Camera Shadow Subtraction for Vehicle Object Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting objects in a motor vehicle's environmental region, such as ultrasonic sensors and rear view cameras, face challenges in accurately determining object distance and differentiating between obstacles and non-obstacles, leading to false alarms and incomplete information display, especially with non-flat or large objects and objects above the horizon.
Innovation Solution
A camera system that uses illumination to generate shadow images, subtracts these from non-illuminated images to create ratio images, and identifies step edges along epipolar lines to generate an edge image, allowing for precise detection of objects with height above the ground, thereby distinguishing obstacles from ground markings and providing accurate distance calculations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If a rear view camera displays images on a display, then the driver can see the environmental region, but the driver cannot accurately estimate object distance due to non-linear relative distances
Solution Approach 1:
The patent applies Inverse Perspective Mapping (IPM) to transform the 2D camera image into a 3D-like bird's eye view representation. This dimensional transformation allows distances to be displayed linearly and accurately, enabling the driver to correctly estimate object distances while maintaining the visual information from the camera.
2Reliability
If ultrasonic sensors are used to detect objects, then object detection is achieved, but multiple sensors are required to monitor the entire environmental region and flat objects are missed
Solution Approach 1:
The patent makes the camera system multi-functional by combining traditional imaging with shadow-based detection capabilities. The same camera that captures visual information also detects object height through shadow analysis, eliminating the need for separate ultrasonic sensors and providing both presence and height information with a single device.
Solution Approach 2:
The patent utilizes optical properties (shadow formation and light absorption) to detect object characteristics. By analyzing the shadow patterns and intensity variations in the captured images, the system determines object height and presence, converting optical information into detection data without requiring additional sensors.
3Measurement precision
If automatic object detection is applied, then objects are detected and displayed, but height values cannot be assigned and ground markings are falsely identified as obstacles
Solution Approach 1:
The patent introduces shadow analysis as an intermediary mechanism between the camera image and object detection. The shadow pattern serves as a mediator that provides height information about detected objects, allowing the system to distinguish between actual obstacles with height and flat ground markings, thereby improving classification accuracy.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This method reduces false alarms by accurately identifying obstacles with height, enhancing driver safety by providing precise and reliable object detection and distance information, reducing the need for multiple sensors and improving reaction time.
Implementation Method 1
A first image of the environmental region is provided by a camera of the camera system with illumination of the environmental region by means of at least one light source of the motor vehicle such that a shadow of the object caused by the illumination is depicted in the first image
Data Source
AI summary
A method for detecting an object in an environmental region of a motor vehicle by a camera system of the motor vehicle is disclosed. A first image of the environmental region is provided by a camera of the camera system with illumination of the environmental region using a light source of the motor vehicle such that a shadow of the object caused by the illumination is depicted in the first image. A second image of the environmental region is also provided by the camera in an inactive state of the light source, in which the illumination is omitted. Then, a subtraction image is generated by subtracting one of the images from the other image by an image processing device of the camera system. Then, an edge image is generated from the subtraction image, and the object is detected based on edges of the edge image by the image processing device.