We are excited to add Storm Intelligence to our suite of solutions. During recent hurricanes, we successfully deployed our Inspect system to capture and analyze data in near real-time, enabling faster damage assessments that helped utilities reduce their estimated time of restoration (ETR). We’re committed to supporting faster recovery efforts to help those profoundly affected by these extreme storms.

New: Transformer damage detection
This new model detects transformers with rust/corrosion and oil leaks with 90% accuracy.

New: Animal guard detection
This model detects animal guards on important components, including cutouts and arresters, with 85% accuracy.

New: Asset tag detection + text extraction
Automatically detect and extract asset tag information for poles and bypass switches with 86% accuracy. This new model will make it easier to match Noteworthy-detected assets to existing customer GIS records.

Improved: Component detector v2 (beta)
The latest version adds new components, including capacitor banks, reclosers, control boxes, riser heads, bushing insulators, and spool insulators, increasing the total number of components identified to 21.

New: Lighting asset tag detection + text extraction
Similar to our asset tag detector for power distribution, this model automatically detects and extracts text information from lighting asset tags.

Improved: Light fixture type classification
This updated model identifies the following types of light fixtures with an average accuracy of 83%: street lights, area lights (sub-types: security lights, floodlights, other), pendant lights, post-top (sub-types: colonial and acorn), and high-mast lights.

Improved: NEMA label wattage and LED text extraction
This model detects NEMA labels on the underside of lights and extracts wattage information with 88% accuracy. Additionally, it extracts text reading “LED.”

New: Vine detection (beta)
Vines growing on poles can pose a significant outage risk. This new model detects vines on poles or associated pole-top components.

New: Attachment counting (beta)
This model detects attachment brackets, separates cables and cable bundles, and counts the number of attachments within the telecom space. This helps validate third-party inventories and identify unauthorized attachments.

Inspect Cloud 2.0 is now available in beta! This version offers a faster user experience, cleaner UI, and new features that we're excited to share (with more to come).

New: Real-time updates
Last quarter, we introduced over-the-air (OTA) data uploads, where images and asset information are automatically transferred from Inspect Edge devices to the cloud. This quarter, we’ve taken this feature one step further. Users can now see new assets, geolocation coordinates, images, and model results in Inspect Cloud in real-time as the devices are in the field.

New: Map-centric view
The new Asset Explorer page always includes a map with all asset markers, making it easier to navigate from asset to asset based on location or street.

New: Customer region matching
In addition to matching existing customer asset GIS records, we can now match Noteworthy-detected assets to user-specified service territories/regions.

New: Asset filter functionality
Users can now build customized, multi-condition queries that support both “AND” and “OR” conditions using the improved filter UI.

New: Consolidated asset details display
Model results are now consolidated into one section on the Asset Details page, so users can view all associated AI insights in one place.

Following the release of Inspect Edge 2.0 last quarter, this quarter’s efforts focused on optimizing performance, enhancing data precision, and streamlining system processes to ensure smoother operation and greater reliability.

Improved: Asset tracker performance
Enhanced object tracker better compensates for rapid camera motion, improving the accuracy of asset tracking across frames, enabling better geolocation and fewer missed assets.

Improved: Better file handling
Faster, more efficient file uploads to the cloud, with local archiving of uploaded files for better organization.

Improved: More accurate ignition status handling
Improved reliability of vehicle ignition status readings to prevent incorrect Edge shutdowns.

Improved: Memory management
Reduced memory usage and doubled metadata cache size to prevent data loss.

Improved: System stability
Several bug fixes have been implemented, including resolving issues with partial asset capture and periodic camera failures.