Vehicle Camera Blockage Detection for Selective Sensor Cleaning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting blockages in camera sensors of vehicles are not precise enough, leading to unnecessary cleaning and inefficiencies in advanced driver assistance systems.
Innovation Solution
A method using a neural network algorithm to segment and classify camera images, determining the degree and type of blockage, and transmitting cleaning information to a dedicated cleaning device based on these results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a cleaning device is activated whenever a blockage is detected, then the camera sensor is cleaned more frequently, but unnecessary cleaning occurs reducing system efficiency
Solution Approach 1:
The patent segments the blockage detection task into two independent neural network processes: segmentation to identify blocked regions and classification to determine blockage type. This segmentation allows each process to specialize, improving detection accuracy while enabling selective cleaning activation based on blockage type, thus avoiding unnecessary cleaning operations.
Solution Approach 2:
The patent applies local quality by treating different blockage types (rain, snow, dirt, fog) with different cleaning strategies. The classification module identifies the specific type of blockage, and the control unit activates cleaning only when appropriate (e.g., not activating for fog when a wiper would be ineffective), optimizing cleaning efficiency based on local conditions.
2Measurement precision
If a neural network algorithm is used to segment and classify camera images, then blockage detection precision is improved, but computational complexity increases
Solution Approach 1:
The patent divides the complex blockage detection task into two separate neural network modules: a segmentation network for identifying blocked regions and a classification network for determining blockage type. This segmentation of the algorithm reduces the complexity of each individual network while maintaining or improving overall detection precision through specialized processing.
Solution Approach 2:
The patent introduces an intermediary control unit that receives outputs from both the segmentation and classification neural networks. This control unit integrates the information from both modules and makes the final decision on cleaning activation, acting as a mediator that simplifies the overall system decision-making process while leveraging the precision of both neural network outputs.
3Area of stationary object
If the camera sensor is placed at the outer surroundings of the vehicle, then the field of view is improved, but the transparent component is more exposed to blockage
Solution Approach 1:
The patent implements a feedback mechanism where the neural network continuously analyzes camera images to detect blockages. When a blockage is detected, the system activates the cleaning device to remove it, and then continues monitoring to confirm the blockage is cleared. This closed-loop feedback ensures the camera maintains its optimal field of view by actively responding to blockage conditions.
Solution Approach 2:
The system employs self-service by automatically detecting blockages and activating cleaning operations without requiring manual intervention. The neural network continuously monitors the camera sensor condition and triggers cleaning when necessary, allowing the system to maintain its own operational effectiveness autonomously.
Data Source
AI summary
Method for determining cleaning information for a camera sensor which includes a blockage on a transparent camera sensor component in an optical camera sensor path, the method including:processing a captured camera image with a neural network algorithm which determines as an output a degree of camera sensor blockage by segmenting a part of the camera image and a camera sensor blockage class by classification of a part of the camera image,determining cleaning information dependent on the camera sensor blockage degree and the camera sensor blockage class, wherein the cleaning information describes that a cleaning of the camera sensor is required if a cleaning criterion is assigned to the determined camera sensor blockage class and if a degree threshold is exceeded by the determined camera sensor blockage degree, andtransmitting the cleaning information to a camera sensor cleaning device to clean the camera sensor.