Vehicle Camera Signal Processing With Local-Cloud AI Split
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle signal processing devices face limitations in expanding functions or services due to the need for downloading AI models from the cloud using limited computing resources, and there is a challenge in efficiently processing camera data for various services while ensuring privacy and stability.
Innovation Solution
A vehicle signal processing device that includes a processor capable of performing partial image processing locally and controlling cloud-based processing, with features like data splitting, calibration, and privacy management, to efficiently handle camera data and execute services.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If AI models are downloaded from the cloud to the vehicle, then image processing capability is improved, but the limited computing resources of the vehicle are overwhelmed and function expansion is limited
Solution Approach 1:
The patent divides image processing into two segments: complex AI-based processing is performed on the cloud server, while basic processing and result integration are performed on the vehicle's processor. This segmentation allows the vehicle to benefit from advanced AI capabilities without needing to download entire AI models, thus avoiding overwhelming the vehicle's limited computing resources while still achieving high-level image analysis functionality.
2Power
If all camera data is processed by the cloud, then processing power is sufficient, but network dependency increases and processing latency occurs
Solution Approach 1:
The vehicle's processor performs preliminary processing on camera data before transmitting to the cloud. This preliminary action includes basic image processing and preparation steps that can be completed locally without network dependency, reducing the amount of data that needs to be transmitted and processed remotely, thereby reducing overall processing latency while maintaining sufficient processing power through cloud collaboration.
3Adaptability or versatility
If camera data is transmitted to the cloud for processing, then advanced services are enabled, but privacy data exposure risk increases
Solution Approach 1:
The patent extracts and processes privacy-sensitive information locally on the vehicle's processor before transmitting only the necessary processed data to the cloud. This extraction approach enables advanced services by utilizing cloud computing power for non-sensitive analysis while removing privacy risks by keeping sensitive data processing within the vehicle's secure environment.
4Speed
If image processing is performed entirely on the vehicle, then response time is fast, but the complexity of the vehicle's processing system increases
Solution Approach 1:
The processing system is segmented into local and cloud components, where the vehicle's processor handles time-critical basic processing and result integration to maintain fast response times, while complex AI processing is segmented to the cloud. This segmentation avoids the need for the vehicle to possess complete AI processing capabilities locally, thus preventing excessive system complexity while preserving essential response speed.
Data Source
Figure 1
Figure 2A~2B
Figure 2C~3A
AI summary
A signal processing device of a vehicle and a vehicle display apparatus including the same according to an embodiment of the present disclosure include a processor configured to receive camera data from a camera in a vehicle and to transmit at least some of the camera data or information related to the camera data to a cloud, wherein the processor is configured to perform a portion of image processing of the camera data and to control the cloud to perform another portion of the image processing of the camera data, and to execute at least one service based on result data of the image processing performed by the processor in the vehicle, and result data of the image processing performed by the cloud. Accordingly, it is possible to provide service by efficiently processing the camera data.