Overhead Line Object Detection Using 3D Parallax Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting objects on installations, such as overhead lines, using aerial photographs are prone to errors due to the lack of depth information in 2D images, leading to false positives and increased manual processing effort, as they cannot reliably distinguish objects on the line from those on the ground.
Innovation Solution
The method incorporates 3D information using the parallax effect to differentiate between objects on and under installations by combining 2D and 3D detection, reducing false alarms and computational complexity through semantic segmentation and LIDAR data, allowing for automatic and reliable object detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 2D aerial photographs are used for object detection, then the detection process is simple and quick, but depth information is lost leading to false positives and inability to distinguish objects on line from ground objects
Solution Approach 1:
The patent transitions from 2D image analysis to 3D spatial analysis by integrating LIDAR point cloud data with aerial photographs. The LIDAR data provides depth information and three-dimensional coordinates, enabling the system to distinguish whether detected objects are on the overhead line or on the ground below, thereby resolving the false positive problem while maintaining manageable system complexity through efficient data fusion algorithms
2Reliability
If manual evaluation of recorded images is performed, then object detection accuracy can be improved, but processing time and labor costs increase significantly
Solution Approach 1:
The patent replaces manual visual evaluation with an automated computer-based detection system that processes LIDAR point cloud data and aerial photographs. The system uses algorithms to automatically identify objects, determine their spatial locations, and classify them as being on the line or on the ground, achieving both high reliability through 3D analysis and high productivity through automated processing without manual intervention
3Measurement precision
If 3D LIDAR data is integrated with 2D images for object detection, then false positives are reduced and object localization is improved, but computational complexity increases
Solution Approach 1:
The patent performs preliminary processing of LIDAR point cloud data by generating a three-dimensional representation of the overhead line structure before object detection. This pre-established spatial model allows the system to efficiently compare detected objects against known line positions, reducing the computational burden during actual detection while maintaining high position accuracy through the pre-computed 3D spatial relationships
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach significantly reduces false positive detections and manual processing effort, enabling immediate identification of hazards and damage on installations, with lower computational requirements and more accurate object localization.
Implementation Method 1
a three-dimensional representation of the installation is obtained by means of a light detection and ranging (LIDAR) sensor
Implementation Method 2
Due to the parallax effect, objects under the installation, e.g. an overhead line, are represented in the images at different positions in relation to the line
Data Source
AI summary
A method for detecting objects on systems, includes providing a three-dimensional representation of the system, wherein the position and orientation of the representation and the system are known, and capturing a first image and a second image of the system, the two images being captured from different positions above the system. For a plurality of sections of the system, a respective comparison of the first and the second image is carried out using a parallax effect. If the images in a region surrounding the system match, an object is detected on the system.