Computer Vision and Machine Learning are core elements of the Artificial Intelligence Spectrum.

From industrial automation to increasing capability of autonomous systems, computer vision is a powerful tool for any robotics application.

By analyzing video and images we can work out what unique features an object has that differentiates it from the rest of the frame. This is a relatively easy task in an indoor environment, but increases in complexity outdoors or in an unknown environment. Tracking a moving target from a moving camera (e.g. mounted on a UAV) increases the complexity of the problem to its extreme.

Despite the challenges, we are happy to announce we have solved detection and tracking from UAV footage and are now exploring the applications of the capability in the agricultural sector.

With Machine Learning a computer can be ‘taught’ to recognise an object or pattern. Further, we are currently implementing Machine Learning for data analytics and predicting online consumer interaction.

By further training a UAV with machine learning, we can teach it to recognise different objects or animals that it detects and tracks using computer vision techniques. This is a powerful capability that can tie in with our custom autonomous control systems to achieve smarter UAV autonomy.