Camera-Based Object Detection Using Segmented Regions of Interest
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional distance detection apparatuses in vehicles are limited in providing richer information such as object type and motion state, and are prone to misjudgment and high setup costs, particularly in addressing the increasing number of rear-end collision accidents.
Innovation Solution
An object detection device and method utilizing a camera, processor, and deep neural network learning model to analyze images, define regions of interest, and output position and size information of target objects, enabling reliable detection of objects in front, including their type and distance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If traditional distance detection apparatus (radar) is used, then distance sensing function is provided, but the apparatus cannot provide richer information such as object type and motion state
Solution Approach 1:
The patent segments the detection task into multiple specialized modules: a first detection module for detecting objects in a first region, a second detection module for detecting objects in a second region, and a third detection module for detecting objects in a third region. Each module is optimized for specific detection requirements, enabling comprehensive information extraction including object type, motion state, and distance while maintaining high detection accuracy across different scenarios
Solution Approach 2:
The patent implements a multi-functional detection system that can simultaneously perform distance sensing, object type recognition, and motion state detection. The detection apparatus integrates multiple detection modules that work together to provide comprehensive information about target objects, making it versatile for various detection needs beyond simple distance measurement
2Reliability
If traditional distance detection apparatus is used, then distance sensing is provided, but the apparatus is susceptible to misjudgment
Solution Approach 1:
The patent employs feedback mechanisms where detection results from multiple modules are integrated and cross-validated. The system uses detection data from different regions and modules to相互 verify and supplement each other, reducing misjudgment risks and improving both reliability and measurement precision through collaborative decision-making
Solution Approach 2:
By dividing the detection field into multiple regions (first region, second region, third region) and using specialized detection modules for each, the system achieves higher precision in each specific region while maintaining overall reliability through the coordinated work of all modules
3Ease of manufacture
If traditional distance detection apparatus is used, then distance sensing function is provided, but the setup cost is high
Solution Approach 1:
The patent adopts cameras as the primary detection device, which are significantly cheaper than traditional radar systems. The camera-based approach uses image processing and deep learning algorithms to achieve reliable detection of object type, motion state, and distance, providing a cost-effective solution that maintains high detection reliability through intelligent image analysis rather than expensive hardware
Data Source
AI summary
There is provided an object detection device and an object detection method. The processor of the object detection device defines respective overall image areas of a plurality of first sensed images from a plurality of original sensed images as first regions of interest; the processor defines respective partial image areas of a plurality of second sensed images from the plurality of original sensed images as second regions of interest, and crops out a plurality of third sensed images; the processor inputs the plurality of first sensed images and the plurality of third sensed images to a deep neural network learning model, so that the deep neural network learning model outputs image information of a target object image in the plurality of first sensed images and the plurality of third sensed images, respectively. By this, a function of detecting an object in front with high reliability is provided by means of image detection.