Motion Detection in 3D Data Reconstruction via Temporal Pixel Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional stereo vision systems are limited in their ability to detect motion during three-dimensional data reconstruction of objects, leading to invalid pick points and reduced throughput in applications like item picking, where object movement can invalidate pick points.
Innovation Solution
A system that generates temporal pixel images from stereo image sequences, calculates derived values such as temporal averages and deviations, and determines correspondence data to indicate the likelihood of motion in a scene, leveraging existing 3D reconstruction data to quickly detect motion without the need for computationally expensive optical-flow approaches.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional stereo vision systems are used for 3D reconstruction, then the system structure is simple, but the ability to detect motion is limited
Solution Approach 1:
The patent segments the image data by creating temporal pixel images that separate motion information from static scene information. Each temporal pixel contains intensity values across multiple time points, allowing the system to analyze motion independently while maintaining the overall stereo vision structure.
Solution Approach 2:
The patent adds a temporal dimension to the traditional spatial image data by creating temporal pixel images with multiple intensity values over time. This transforms 2D spatial information into 4D spatio-temporal information, enabling motion detection without adding physical sensors.
2Measurement precision
If optical-flow approaches are used to detect motion, then motion detection accuracy is improved, but computational cost increases
Solution Approach 1:
The patent extracts only the essential motion information needed for pick point validation by analyzing temporal pixel variations. Instead of performing full optical flow computation, it extracts motion indicators directly from temporal pixel intensity changes, significantly reducing computational requirements while maintaining sufficient accuracy for the application.
Solution Approach 2:
The patent performs partial motion analysis by focusing only on regions and temporal variations relevant to pick point validation. It does not compute complete motion fields but rather sufficient motion indicators to determine whether objects have moved since 3D data capture, reducing computational effort to the minimum necessary.
3Speed
If motion detection is not performed, then processing speed is fast, but pick points become invalid due to object movement
Solution Approach 1:
The patent performs preliminary motion detection by analyzing temporal pixel images before finalizing pick points. This early motion check allows the system to identify and exclude moved objects from pick point generation, ensuring that only valid pick points for stationary objects are provided to the picker robot.
Solution Approach 2:
The patent introduces temporal pixel images as an intermediary data structure between the stereo image capture and pick point generation. These temporal pixels serve as a mediator that efficiently conveys motion information without requiring complex processing, enabling fast validation of pick point validity.
Data Source
AI summary
In some aspects, the techniques described herein relate to systems, methods, and computer readable media for detecting movement in a scene. A first temporal pixel image is generated based on a first set of images of a scene over time, and a second temporal pixel image is generated based on a second set of images. One or more derived values are determined based on values of the temporal pixels in the first temporal pixel image, the second temporal pixel image, or both. Correspondence data is determined based on the first temporal pixel image and the second temporal pixel image indicative of a set of correspondences between image points of the first set of images and image points of the second set of images. An indication of whether there is a likelihood of motion in the scene is determined based on the one or more derived values and the correspondence data.