Airvision is a platform where computer vision models for drone applications can be developed easily. This way, companies can make use of the power of AI easier, quicker and cheaper.
The typical computer vision training pipeline contains 6 steps, which are descriped in the schema below. All of these steps can be executed in the Airvision platform, in a way that it perfectly fits your specific project requirements.
Steps 1 to 4 are used for training a computer vision model. In these steps you train a specific model for your specific situation. This is a process that you have to go through only once.
Once you’re model is developed and deployed, step 5 makes sure that new images can be analyzed automatically within seconds by the model that was developed in the training step. Step 6 helps you to improve the performance of the model even more, by providing an easy interface for giving feedback on the model results.
(1) Data import
Upload your drone footage to the Airvision platform. Airvision can provide API connections with the most used flight management systems to make synchronisation of data as easy as possible.
(2) annotation
Annotation is the process of manually labelling images in a way that an AI model can learn from it. We have multiple ways to make this process as fast as possible.
(3) model training
An AI model is trained that is constructed conform the provided data, annotations and user input. The Airvision platform automatically trains state-of-the-art computer vision models for your specific case.
(4) Deployment
After the model is trained, it will be deployed in the cloud with the press of a button. This way, you don’t have to worry about GPU’s, cores, containers or other technical stuff.
(5) INFERENCE
Model inference is the proces of running live data into an AI algorithm to calculate an output. This way, your model is able to generate predictions in the blink of an eye. The model inference is automatically set up by the Airvision platform.
(6) feedback
Once the model is in production, you can give feedback on the models output to improve the performance even more. This way, the model learns from it’s mistakes and becomes better over time.
USE AirVISION FOR YOUR COMPANY?
Contact Airvision today to make use of the power of AI for your use case fast and easy.