Computer Vision News - July 2016

CVN: But you do not need computer vision technology per se to fly a drone, do you? Modali: Not at all: the hardware needed to fly a drone is already available. With those pieces of hardware you build a drone and you can fly to take pictures. That alone does not require computer vision knowledge. But if you want the drone to do something meaningful, then you have to think at vision technology . CVN: Is it in this area that our industry can intervene and create significant added value? Modali: Absolutely! Vision is the space where you can come in and say: with vision technology you can do several things that a short time ago were not possible. That is the service that you are able to create to add value on top of the drone and that is a monetization opportunity. Therefore, with vision technology you are solving two issues: the first is that you are getting more data because you are in closer proximity to what you are inspecting; the second is that you rely less on piloting skills and reduce piloting errors, so that you are effectively reducing the total cost of operation. And when you reduce the cost of operation you generate added value and a monetization opportunity. CVN: What are the main algorithmic techniques which are needed to get there and what are the challenges that need to be dealt with in order to be successful? Modali: I think that in terms of algorithms, the sky is the limit: it really depends on your imagination finding out how you can apply a certain algorithm to solve a given problem. Typical problems are having a stable flight, hovering in place and being able to do such things as getting closer to the inspection sites, flying beyond the line of sight and inspect indoors where GPS signal is not available. Computer vision can help in all these areas . “ In terms of algorithms, the sky is the limit ” Computer Vision News Guest 17 Drone with camera, ready for aerial photography Guest