Software Design
Design and develop softwares solutions. I use mainly Python, jupyter, docker and the cloud.
Machine Learning Engineering
Good knowledge of applying machine learning to signal processing. It is completed with a good level in maths.
Hyperspectral Images Analysis
Manage the setup for analysing Hyperspectral data sets. The setup can be in a cloud. The service include python/cython programming, (maths based) algorithms and software design, use of numpy, scipy, matplotlib, jupyter notebook, scikit-learn and many more python libraries or others like BLAS. Cloud computing is a solution to process large data sets. The cloud computing service use virtual Linux machines or Docker containers.
Benefits
Less time spent analyzing images. You get the results faster.
We manage the setup for you! Including R.D.
Based on existing hyperspectral softwares libraries. You do not have to reinvent the wheel.
Features
For hyperspectral processing
Analyse your images using Machine Learning and avoid complex modeling.
Python power
Easy to use and yet speedy tanks to numpy, multi-processor, MKL and GPU.
Run on jupyter notebook
You can experiment and share your works.
Cloud computing
Hyperspectral needs computing power. We manage it for you.
Scikit learn
Wrap scikit learn to make hyperspectral images analysis more powerfull.
Use matplotlib and ggplot
Use the powerful matplotlib and plotnine (ggplot) graphic libraries to output the results.
Gradient Boosting
Gradient Boosting is very effective on hyperspectral images analysis. The library integrate XGBoost and LightGBM.
Signals processing algorithms
PySptools add many signals processing algorithms specific to hyperspectral like SAM, N-FINDR, unmixing and many others.
Read the doc
The system is well documented and will improve with time.