Skip to main content
All CollectionsInspect Release Notes
Inspect Release Notes: July 2024
Inspect Release Notes: July 2024

Several new and improved AI models, including cutout and insulator composition, light type classification, and estimated vegetation volume.

B
Written by Ben Candea
Updated over 6 months ago

Our teams have been hard at work over the last few months, releasing several new and improved AI models, adding new features to our Cloud application, and updating our Inspect Edge system.

If you’re interested in validating a new model (or any others) with us, reach out to our team at [email protected]!

Introducing New Product Categories

We’re slightly rebranding our AI models by grouping them into product categories that better align with how utility operations teams think about these use cases. Going forward, you’ll see us reference these product categories when we communicate new updates and improvements in release notes.

  1. Power distribution intelligence: includes our distribution asset inventory and inspection models, which lower inspection costs and increase situational awareness of your assets.

  2. Lighting intelligence: includes our lighting audit products, such as classifying fixtures, light emission types, and detecting NEMA labels.

  3. Vegetation intelligence: includes our vegetation volume estimation model, supporting maintenance corridor condition assessments.

  4. 3rd-party and joint-use intelligence: includes our models for attachment inventory, clearance, and make-ready workflows.

AI Model New Releases + Enhancements

Power Distribution Intelligence

New: Cutout and insulator composition (beta)

This new model classifies cutouts and insulators as either porcelain or polymer, allowing you to inventory where the different types are located. Polymer insulators better withstand extremely cold or highly humid climates than porcelain, and are almost impervious to cracking.

New: Pole composition (beta)
This new model detects the pole materials of distribution assets, such as wood, steel, and concrete. We’ll soon add a filter in Inspect Cloud so you can easily understand where upgrades may be needed.

Improved: Pole lean
This updated model better segments poles from other objects in an image for more accurate measurements and improves the angle estimation by ~10%!

Lighting Intelligence

New: Light emission type classification
This new model classifies light bulb types by LED or non-LED. Surface lighting asset geolocations by emission type to make it easy to conduct lighting audits and plan LED upgrades.

New: NEMA label detection + extraction (beta)
This model identifies NEMA labels on the underside of cobra head lights and reads the wattage information, refreshing your lighting inventory.

Vegetation Intelligence

New: Vegetation volume estimation (beta)
This new model quantifies the vegetation surrounding a pole within 20 and 50 feet, supporting tree trimming and vegetation management efforts. A key focus of electric utilities' extreme weather mitigation efforts, this capability helps to ensure reliability and safety on the grid by prioritizing which trees require trimming the most, while helping to validate that trimming work was performed to specification.

Inspect Edge

Improved: Automatic upload of Edge data to the Cloud
Images and related metadata are now sent to the cloud within minutes of capture, decreasing this time by over 50%, by employing advanced selection and compression techniques.

Inspect Cloud

New: Filter by pole lean
Leverage Inspect Cloud to filter for poles leaning severely and quickly mobilize your crews to repair them.

Coming soon: Inspect Cloud v2

We’ve listened to your feedback and have been hard at work coding v2 of Inspect Cloud! This new version adds dashboards, vehicle statistics, and more. We’ll let you know when this version is ready to use.

Other Updates

New: Single-tenant architecture in the Cloud
This update isolates customer data and related processing to its own environment, enhancing data security and protection in the Cloud.

Did this answer your question?