Blog · 9 MIN READ
No More Early Adopters: Data Solutions for Visual Inspection Robots
Posted on April 24
No More Early Adopters: Data Solutions for Visual Inspection Robots
(Note: Larry Barnard is a HUVR advisor with over 30 years in the field)
A picture is worth a thousand words. In many cases, ten of those words are “What are we going to do with all these pictures!?” This article will take a brief look at the data management side of commercial grade robots; specifically drones, or small unmanned aircraft products, but the principles are the same for any inspection tool that delivers large amounts of data (or large files) that needs to be aggregated.
The struggles of early adoption
For organizations choosing to operate robotics in-house, the journey typically starts with the shiny new gizmo: the hardware. Unfortunately, as with many decisions driven by new tech, it turns out that the cool robot is just half (or less) of the comprehensive solution. Many early adopters choked on the volume of data that their robotics generated. Teams quickly realized the need to develop or purchase a sustainable, scaleable and contextual data solution.
Depending on the size of the organization, early data solutions were akin to a photo-sharing site at best; at worst, they relied on using the corporate One Drive solution or email attachments. Often data was hidden in a series of folders on a partitioned drive, necessitating a lengthy search of the entire directory once you find out what the folder name is (assuming strong naming conventions were used from the start). Given the volume of data, individual file sizes (high res photos or video), choked bandwidth, variable latency and bulky back-end computer processes, efficiently sharing UAV/drone-collected data was an exercise in frustration and inefficiency, leaving data extremely siloed with no clear connection to the rest of the inspection data or maintenance management system.
Regardless of the type of robot (flying, swimming, crawling, fixed), once an organization decides to invest resources in robotics solutions, it is imperative to invest in a data solution in parallel to the robotics hardware being selected. For all the amazing capabilities and benefits that robotics hardware can deliver, those benefits will ultimately fall short without a reliable data management workflow.
OK, so what does this solution look like?
A quality, well-supported, comprehensive data solution for media and measurements to manage your drone and robotics data is indispensable for delivering on the promise of robotics. It is also worthwhile to future-proof your robotics inspection efforts by selecting a comprehensive data solution that can ingest non-robotic image files. This includes data from mobile devices, inspection cameras, ultrasonic thickness measurements, optical gas imaging data, or even digital x-ray data.
A comprehensive data solution will not only provide state-of-the-art cloud storage (compliant, low-latency, scale, solid uptime, etc.), but will also provide an intuitive interface to:
- manage the data (imagery, ultrasonic thickness, 3D files, etc.)
- enable annotation
- Provide a means to track status of an inspection from request, to finding, to resolution
- allow local naming conventions
- have role-based data views
- provide a variety of analytics to find problems before they find you
- Allow import of existing inspection documents, and export to legacy inspection and maintenance applications
- Ability to generate configurable reports or other communication documents
A solution to handle your robotics-generated data is an absolutely necessary step in the evolution of your robotics inspection program. The above might be seen as the minimum requirements, but a more forward-looking approach to managing your image data may include some or all of the following features and functions;
- A secure platform, including the ability to isolate contract work to just the contractor being used. Keep the remaining company data isolated.
- A compliant, performant, scalable, secure and well funded/supported data solution provider.
- Import and integration of existing inspection documentation to support the specific inspection image set (history, TMLs, digital xrays, etc.).
- Ability to annotate your data (images!) from a lightweight mobile app in the field, or from your desktop workstation and web-based application.
- Artificial Intelligence/Machine-learning algorithms that perform the first-pass analysis of images to reduce/eliminate requirements for employee time to scan the data manually.
- Standardized reporting to a contract service provider, a site, or an entire enterprise. Same format, information, and output data. Ability to somewhat customize reports based on asset type
- Onboard analytics to help visualize the data (leading, lagging indicators, etc.)
- Connectivity between your legacy maintenance and inspection software and your new robotic data: API’s to allow you to export data into your legacy systems (maintenance, inspection, etc.). Use your old work order system, and have your new robot data solution help fill in some of the needed information fields. This approach can also work to feed near real-time data to your digital twin.
- Mapping applications to help visualize locations and alerts. Especially important for numerous but similar assets (power poles, tanks, etc.).
- Personalized dashboards based on role (drone pilot, engineer, inspector, team lead, big boss, aggregate sites, corporate level, etc.). One size does not fit all.
- Ability to upload supporting documentation (drawings, checklists, permits, JSA’s, etc.).
- Ability to apply specific documents/ forms to specific tasks.
- Ability to provide bulk upload of information such as facility assets, inspection types, equipment names, product types, facility locations and reduce the manually intensive steps to set up your new system.
That seems like a lot. Do I really need all that?
You will never have less data or fewer images collected.
Invariably, when an organization starts using drones (UAV’s) or other robotics, the benefits of performing inspections with them are immediate, impressive and almost always lead to novel ways to use the robots. These ideas will come from creative practitioners, project managers, engineers, inspectors, and other boots-on-the-ground folks who have seen what is possible, and perhaps have experienced how safely and efficiently their data can be collected. The incredibly high-resolution photos, videos, and infrared files consistently shock and impress the person receiving this data. People will want to utilize, and find new ways to utilize the robots for data collection as part of their jobs/tasks.
You will want those same efficiency gains from the data and to provide the ability to:
- enable the workforce to request an inspection (due dates, description of the need, compliance vs. reliability based, identification of specific asset and specific target on the asset).
- Enable two-way sharing of data between your image based inspection application and your legacy software systems already in use (compliant import/export).
- Easily move large files from hardware to software, from cloud to desktop, from project to project, and customer to customer safely and quickly.
- Enable specific data to be exported to work order management systems to better communicate issues/tasks to planners and maintainers.
- Provide a safe and secure storage solution for potentially sensitive image based data of your facility assets.
Conclusion: The age of early adoption is over
Since the heavier use of drones and robotics started around 2018/19, there is no longer an excuse to focus on the tool’s capabilities and worry about the data later on. Later on is now. Drones in a wide variety of industries are commonplace and accepted: no more early adopters.
New workflow applications provide an opportunity–a sort of reset to how things have been done. This is a great opportunity to select a solution that can work with all the cutting edge data collection hardware and provide a solution to employees and contract partners with less limitations and much more functionality and efficiency.