Digitization of everyday things has introduced a proliferation of incoming data, streaming in from every angle. Smart-Edge simplifies and accelerates data-driven decisions by relaying only actionable insights.
Image Classification and Analytics
Analog gauges designed for legacy equipment can often be costly to modernize. My algorithms are trained to convert analog gauge data into digital streams of data for existing and remote pipeline infrastructures.
Image-based solutions solve many of the manual issues associated with analog devices.
Some of these issues include who’s reading the meters, how they are being read, and how much they cost to maintain.
Getting Accurate and Consistent Readings Is Not Easy
Who reads the meters?
- Consistency across operators
How is it read?
- Parallax errors + environment factors
How much does it cost?
- Money matters
Simulation Saves Time and Money
My crew developed a state-of-the-art solution that uses simulation to reduce the need for hundreds of images per gauge (which can be expensive to obtain) down to fewer than a handful, dramatically accelerating the training process. This simulation approach rapidly and accurately trains on any analog device to generate a digital signal.
These models are then deployed to the edge (on camera). This enables on-device processing, reducing cloud costs as each camera transfers only the digital results of interest. Using a federated process, these devices also continue to learn.
This Smart-Edge Solution:
- Converts analog gauge data to provide near real-time digital outputs
- Allows for swift implementation thanks to its unique training approach
- Extends the life of aging manufacturing equipment, reducing cost of ownership
- Detects anomolies and creates alerts to focus attention only on 'out-of-spec' conditions … priceless