Promotion of innovative DeepFlow software on foreign markets in order to boost the company’s exports
The goal of the project was to promote DeepFlow software on foreign markets. This is innovative Earth Observation (EO) software which enables its users to utilize deep neural networks effectively in order to analyse satellite images. Owing to the advanced technology of deep learning, the machine is able to analyse large datasets and helps discover similar patters in hundreds or thousands of photos. However, these data require preparation and further processing, which can be achieved with the use of deep networks and DeepFlow image correction mechanisms. Furthermore, owing to the integration of DeepFlow with Copernicus software and DIAS platform, it is possible to access free satellite images from Sentinel and Landsat satellites. Thus, DeepFlow functions comprise:
- access to data from Sentinel and Landsat satellites
- network training with large training sets and finding similar patterns in thousands of new input images
- pre-processing and further transformation of images.
DeepFlow is a tool designed for huge datasets, especially satellite images. Owing to the advanced technology of deep learning, the machine is able to analyse large datasets, thus reducing the time necessary to extract information from the observed areas.
The learning process is automated and aimed at minimization of manual labour, while maintaining the universality of the solution. DeepFlow offers an opportunity for integration with Copernicus software which supplies ca 9 petabytes of free data per year. However, these data require further processing so that recurrent patterns could be found.
Your project description
To add a photo, description, specify a location or attach other materials about the project, you should fill out the form below. After clicking the “send” button, a message will be sent to you e-mail asking for confirmation. You must accept to submit the form. The EU Dot Map Guidelines and consent to the processing of personal data. materials will be sent to the editor of the site, who will publish it on the website after verification test.