top of page

Manifold Analytics Pilot Program

Crowdsourced Energy Distribution Inpsection

Overview

Manifold Analytics is looking for partner relationships to test a new pilot program utilizing commercial grade cameras and incidental collections to provide detailed analysis to electricity providers and electrical cooperatives.  Since the majority of an electrical company’s distribution network resides along local roads and highways, there is a significant opportunity to collect continuous metrics on the distribution infrastructure with a minimal cost to utility companies or cooperatives.

PoleDetectionOverview.jpg
PoleMultiView.png

This is an example of a utility pole captured by different vehicle cameras on different days. Note: The lens distortion of the wide angle lenses has not yet been removed.

Data Collection

Manifold will supply a camera system to several drivers and have them passively collect around the coverage area.  This data is downloaded on to Manifold servers each day and processed on our internal cloud to extract information about infrastructure and accumulate a detailed database of the utility’s infrastructure. 

To keep cost and complexity down, Manifold is evaluating various consumer grade camera systems in the < $1000 range with a number of different modalities.

CameraExamples.png

Full Spherical Capture

This video was captured with a GoPro Fusion 360 camera in a full spherical capture mode at 30 FPS.

Hemispherical Capture

Traditional Capture

This video was captured with a Rylo360 camera in hemispherical capture mode at 25 FPS

This video was captured with a gimbal stabilized DJI Osmo camera at 30 FPS

Object Classification 

One of the critical aspects of the roadside inspection approach is neural network-based object detection and classification.   The initial detections from classification, helps determine what other algorithms will be run downstream.

ClassificationExamples.png

These examples show the object detection of basic utility components such as poles, cutouts, and transformers. 

Rogue Connections

In addition to power, most utility poles rent space to secondary providers like cable, phone, fiber optics, etc. 

The communication between utility companies, energy cooperatives, and these secondary providers 

LineDetection.jpg

In addition to power, most utility poles rent space to secondary providers like cable, phone, fiber optics, etc. 

The communication between utility companies, energy cooperatives, and these secondary providers 

Line Sag

Pole Lean

In addition to power, most utility poles rent space to secondary providers like cable, phone, fiber optics, etc. 

The communication between utility companies, energy cooperatives, and these secondary providers 

Vegetation Encroachment

In addition to power, most utility poles rent space to secondary providers like cable, phone, fiber optics, etc. 

The communication between utility companies, energy cooperatives, and these secondary providers 

Post Storm Assessment

In addition to power, most utility poles rent space to secondary providers like cable, phone, fiber optics, etc. 

The communication between utility companies, energy cooperatives, and these secondary providers 

Geospatial Tools

Manifold KML allows a user to interact with the collected data and view information from a large system overview, down to sub-inch views all within Google Earth.

Rail KML 1.jpg
google-earth-icon.png
Google Earth KML
Clicking will Download KML File

Note: The imagery in this example KML has been significantly

reduced in resolution to minimize download time.

Rail Gage Measurement

Frame00214.jpg

The Manifold gage measurement algorithm can determine the distance between the rails on a track to within ¼”.  These measurements are made using a high end photogrammetric aerial camera and Rosetta image processing algorithms.

 

 

Rail Gage Inspection Examples

Frame00129.jpg
Frame00120.jpg
Rail-Gage

Rail-Gage

Watch Now

Crosstie Orientation

Crosstie orientation analysis evaluates the rail crossties for signs of movement relative to the track.  This allows the rail company to inspect for issues like crosstie degradation, track warping, and ballast settling without having to send a technician to inspect.

The green box ends at the end of each crosstie indicate the tie was detected and is within spec.  The ties that are highlighted yellow or red indicate ties that require repair.

Frame00214.jpg

 

 

Rail Crosstie Orientation Inspection Examples

Frame00216.jpg
Frame00575.jpg
Crosstie Orientation
Play Video

This video shows an older iteration of the crosstie orientation algorithm with regional statistics and mapping overlays applied.

Crosstie Spacing

Frame00308.jpg

Crosstie spacing analysis evaluates the rail crossties for signs of movement relative to each other. the track.  This allows the rail company to inspect for issues like crosstie degradation, track warping, and ballast settling without having to send a technician to inspect.

The spacing measurements (shown between the ties) are colored to indicate the type of spacing issue. Red and yellow indicate spacing that is greater than expected and cyan and blue indicate spacing that is narrower than spec.

 

 

Rail Crosstie Spacing Inspection Examples

Frame00527.jpg
Frame00892.jpg
Track-CrosstieSpacing1
Play Video

This video shows an older iteration of the crosstie spacing algorithm with regional statistics and mapping overlays applied.

Crosstie Length

Manifold rail inspection algorithms measure the lengths of the crossties to detect shortened or broken ties.  Ties that are abnormally short can represent locations where crosstie ends have broken or become buried by ballast material.

This tie has been flagged as too short.  On closer inspection, the end of this tie has broken and begun to fill in with ballast material.

Rail KML - Length.jpg

 

 

Rail Crosstie Length Inspection Examples

Frame00345.jpg
Frame00573.jpg
Frame00755.jpg

Crosstie Split Detection

TieCrackDetection-00598-(11)FullImageCra

Manifold rail inspection algorithms inspect the surface of rail crossties to determine if they are cracking or completely split.  Additionally, they are inspected for ballast material rising through the crack, indicating a complete break.

The ties are flagged in regions where the splits are detected. Regions of sufficient length and width are highlighted red.

TieCrackDetection-00598-(10)CrackPgons14

Baseplate and Spike Analysis

The baseplate algorithm detects the positions of the plates at the end of each tie.  The spikes are then analyzed based on a CAD model to assess if they are installed correctly.

CloseUp.jpg
SpikeDetection-00500-(03)CorrHoles15.jpg

The blue boxes show base plate detection.  The green circles indicate that a spike was found at a hole location and a red ‘x’ indicates that the hole is unused.

Frame00501.jpg

Vegetation Encroachment

Frame651.jpg

The vegetation encroachment algorithm scans the track surface for signs of vegetation that is growing inside the right of way.  This algorithm detects both trees/shrubs that are growing into the right of way, as well as significant growth through the ballast .

 

 

Rail Vegetation Encroachment Inspection Examples

Track-RightOfWay1
Play Video

Remote Asset Detection

Manifold uses geofencing to tag regions of interest along the flight path.  Images collected inside a geofence are provided as a high-resolution data package that can be used for visual inspection of a hard-to-reach asset.

Frame00775.jpg

Aerial Anaglyph

Anaglyph-SwitchPoints.jpg

Manifold combines sequential aerial images (or video frames) using precise geospaital projections to generate stereo viewable imagery. .

bottom of page