Vehicle Object Detection Using Depth Distance Change Rate
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object detection technologies fail to adequately inhibit erroneous detection of objects, particularly reflections of the host vehicle, due to limitations in obtaining distance information.
Innovation Solution
An object detection apparatus and method that uses a camera and ECU to capture images, obtain depth distances for each pixel area, extract potential objects, calculate changes in depth distances over time, and determine if adjacent objects are the same by comparing their change rates, thereby preventing the host vehicle's reflection from being detected as a separate object.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If only a specific area is omitted as a target for which distance information is to be obtained, then the device complexity is reduced, but the reliability of object detection deteriorates because reflections in other areas cannot be managed
Solution Approach 1:
The patent changes the parameter being monitored from static depth distance to dynamic change rate of depth distance over time. By calculating how the depth distance changes between consecutive frames and comparing this change rate to expected values for real objects, the system can distinguish reflections from actual objects without adding complex hardware or excluding specific image areas.
2Measurement precision
If the amount of change over time in depth distance is used to distinguish reflections from actual objects, then the measurement precision of object detection is improved, but the loss of information increases due to requiring temporal comparison data
Solution Approach 1:
The patent utilizes continuous capture of image frames over time, maintaining a steady stream of depth distance measurements. This continuous data flow allows the system to calculate change rates between consecutive frames, providing the temporal information needed to distinguish reflections from real objects while using the naturally occurring continuous imaging process.
Data Source
AI summary
An object detection apparatus includes a camera and an ECU. The camera is mounted to a vehicle and captures an image. The ECU is configured to obtain a depth distance for each pixel area of the image based on the image; extract at least one potential object based on the depth distance; calculate an amount of change over time in the depth distance of the at least one potential object; and detect an object around the vehicle from among the at least one potential object using the amount of change over time. When a first potential object and a second potential object that are adjacent in the image are extracted, and a second amount of change over time is twice a first amount of change over time, the second potential object is determined to be the same object as the first potential object.