EV Charging Station Image Tagging for Vehicle Recognition Training
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing methods for training machine learning models to recognize electric vehicles are inefficient due to the need for extensive human tagging of images from continuous video streams, which is costly and time-consuming, and require sorting through numerous irrelevant frames.
Innovation Solution
Electric vehicle charging stations equipped with cameras and sensors identify and select only relevant frames containing electric vehicles, using changes in station status to automate the tagging process, reducing the burden on human taggers and enabling efficient data collection for training models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If all frames from continuous video stream are used for training, then training data quantity increases, but human tagging burden and time consumption increase significantly
Solution Approach 1:
The patent extracts only the relevant frames containing electric vehicles from the continuous video stream using change detection algorithms. By monitoring status changes at the charging station (such as vehicle arrival, departure, or charging state changes), the system identifies and extracts only those frames that depict electric vehicles, eliminating the need to manually tag all frames in the video stream.
Solution Approach 2:
The system performs automatic frame selection and tagging without human intervention. The change detection mechanism automatically identifies frames with electric vehicles by comparing consecutive frames or monitoring charging station status, and the system automatically tags these frames using the detected status information, making the entire process self-serving and eliminating manual human tagging.
2Measurement precision
If manual tagging of all video frames is performed, then labeling accuracy improves, but cost and time consumption increase
Solution Approach 1:
The patent replaces the mechanical process of manual human tagging with an automated computational system. Change detection algorithms and status monitoring mechanisms automatically identify frames containing electric vehicles and generate labels based on charging station status changes, substituting human labor with automated image processing and status detection systems.
3Reliability
If continuous video streaming is monitored for all frames, then no relevant images are missed, but data processing burden increases
Solution Approach 1:
The system extracts only the essential information needed for training by monitoring status changes at the charging station. Instead of processing all video frames, the system identifies status change events (such as vehicle connection, disconnection, or charging state changes) and extracts only the frames corresponding to these events, significantly reducing data processing requirements while maintaining reliability.
Data Source
AI summary
The disclosed embodiments provide a method performed at a computer system that is in communication with an electric vehicle charging station (EVCS). The EVCS includes a camera for obtaining images in a region proximal to the EVCS. The method includes capturing, using the camera, a plurality of images of electric vehicles, each image in the plurality of images being an image of a respective electric vehicle. The method further includes, for each respective image of the plurality of images of electric vehicles: determining, without user intervention, a characteristic of the respective electric vehicle; and tagging, without user intervention, the respective image with the characteristic of the respective electric vehicle. The method further includes training a first machine learning algorithm to identify the characteristic of other electric vehicles using the tagged plurality of images.