2D Camera Fill Volume Detection for Earthmoving Loaders
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for optimizing the loading phase of earthmoving machines like wheel tractor scrapers are inefficient due to reliance on expensive three-dimensional perception sensors, which are cost-prohibitive for fleets and computationally expensive, leading to suboptimal loading and increased operational costs.
Innovation Solution
A method using a two-dimensional camera system with a neural network to determine the fill volume of a payload carrier by segmenting images and identifying regions with fill material, providing real-time feedback to operators or autonomous systems to optimize loading cycles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If three-dimensional perception sensors (LiDAR or stereo camera) are used to monitor the bowl, then loading efficiency is improved, but system cost increases significantly
Solution Approach 1:
The patent replaces expensive three-dimensional perception sensors with inexpensive two-dimensional cameras that can be deployed across entire fleets. The 2D camera system achieves sufficient loading monitoring capability at a fraction of the cost of LiDAR or stereo camera systems, making fleet-wide deployment economically viable.
Solution Approach 2:
The patent uses two-dimensional image copies from standard cameras to represent the three-dimensional bowl contents. By processing 2D images through neural networks and image analysis algorithms, the system extracts fill level information without requiring direct 3D sensing, effectively creating a computational copy of the 3D state from 2D data.
2Measurement precision
If three-dimensional image processing is used to determine fill volume, then measurement accuracy is improved, but computing power requirements increase
Solution Approach 1:
The patent extracts only the essential information needed for fill volume measurement from the image data, rather than performing complete three-dimensional reconstruction. By focusing on identifying fill material presence in segmented regions and calculating volume from 2D projections, the system achieves sufficient accuracy with reduced computational processing.
Solution Approach 2:
The patent replaces computationally intensive three-dimensional image processing algorithms with a simplified approach using two-dimensional image analysis and neural networks. The system substitutes complex 3D reconstruction mechanics with 2D pattern recognition and volumetric estimation based on image segmentation, significantly reducing computing power requirements.
3Productivity
If the payload carrier is fully loaded to maximize productivity, then output is improved, but machine wear and fuel consumption increase
Solution Approach 1:
The patent implements real-time feedback on payload carrier fill levels to operators or autonomous control systems. By continuously monitoring fill status through the 2D camera system and providing immediate information, the system enables optimization of loading cycles to achieve sufficient load levels without excessive overloading, balancing productivity with fuel efficiency and machine wear reduction.
Data Source
AI summary
A method for loading a payload carrier of a machine includes receiving, from a camera on the machine, a two-dimensional image of an interior of the payload carrier as material is loaded into the payload carrier. The method further includes sectioning the two-dimensional image into a plurality of regions, and determining, for individual of the regions, whether the region includes a representative of fill material. The method may also include determining a fill volume based on the regions having fill and those not having fill.