Industrial Robot Training Data Generation Using Synthetic Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating training data sets for industrial robots are lengthy and complex, making it difficult for users to easily and efficiently train robots for specific tasks.
Innovation Solution
A method involving the processing of initial imaging information to generate further imaging information, which is then used to create a training data set through automated image processing, utilizing cloud computing and deep learning to extend the information content and facilitate efficient training of industrial robots.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional methods are used to generate training data sets for industrial robots, then the training data sets can be generated with sufficient quality, but the process becomes lengthy and complex
Solution Approach 1:
The patent uses image synthesis technology to create artificial images that copy and extend real-world visual data. Instead of requiring actual physical objects and complex capture processes, the system generates synthetic images that replicate real object appearances, allowing parallel processing of multiple training samples simultaneously and dramatically reducing generation time while maintaining data quality
Solution Approach 2:
The patent replaces mechanical data collection processes (physical object handling, camera capture, manual annotation) with computational methods. Image synthesis algorithms generate training data through software processing rather than physical operations, enabling automated, rapid creation of training sets without human intervention or complex mechanical setups
2Manufacturing precision
If traditional methods are used to generate training data sets for industrial robots, then the training data sets can be generated with sufficient quality, but the process becomes complex
Solution Approach 1:
The system creates synthetic training data by copying and transforming existing image data through computational algorithms. This approach simplifies the generation process by eliminating the need for complex physical setups, multiple capture devices, and manual processing steps, while still producing high-quality training data through sophisticated image synthesis
Solution Approach 2:
The patent divides the training data generation process into independent computational modules: image synthesis, data augmentation, and annotation generation. Each module processes data independently and can be executed in parallel, reducing overall process complexity while maintaining data quality through systematic processing of individual components
3Reliability
If more training data is generated to improve robot performance, then the robot's task execution improves, but the generation process becomes more time-consuming
Solution Approach 1:
The patent enables continuous generation of training data through automated image synthesis that operates without interruption. The system can generate multiple training samples simultaneously and continuously, providing an unlimited supply of high-quality training data without the time penalties associated with traditional sequential data collection methods
Solution Approach 2:
By using image synthesis to copy and replicate real-world visual data, the system can rapidly generate large volumes of training data without requiring proportional increases in physical capture time. The computational processes create multiple training samples from single real-world observations, dramatically increasing data quantity while maintaining generation speed
Data Source
AI summary
A method for generating a training data set for training an industrial robot which can be trained based on a corresponding training data set, comprising: providing a first imaging information, which describes a first one- or multi-dimensional image of an object which is to be relocated by means of an industrial robot which is to be trained on the basis of the training data set to be generated; processing the first imaging information to generate further imaging information, which describes at least one artificially generated further one- or multi-dimensional image of the object which is to be moved by means of an industrial robot which is to be trained on the basis of the training data set to be generated; and processing the further imaging information to generate a training data set for training an industrial robot which can be trained on the basis of the training data set.