Iron Ore Detection
The client is a firm who builds economic indicators using machine learning and remote sensing technologies. The client wants to build a solution which can detect iron ore volumes from high-resolution satellite imagery.
- To develop a computer vision architecture to obtain the iron ore volume from high-resolution satellite imagery.
- To suggest methods to productionalize the process starting from raw data to get iron ore volume.
- All required information about data are collected from the client .
- Downloaded the images from sentinelhub and understood the data, image extraction process and image annotation.
- Researched the techniques to annotate the images.
- Researched and applied the possible techniques or architecture can be used to detect iron ore.
- Leveraged deep learning classification model to classify images with or without iron ore.
- Leveraged deep learning segmentation model to segment the iron ores.
Iron ore volume can be detected and calculated using developed solution rather than manually measuring the iron ores.