Open navigation

Power & Utilities Analytics

Delair.ai provides a range of powerful analytics specifically designed for the Power&Utilities use cases, and leveraging the variety of supported sensors: LiDARs, thermal and RGB cameras.

These analytics adress the requirements of :

  • Classifications, notably as a part of asset digitization operations
  • Conductors vectorization, as a part of asset digitization operations and vegetation management
  • Vegetation Encroachment : assessment in the vicinity of power lines to manage pruning operations
  • Solar Plant Thermal Inspection

1. Power & Utilities Classification Analytics

According to the kind of classification analytic ordered, more or less classes will be delivered in the platform, from standard ones to custom classifications.

1.1. Basic LiDAR Point Cloud Classification

1.1.1. Usage

This analytic provides a classification of objects from a point cloud generated from LiDAR data. It classifies the power line asset and surrounding environment in 5 classes (4 standard + 1 other):

  1. Ground
  2. Vegetation
  3. Poles
  4. Conductors
  5. Other Objects Class = All other than described in 1 to 4 above

1.1.2. Required Input

A point cloud already pre processed (step previously carried out with a dedicated LIDAR sensor's manufacturer software) in order to obtain from flight raw data a cloud point in one of the formats listed below. Please note that raw data point cloud processing must include the correction of GPS laser points coordinates with IMU information, and also a first quality check for noise reduction on point cloud. The supported input formats and standards are :

  • Las 1.2 (*.las)
  • Las 1.4 (*.las)

1.1.3. Viewing Classes

Select first a 3D view from the project main window :

Once the classification results have been made available in the project, viewing classes is enabled by first selecting the classes-carrying Layer, displaying its properties in the right panel and selection Classes from the Material section. In  the Example hereafter, Lidar is the layer that has the classification built. The point cloud will turn into colored mode according to the classification chart.

By opening the Classification drop-down tab, the classification chart can be browsed. In the example herafter, we have for instance Grounclass in brown, Medium vegetation class in medium green, and High vegetation in light green.

1.1.4. Deliverables

For each 5 km of power line, a Basic Lidar "Clas" file named with the prefix B_LiDAR_Cl, plus a code composed by adding to power line's unique code (provided by the user and with a reference length of 9 digits) two more digits for defining the portion of line considered.

For example : a power line of 48 km will be delivered in 10 files that will be distinguished at code level by the last two digits while the prefix plus the first 9 will remain the same. Each file contains one georeferenced classified cloud point featured by 4 classes plus a fifth one to define other objects. For double circuit lines placed on the same poles and therefore interested by the same corridor, we have only one file delivered each 5km. Files are delivered in Las 1.2 standard (*.las).

1.1.5. Export

From the Download section, point clouds can be exported from the All other files section, for use in third-party software.

1.2. Advanced LiDAR Point Cloud Classification

1.2.1. Usage

This analytic provides a classification of objects from a point cloud generated from LiDAR data. It classifies the power line asset and surrounding environment in 9 classes (8 standard + 1 other):

  1. Ground
  2. Vegetation
  3. Poles
  4. Conductors
  5. Buildings
  6. Crossing Lines
  7. Roads
  8. Railways
  9. Other Objects Class = All other than described in 1 to 8 above

1.2.2. Deliverables

For each 5 km of power line a Advanced Lidar "Clas" file named with the prefix A_LiDAR_Cl, plus  a code composed by adding to power line's unique code (provided by the user and with a reference length of 9 digits) two more digits for definying the portion of line considered. For example, a power line of 48 km will be delivered in 10 files that will be distinguished at code level by the last two digits while the prefix plus the first 9 will be the same. Each file contains one georeferenced classified cloud point featured by 8 classes plus a ninth one to define other objects. For double circuit lines placed on the same poles and therefore interested by the same corridor, we have only one file delivered each 5km. Files are delivered in Las 1.2 standard (*.las).

1.2.3. Required Input

A point cloud already pre-processed (step previously carried out with a dedicated LIDAR sensor's manufacturer software) in order to obtain from flight raw data a point cloud in one of the formats listed below. Please note that raw data point cloud processing must include the correction of GPS laser points coordinates with IMU information,  and also a first quality check for noise reduction on point cloud. The supported input formats and standards are :

  • Las 1.2 (*.las)
  • Las 1.4 (*.las)

1.3. Basic RGB Classification

1.3.1. Usage

This analytic provides a classification of objects from a RGB dataset. It classifies the power line asset and surrounding environment in 5 classes (4 standard + 1 other) :

  1. Ground
  2. Vegetation
  3. Poles
  4. Conductors
  5. Other Objects Class = All other than described in 1 to 4 above

1.3.2. Deliverables

For each 5 km of power line a Basic Lidar Clas file named with the prefix B_RGB_Cl, plus  a code composed adding to power line's unique code (provided by the client and with a reference length of 9 digits)  two more digits for definying the portion of line considered. ( eg: a power line of 48 km will be delivered in 10 files that will be distinguished at code level by the last two digits while the prefix plus the first 9 will be the same). Each file contains one RGB georeferenced classified cloud point featured by 4 classes plus a 5th to define other objects not included in the first four. For double circuit lines placed on the same poles and therefore interested by the same corrior we have only one file delivered each 5km. Files are delivered in Las 1.2 standard (*.las).

1.3.3. Required Input

A 3D georeferenced cloud point obtained through stereophotogrammetry. As a reminder please consider that this technique estimates the three-dimensional coordinates of points on an object employing measurements made in two or more photographic images taken from different positions. Common points are identified on each image. A line of sight (or ray) can be constructed from the camera location to the point on the object. 

It is the intersection of these rays (triangulation) that determines the three-dimensional location of the point.

Please note that raw data cloud point processing  must include correction of GPS laser point coordinates with IMU info and also a first quality check for noise reduction on cloud point.

1.4. Custom Classification

Delair can provide customized classification analytics upon your specific request, for your own use case, especially if other classes of objects or more than 9 classes are required, and with the level of privacy required for the application.

  • If you are not already a delair.ai user, please contact us here.
  • If you already use delair.ai, please contact your Sales Representative at Delair to discuss your requirement.

2. Vectorized Conductors

2.1. Usage

This analytic provides a 3D georeferenced vector file for power lines catenaries. More in detail, in the same file, we have a 3 vectors (one for each phase of the line).

2.2. Required Input

This tool uses as input a classified point cloud; the analytic works on point clouds classified as "conductors" and provides as output a 3D georeferenced vector file.

2.3. Viewing Vectorized Conductors

Select the 3D View in project main window, then from the Layers panel, enable the one that contains the results of vectorization. In the example below, user named it Digitized Conductors :

2.4. Deliverables

For each 5 km of power line a vectorized conductors file named with the prefix VCond, plus a code composed by adding to power lines unique code (provided by the user and with a reference length of 9 digits) two more digits for defining the portion of line considered.

For example a power line of 48 km will be delivered in 10 files that will be distinguished at code level by the last two digits while prefix plus the first 9 will be the same.

Each file contains three (one for each phase of the power line), 3D georeferenced vectors with a maximum length of 5km.

For double circuit lines we have two files each 5 km.

Each file is delivered in two formats (.dxf and .shp).

3. Vegetation Encroachment Analytics

3.1. Usage

This analytic provides a detection of vegetation surrounding power lines with advanced features to visualize the Risk Map. The Risk Map layer highlights crtitical zones where vegetation is detected too close to the digitized conductors, according to a color chart for the criticity in terms of distance.

Starting from a Lidar or a photogrammetry point cloud, with already classified data, and using a vector as reference for the analysis, the tool allows the user to identify and group all the dangerous trees that are nearby a power line.

The deliverables enable to take decisions for planning the pruning operations in terms of volume to prune and priority zones.

3.2. Required Input

  • Coordinate system for the project
  • Classified Cloud point to be analyzed ( basic classification is sufficient)
  • Vectorized Conductors to be used as reference

Additional attributes for conductors like:

  • Coordinates system for conductors
  • Temperature for conductors 
  • Reference voltage for conductors 
  • Reference distance to be used for analysis

3.3. Browsing the Vegetation Risk Map

Enable the Lidar Point Cloud, Digitized Conductors, and Vegetation Encroachment from Layers panel then zoom in point cloud for the areas where the vegetation encroaches the power lines trajectory. The distance of vegetation to lines is displayed according to a colored chart in relationship with the defined buffers. In the example below, red points highlight the closest vegetation detected, whereas blue ones are farther ones.

3.4. Deliverables

3.4.1. Text Collision Report

This report available in text format (*.txt and *.csv) shows all the portions of vegetation that are included inside the buffers used for the analysis. 

3.4.2. Map Collision Report

This report shows in a map format (*.kml and *.shp) all the portions of vegetation that are included inside the buffers used for the analysis. Each point of the LiDAR cloud included in the buffer is a point in the map. Since, according to this definition, the number of points in the map may change a lot on the basis of point cloud density, it must be considered that each point representing a collision in the map must not cover an area bigger than 20 sq.cm. 

3.4.3. Collision Detailed Report

This report is a collection of .pdf files containing a detailed view of a given collision.

3.4.4. Falling Tree Report

This report available both in text format (*.txt and *.csv) and in map format (*.kml and *.shp), shows all the portions of vegetation that can be dangerous for falling tree risk.

3.4.5. Pruning Report

This report available both in text format (*.txt and *.csv) and in map format (*.kml and *.shp), shows all the portions of vegetation that are included inside the buffers used for the analysis, grouped according the user requirements.

5. Solar Plant Thermal Inspection

5.1. Usage

This analytic serves for the two purposes of successively providing :

  1. An automated photovoltaic arrays boundaries detection (array vectorization)
  2. An automated hot spot detection within these arrays, assessing for abnormal condition (high temperature) from an automatically-computed threshold.

5.2. Required Input

A thermal map must  first have been either processed or imported in delair.ai, for the algorithms to be run onto.

5.3. Viewing Vectorized Arrays

From a 2D View, enable the Layer that contains the vectorized arrays. In the example below, the Digitized Arrays layer will display them as white rectangles overlaid to Thermal Map layer.

5.3. Viewing Detected Hot Spots

From a 2D View, enable the Layer that contains the hot spot detections :

5.4. Deliverables

A unique georeferenced layer for each plant is provided in GeoJSON format.

Hot Spots detections are provided in geoJSON format as well.

Did you find it helpful? Yes No

Send feedback
Sorry we couldn't be helpful. Help us improve this article with your feedback.