask only those yellow bits and derive a clear urban footprint. I’ve played a bit with subsetting and threshold values here, and Tadaaam!:

 

subset_0_of_s1a_iw_slc_1ssv_split_orb_stack_coh_deb_ml_tc_stack_urban_footprint.jpg
The final result is a bit messy but beautifully distinct.

 

Now, there is a glitch. The whole process has been done using SNAP and although there is an option to export mask pixels, apparently it is not working for everybody and it is still under investigation in an open issue. There are numerous workarounds, but I haven’t identified one for me, yet. Hopefully, this gets solved, as I got super excited about the possibilities to use these newly resulted products.

 

bucharest_white
This is how the big black spot looks if you zoom in.

 

Cheers,

Cristina