Safety Belt Detection Using Grid Classification Network
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional safety belt detection methods in Driver Monitoring Systems (DMS) face challenges with low accuracy and high computational complexity, making it difficult to determine if a driver is wearing a safety belt correctly, especially due to complex backgrounds, lighting conditions, and the slender shape of safety belts.
Innovation Solution
A safety belt detection method utilizing a deep learning-based neural network with a global dichotomous branch network and a grid classification branch network, which quickly determines if a driver is wearing a safety belt and identifies its position, reducing computational complexity by dividing the image into grids for classification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional deep learning methods are used for safety belt detection, then detection capability is provided, but detection accuracy is low and computational complexity is high
Solution Approach 1:
The patent divides the image into multiple grid cells and processes each grid independently through classification. This segmentation approach reduces the computational complexity by breaking down the large image processing task into smaller, manageable units, while maintaining detection accuracy through systematic evaluation of each grid region for safety belt presence.
Solution Approach 2:
The patent transforms the image processing problem from a single holistic classification to a multi-dimensional grid-based classification system. By organizing the image into a grid structure and evaluating each grid separately, the system achieves both reduced computational complexity and maintained accuracy through distributed processing across multiple dimensions.
2Loss of information
If traditional detection methods are used, then safety belt detection is performed, but detailed information of the safety belt cannot be obtained
Solution Approach 1:
The patent segments the image into multiple grid cells, allowing detailed examination of specific regions where safety belts may be located. This segmentation enables the system to obtain detailed information about safety belt position, orientation, and characteristics by analyzing individual grids, while managing system complexity through structured processing.
Solution Approach 2:
The patent applies local quality analysis by evaluating each grid cell independently for safety belt characteristics. This allows the system to obtain detailed local information about safety belt presence, position, and appearance in specific regions, while maintaining overall system manageability through standardized local evaluation procedures.
3Measurement precision
If traditional methods are used for safety belt detection, then detection is performed, but it is difficult to judge correct wearing
Solution Approach 1:
The patent divides the image into grids to systematically evaluate the safety belt's position and orientation relative to the driver's body. This segmentation enables precise judgment of correct wearing by checking specific grid regions for proper safety belt placement, while managing detection difficulty through structured evaluation criteria.
Solution Approach 2:
The patent applies local quality assessment by examining each grid cell for specific safety belt characteristics and positioning. This allows accurate judgment of correct wearing by evaluating local grid regions for proper safety belt orientation and location, while reducing overall detection difficulty through standardized local evaluation protocols.
Data Source
AI summary
A safety belt detection method, apparatus, computer device and computer readable storage medium are disclosed. The safety belt detection method includes the steps as follows. An image to be detected is obtained. The image to be detected is inputted into a detection network which includes a global dichotomous branch network and a grid classification branch network. A dichotomous result, which indicates whether a driver is wearing a safety belt and is output from the global dichotomous branch network, is obtained. A grid classification diagram, which indicates a position information of the safety belt and is output from the grid classification branch network, is obtained based on image classification. A detection result of the safety belt, indicating whether the driver is wearing the safety belt normatively, is obtained based on the dichotomous result and the grid classification diagram.