Bounding Box Correction for Accurate Object Position Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object tracking technologies, such as those using the Kalman filter, face inaccuracies due to objects being obscured or extending beyond the frame of acquired images, leading to errors in position estimation.
Innovation Solution
A method and device for correcting or updating bounding boxes based on their location within or outside the image frame boundaries, using image processing to adjust the size and position of bounding boxes to improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If object detection is performed using standard image processing, then object position can be detected, but position estimation accuracy deteriorates when objects are obscured or extend beyond frame boundaries
Solution Approach 1:
The patent applies preliminary action by predicting the object's position before it is fully detected. When an object is partially obscured or extends beyond the frame, the system uses motion models and historical data to predict where the object should be, then adjusts the bounding box accordingly before final position estimation occurs.
Solution Approach 2:
The patent implements feedback mechanisms where the system continuously monitors detection confidence and position estimates. When objects are partially detected or extend beyond boundaries, the system feeds back to adjust the bounding box using prediction models and historical tracking data, iteratively improving position accuracy.
2Measurement precision
If bounding box is updated based on current image detection, then response time is reduced, but position estimation accuracy deteriorates when objects are partially obscured
Solution Approach 1:
The system performs preliminary position prediction based on motion models and historical data before updating the bounding box with current image detection results. This allows the system to maintain accurate position estimates even when current detection is incomplete, reducing the effective time delay while preserving accuracy.
Solution Approach 2:
The patent implements dynamic adjustment of the bounding box by blending current detection data with predicted positions based on object motion characteristics. The system dynamically weights between immediate detection and predicted position based on detection confidence, allowing timely updates while maintaining accuracy when objects are partially obscured.
Data Source
AI summary
Disclosed is a device configured to estimate a position of an object. The device includes: an image sensor configured to acquire an image around a road; and a processor configured to detect the object from the acquired image. The processor may be configured to: determine whether a portion of a first bounding box of the detected object is located within a predetermined region of the acquired image or whether a portion of the first bounding box is outside a frame boundary of the acquired image; and update the first bounding box to a second bounding box based on that the portion of the first bounding box is located within the predetermined region of the acquired image or that the portion of the first bounding box is outside the frame boundary of the acquired image.