Data Import and Visit to the Google, Bangalore office.

After the data was arranged in a suitable format and converted into shapefile, it was imported as an earth engine asset. I began the code by importing the Sentinel 2 MSI Level 1C data and the shapefile in the GEE code editor interface. I then filtered the image collection based on the temporal and spatial resolution of the available ground truth data.

The next task was to get a list of a feature collection that did not have clouds over them in the filtered image collection. I used ‘QA60’ band data for the same. I used the max reducer to reduce regions. After filtering the feature collection based on the ‘QA60’ mask, it was intended to get the matching dates for the satellite data and the available ground truth data. I am still struggling to find that part out. But it will be done soon.

By this time, I had spent some 3 weeks in Bangalore. I was getting well-adjusted there when I realized I had to leave already. I had a flight back to Delhi on 30th June, so I wanted to wrap up at least the initial part of the project. So, the next step in the process was to export the satellite band information in the form of the table after which it is planned to chart the pixel values against the ground truth data. This will help in analyzing the relationship between the band and the in-situ data. this step will be initiated once the table-data is exported.

Devaja ma’am had scheduled a visit to the Google office for me and my project mentor, Raj on 25^th June 2019. A meeting with Ujaval sir was also lined up so that we can discuss on what all has been done till then related to the project. I presented my literature review as well, which was followed by suggestions on how I can improve the code that was done until then. 

I was lucky enough to click some of the pictures with Raj and Devaja ma'am. Sharing some of them here.

I heartily thank Devaja ma'am and Ujaval sir for having us at Google office and will be looking forward to meeting them again. 

So, all in all, I had a great time working full-time at WRI, Bangalore. I'll be visiting again at the end of the month of July for the progress presentation. I am really excited about that and will be looking forward to having more insights into the project.

Data Arrangement and Sample Map Preparation

Entering the month of June, that is, the second month into the SoEE program, I decided to work full-time with the organization (WRI) I had applied for. I moved to Bangalore on 6^th June 2019. Before moving there, I discussed with my mentor Raj as to where I should look for a residing place. I decided to move to a PG at BTM Layout, 2^nd phase. I did have issues in adjusting myself there, but that was manageable.

I joined WRI on 6^th June itself. I met my mentor, Raj who introduced me to the other people in his team. The first day I met Samrat and Sahana. Samrat is Director, Urban Water and Sahana is Manager, Sustainable Cities. In the next few days, I got a chance to meet Jaya who is Strategy head, Urban Development. I also got a chance to interact with Akash, who is being hired as a GIS/Remote Sensing Consultant in the team.

Initially, I was assigned to arrange the ground truth data. The data was confidential as it was collected from Namami Gange project. Since it was in excel format, I had to turn it into the earth engine asset. I first filtered the available data based on the Sentinel-2 satellite passing time in the excel sheet itself. Then, I converted that into a shapefile using the ArcGIS platform. The shapefile was later imported as the earth engine asset in the GEE platform.

The data had a temporal resolution from January 2017 to September 2018. I bifurcated the data in two temporal stretches, that is, 2017 and 2018. This was done just to simplify the amount of data to be handled on earth engine initially. Later it is proposed to be combined.

I also generated some sample maps using the C2RCC processor in the SNAP toolbox, just to validate the expected outputs. The maps were for two water quality parameters – Total Suspended Matter (TSM) and Chlorophyll-a (Chl-a). I ran the generation for three study areas (1. Upper Lake, MP; 2. Bellandur Lake, Karnataka; 3. Hussain Sagar Lake, AP) out of which I got visually good results for Hussain Sagar Lake, Hyderabad. For the other two areas, the processor was not able to give out proper results. The visual representation was prepared in the ArcGIS platform and for map representation, I had used Google Earth images as background. I have added a few images here showing the samples.

    Chl-a map for Hussain Sagar dated 16.10.2016 

    Chl-a map for Hussain Sagar dated 29.04.2019 

    TSM map for Hussain Sagar dated 16.10.2019

    TSM map for Hussain Sagar dated 24.04.2019

So, this was about the data arrangement and basic sample output preparation for the project. I’ll talk about the code in my next post.  

First month into the SoEE program

It is more than two months into the SoEE program, so I know I am a bit late to write about my experience in the last two months. But I will anyway try to write in segments what all has been done in the past two months.

So, I had begun my work by some basic literature review. I had started with reading "EVALUATING THE POTENTIAL OF SENTINEL 2 DATA FOR WATER QUALITY ESTIMATION - ISMAIL KARAOUI ET. AL. (2019)" where the authors talked about how Sentinel 2 data can be used to empirically model the water quality parameters in the Bin El Ouidane Reservoir in Morocco. The model equations were derived based on correlation and regression analysis on the in-situ data and the satellite data. I had initially chosen this paper to go through because my ideas related to this project were like the approach followed here.

I also went through some of the papers which described the utility of OLCI Sentinel-3 data so that I get a wide approach of Sentinel products. I concluded, since the resolution of Sentinel-3 products is about 300m, it won’t serve the purpose of the SoEE project.

In addition, I looked for the C2RCC SNAP tool references to have an idea on the machine learning domain of the project. It was initially decided to generate some sample maps for the validation step of the project, which was another reason to explore this SNAP processor.

Some of the other papers which I went through to completely dive into the subject matter of this project are as follows:

·   Mohammad Haji Gholizadeh, Assefa M. Melesse, And Lakshmi Reddi (2016) “A Comprehensive Review on Water Quality Parameters Estimation Using Remote Sensing Techniques.” Sensors 16, 1298

·  Gidudu Anthony, Banura Constance, And Namugga Angella (2016) “Monitoring Water Quality on Lake Victoria Using Modis Imagery.” Ijtd Volume 3, Issue 1

·    Miguel Potes, Gonçalo Rodrigues, Alexandra Marchã Penha, Maria Helena Novais, Maria João Costa, Rui Salgado, And Maria Manuela Morais (2018) “Use of Sentinel 2 – Msi For Water Quality Monitoring at Alqueva Reservoir, Portugal” Proc. Iahs, 380, 73-79

·   Ismail Karaoui, Abdelkrim Arioua, Abdelghani Boudhar, Mohammed Hssaisoune, Sabri El Mouatassime, Kamal Ait Ouhamchich, Driss Elhamdouni (2019) “Evaluating the Potential of Sentinel-2 Satellite Images for Water Quality Characterization Of Artificial Reservoirs:  The Bin El Ouidane Reservoir Case Study (Morocco).” Meteorology Hydrology and Water Management Vol 7, Issue 1

·     Kaire Toming, Tiit Kutser, Rivo Uiboupin, Age Arikas, Kaimo Vahter And Birgot Paavel (2017) “Mapping Water Quality Parameters with Sentinel-3 Ocean and Land Colour Instrument Imagery in The Baltic Sea.” Remote Sens., 9, 1070

·  Brockmann Carsten, Doerffer Roland, Peters Marco, Stelzer Kerstin, Embacher Sabine, Ruescas Ana (2019) “Evolution of The C2rcc Neural Network for Sentinel 2 And 3 For the Retrieval of Ocean Colour Products in Normal and Extreme Optically Complex Waters.” Http://Step.Esa.Int/

·    Kel N. Markert, Calla M. Schmidt, Robert E. Griffin, Africa I. Flores, Ate Poortinga, David S. Saah, Rebekke E. Muench, Nicholas E. Clinton, Farrukh Chishtie, Kritsana Kityuttachai, Paradis Someth, Eric R. Anderson, Aekkapol Aekakkararungroj And David J. Ganz (2018) “Historical and Operational Monitoring of Surface Sediments in the Lower Mekong Basin Using Landsat and Google Earth Engine Cloud Computing.” Remote Sens., 10

In a crux, I went through the in-depth reviewing of related literature in the whole one month into the SoEE program.

First week into Summer of Earth Engine program

I am very new to blogging, so this is going to be very basic. I would include all I can think of while going through the whole project. It'll be like a story, though!

It was just another morning for me. I had started thinking that I must not be shortlisted for the SoEE program because I had not received any mail by then. The results were supposed to be released on 30th April, but they got delayed by 2 days. Anyway, I had already stopped thinking about it when I saw "Congratulations on being...." in my notification panel. I think I got a bit awestruck, maybe? I don't know. I was thrilled, obviously. So, I told my friends, my project guide, family about this wonderful opportunity that was given to me and they were all so happy about it. Yeah! It was a Google internship.

I am thankful to Google and WRI to consider me worthy of this internship. I am very excited to learn about Earth Engine and its implementation in remote sensing. I am highly grateful to my project guide at my college, Dr. Amba Shetty, without whom I wouldn't have come across the GEE platform. 

Into this week of the SoEE program, I have begun with enhancing my insight into the project. My mentor, Mr. Raj Bhagat, along with his other team-mates, will be guiding me throughout this project. I am looking forward to this wonderful opportunity. I hope I am able to carry out my responsibilities towards this project in a very well-organized and sincerest way.