Drone users… It’s time to leave post-processing in the dust
While this plethora of insight can be incredibly valuable for a business, it can also be very overwhelming to understand and make use of. The sea of data either leads to a technician spending hours or even days behind a computer screen digging through images to find relevant information for analysis, or transferring the data to a post-processing program that will also take hours to analyse the information. This just won’t cut it. We live in an instant society where we expect results and insights immediately. Post processing will eventually become obsolete, real-time analysis is the way of the future.
Drones are seemingly making headlines everywhere in recent times, in what can only be described as a drone revolution. This kind of rapid growth in innovation is reminiscent of when smartphones exploded onto the scene, and now as a mature technology, billions of people carry a smartphone with them every day, be they consumers or professionals. When it comes to commercial drones, this widespread adoption is being held back as other supporting technologies try and play catch up.
There is no doubt that drones have greatly improved aspects of a multitude of business workflows. We have all heard the benefits gained from using them, but are you familiar with the painstaking post-flight process that follows?
It’s only as the hype of drones slowly starts to settle that businesses are beginning to realise that the gains they are making in the field, are being lost behind a computer.
Take a wind turbine inspection as an example. Upon completing an inspection of the turbine, the technicians return to the office, place their memory card into their computer to begin the slow and laborious task of combing through the thousands of photos that were taken with the drone to identify faults. Alternatively, if the business has invested in post-processing software, they could be waiting for 8 hours or more just for a 3D representation of their turbines.
This lag between inspection and analytics can be attributed to the fact that we have highly advanced flying robots, yet we are still attaching generic payloads such as handheld DSLR cameras, or other payload technologies that weren’t specifically developed with drones or data analysis as their focus. Not to mention, the software loaded onto laptops or desktop computers are overloaded by the huge amounts of data they are expected to process.
Where do we go from here?
Popular post-processing software programs do a great job of crunching the large amount of data coming in from drone operators, but it takes time. Many companies are beginning to offer cloud-based solutions that seem to be a great method moving forward. The only issue is it still requires an individual to upload a large amount of data once the mission is over, as well as further analysis to be completed by a technician once processing is done.
As an R&D lead company, real-time processing in drones is something that Aerialtronics has identified as a priority and feels strongly about transforming from a theoretical solution to a tangible option for businesses. We are currently in the final-stages of developing an intelligent camera with NVIDIA that combines artificial intelligence and geo-referenced object detection capabilities and will enable users to collect and analyse data in real time — before their drone even touches the ground. Using the wind turbine example from earlier, this would mean the technician inspecting the turbine could receive an instant notification regarding any fault or damage as it is detected by the camera or annotated by the operator. Alongside this, its GIS data would be automatically annotated, saved and pushed out, so the individual making repairs can easily locate it again. This approach works much more efficiently so by the time a mission is over, the technician already has an analysis and can act upon them immediately.
This results in even less downtime than with current drone methods and potentially provides an even higher ROI for businesses. Instant results with actionable solutions.
In addition to this real-time analysis tool, Aerialtronics is also building cloud-based analysis into the Altura Zenith drone by partnering with IBM Watson IoT. By building it into the drone itself and incorporating machine learning, we can offload data, analyse and make predictions in a fraction of the time that post-processing software can do. What if you’re in an area without a reliable internet connection and you need to process gigabytes of data? This is where the real-time processing system is necessary. While some cloud solutions offer a watered-down version of their desktop software offline, it does not allow you to get the complete output that you’re after. This is why a combination of real-time processing and cloud-based processing is necessary — because it provides the user with a full-scope and in- depth analysis of the data collected from a drone. It is the complete solution.
