Road Image State Determination Using Multi-Model Output Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing road state determination systems face challenges in ensuring accuracy when determining road conditions from images captured under various weather and time conditions using a single determination model.
Innovation Solution
A state determination device that combines outputs from multiple determination models, each learned using training data with varying conditions such as different states of moving objects, road conditions, and external environments, to improve determination accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single determination model is used to determine road state from images captured under various weather and time conditions, then the device complexity is reduced, but the measurement precision of road state determination deteriorates
Solution Approach 1:
The patent divides the single determination model into multiple specialized determination models, where each model is trained on images captured under specific weather conditions (e.g., sunny, cloudy, rainy, snowy). This segmentation allows each model to specialize in particular imaging conditions, thereby improving determination accuracy for each condition while maintaining manageable system complexity through modular architecture.
Solution Approach 2:
The patent creates a multi-functional determination system where multiple determination models work together to handle various weather conditions. Each model serves a specific function for particular conditions, but collectively they provide universal coverage for all possible imaging environments, enabling the system to maintain high accuracy across diverse conditions.
2Measurement precision
If multiple determination models trained on different training data are used to improve determination accuracy, then the measurement precision improves, but the device complexity increases
Solution Approach 1:
The patent implements a dynamic model selection mechanism that automatically selects or weights determination models based on the current weather conditions detected in the input image. This dynamic approach allows the system to adaptively use the most appropriate model for each condition, improving accuracy while managing complexity through intelligent resource allocation rather than static multi-model architecture.
Solution Approach 2:
The patent introduces a condition detection module or intermediary layer that analyzes the input image to determine current weather conditions and selects or weights the appropriate determination model accordingly. This intermediary component manages the complexity of multiple models by providing a systematic method for model selection and integration, thereby maintaining system manageability while leveraging multiple specialized models.
Data Source
AI summary
A state determination device according to the present invention includes: a memory; and at least one processor coupled to the memory. The processor performs operations. The operations including: combining output from a plurality of determination models for an input image of a road, the plurality of determination models each learned using training data in which at least one of a state of a moving object mounting an imaging device that acquires an image of the road, a state of the road, and an external environment is different.