thank you for your proposal, I put in cc also the GRASS GIS dev mailing list, as it might be a suitable project candidate if anyone is available for mentoring it. It is usually a bit more difficult to find mentors when the proposal comes from a student and it is not listed in our ideas page, however not impossible, and your idea sounds very interesting. Are you familiar with GRASS GIS?
I’d like to point you out our recommendations for students at https://wiki.osgeo.org/index.php?title=Google_Summer_of_Code_Recommendations_for_Students , particularly our guidelines on how to submit a proposal.
I’m a Computer Science Master’s Degree student whose research if focused on Deep Learning algorithms applied to Remote Sensing. Currently working at the Laboratory of Image Processing and Geoprocessing settled at Federal University of Goiás - Brazil. I’m also member of the High Performance Computing group of the same university (more information here).
Below I present an idea to explain how I can contribute to OSGeo community and I’m seeking for mentors interested in assist my development. Please, feel free to argue me any matter about the project idea.
I would also appreciate a lot if you guys indicate a potential interested mentor to my project idea or a OSGeo Project suitable to it.
Hope there’s some Interested ones out there!
Idea
The increasing number of sensors orbiting the earth is systematically producing larger volumes of data, with better spatiotemporal resolutions. To deal with that, better accurate machine learning approaches, such as Deep Learning (DL), are needed to transform raw data into applicable Information. Several DL architectures (e.g. CNN, semantic segmentation) rely only at spatial dimension to perform, for example, land-cover/land-use (LCLU) maps, disregarding the temporal dependencies between pixels observations over the time. Also, high-res remote sensing data (e.g. Planet, Sentinel) may provide more consistent time-series, that can be use in the identification of important LCLU classes, like crop, pastureland and grasslands.
This potential can be explored using Recurrent Neural Networks (RNN), a specific family of DL approaches which can take into account time dimension. A promising project idea would be implement a RNN approach (e.g. LSTM) to classify a Sentinel time-series, that will organize and preprocess an input data set (e.g. labeled time-series), calibrate and evaluate a RNN model, and finally classify an entire region (i.e. 2 or 3 scenes) to produce a map for one or more LCLU class. It will be great evaluate the accuracy and the spatial consistent of a map produced with a RNN approach.
A simple example on classifying LCLU with two classes (pastureland and non-pastureland):
Target region (input)
Generated LCLU map (output)
Best,
Evandro Carrijo Taquary
Federal University of Goiás
I really appreciate for your feedback! I’m not much familiar with GRASS GIS as I had only developed standalone codes in Python using directly libraries like GDAL and RIOS and used QGIS for layers visualization. But, as I observed here and here, Python scripts can easily be integrated within GRASS GIS and I could seamlessly adapt my programming skills to work with that. That way, I could compromise myself in integrating my already done codes (and new ones) into GRASS GIS software/libraries.
Also, if allowed, I could edit the Ideas’ wiki page to contemplate my own idea, so that it could be visible to a broader audience. If I get a mentor in time, I will make a detailed proposal for the mentors/community be able to understand better the idea.
Thank you very much,
Evandro Carrijo Taquary
Federal University of Goiás
thank you for your proposal, I put in cc also the GRASS GIS dev mailing list, as it might be a suitable project candidate if anyone is available for mentoring it. It is usually a bit more difficult to find mentors when the proposal comes from a student and it is not listed in our ideas page, however not impossible, and your idea sounds very interesting. Are you familiar with GRASS GIS?
I’d like to point you out our recommendations for students at https://wiki.osgeo.org/index.php?title=Google_Summer_of_Code_Recommendations_for_Students , particularly our guidelines on how to submit a proposal.
I’m a Computer Science Master’s Degree student whose research if focused on Deep Learning algorithms applied to Remote Sensing. Currently working at the Laboratory of Image Processing and Geoprocessing settled at Federal University of Goiás - Brazil. I’m also member of the High Performance Computing group of the same university (more information here).
Below I present an idea to explain how I can contribute to OSGeo community and I’m seeking for mentors interested in assist my development. Please, feel free to argue me any matter about the project idea.
I would also appreciate a lot if you guys indicate a potential interested mentor to my project idea or a OSGeo Project suitable to it.
Hope there’s some Interested ones out there!
Idea
The increasing number of sensors orbiting the earth is systematically producing larger volumes of data, with better spatiotemporal resolutions. To deal with that, better accurate machine learning approaches, such as Deep Learning (DL), are needed to transform raw data into applicable Information. Several DL architectures (e.g. CNN, semantic segmentation) rely only at spatial dimension to perform, for example, land-cover/land-use (LCLU) maps, disregarding the temporal dependencies between pixels observations over the time. Also, high-res remote sensing data (e.g. Planet, Sentinel) may provide more consistent time-series, that can be use in the identification of important LCLU classes, like crop, pastureland and grasslands.
This potential can be explored using Recurrent Neural Networks (RNN), a specific family of DL approaches which can take into account time dimension. A promising project idea would be implement a RNN approach (e.g. LSTM) to classify a Sentinel time-series, that will organize and preprocess an input data set (e.g. labeled time-series), calibrate and evaluate a RNN model, and finally classify an entire region (i.e. 2 or 3 scenes) to produce a map for one or more LCLU class. It will be great evaluate the accuracy and the spatial consistent of a map produced with a RNN approach.
A simple example on classifying LCLU with two classes (pastureland and non-pastureland):
Target region (input)
Generated LCLU map (output)
Best,
Evandro Carrijo Taquary
Federal University of Goiás
I really appreciate for your feedback! I’m not much familiar with GRASS GIS as I had only developed standalone codes in Python using directly libraries like GDAL and RIOS and used QGIS for layers visualization. But, as I observed here and here, Python scripts can easily be integrated within GRASS GIS and I could seamlessly adapt my programming skills to work with that. That way, I could compromise myself in integrating my already done codes (and new ones) into GRASS GIS software/libraries.
That’s really good for you to familiarize with writing code in GRASS, as that would give your idea more chances.
Also, if allowed, I could edit the Ideas’ wiki page to contemplate my own idea,
No, that is not necessary
so that it could be visible to a broader audience. If I get a mentor in time, I will make a detailed proposal for the mentors/community be able to understand better the idea.
Actually it works the other way around, you’re supposed to write and submit your proposal at this stage, so that potential mentors can comment on it and give you feedback, and decide whether it’s feasible and if they want to mentor it. Please, follow the guidelines I linked to you to submit a proposal, and share it publicly also in GRASS mailing list for developers to comment.
Good luck!
Thank you very much,
Evandro Carrijo Taquary
Federal University of Goiás
