Blind Spot Detection Using CNN Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current blind spot monitoring systems face challenges in accurately detecting vehicles in blind spots due to high computational loads, limited detection rates, and dependence on external environments, making real-time detection difficult in embedded systems.
Innovation Solution
A method using a learned convolutional neural network (CNN) to segment rear images, calculate free space ratios in blind spots, and warn drivers of lane change risks by analyzing movement information and threshold values, allowing for efficient detection with reduced computational requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional algorithms are used for blind spot detection, then detection accuracy can be maintained, but computational load becomes too heavy for real-time detection in embedded systems
Solution Approach 1:
The patent segments the blind spot area into multiple sub-blind spots and calculates free space ratios for each segment independently. This segmentation approach reduces the overall computational load by dividing the complex detection task into smaller, more manageable calculations that can be performed in real-time on embedded systems while maintaining detection accuracy through aggregated segment analysis.
2Reliability
If more visual processing is performed to achieve accurate detection, then detection reliability improves, but computational requirements increase making real-time detection difficult
Solution Approach 1:
The patent extracts only the essential visual information needed for detection by calculating free space ratios in segmented blind spot areas. Instead of performing comprehensive visual processing on entire images, the system extracts and processes only the relevant blind spot regions, reducing processing complexity while maintaining detection reliability through focused analysis of critical areas.
3Adaptability or versatility
If convex mirrors are added to side view mirrors to eliminate blind spots, then detection coverage improves, but driver strain increases due to the need to constantly monitor multiple mirrors
Solution Approach 1:
The patent implements a system that automatically detects and monitors blind spot areas without requiring driver intervention. The alarm device autonomously processes visual information, calculates free space ratios, and provides warnings when vehicles are detected in blind spots, eliminating the need for drivers to manually check multiple mirrors and reducing driver strain while maintaining comprehensive detection coverage.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method for warning a vehicle of a risk of lane change is provided. The method includes steps of: (a) an alarm device, if at least one rear image captured by a running vehicle is acquired, segmenting the rear image by using a learned convolutional neural network (CNN) to thereby obtain a segmentation image corresponding to the rear image; (b) the alarm device checking at least one free space ratio in at least one blind spot by referring to the segmentation image, wherein the free space ratio is determined as a ratio of a road area without an object in the blind spot to a whole area of the blind spot; and (c) the alarm device, if the free space ratio is less than or equal to at least one predetermined threshold value, warning a driver of the vehicle of the risk of lane change.