Machine Vision Readout for Faster 3D Imaging of Low-Information Scenes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional machine vision systems for forming 3D range images are inefficient due to the need to capture large 2-dimensional intensity images for each line of physical coordinates, resulting in significantly longer capture times, making laser-line based methods too slow for many industrial applications.
Innovation Solution
The system employs a method where pixel signals are accumulated and processed using uncorrelated control signals representative of random basis functions and filtering functions, allowing for the aggregation and digitization of image information in a way that reduces redundant readout and conversion, enabling faster data acquisition by compressive sensing techniques.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional laser-line based 3D image formation methods are used, then complete 2-dimensional intensity images are captured for each line of physical coordinates, but the capture time becomes excessively long (up to 100 times longer than intensity image acquisition)
Solution Approach 1:
The patent extracts only the essential information needed for 3D measurement from the full image data. Instead of capturing and processing complete 2D intensity images, the system uses a reduced set of measurements that directly encode the spatial coordinates of illuminated points, eliminating redundant data acquisition and processing time
Solution Approach 2:
The patent applies partial action by using a subset of pixel elements (e.g., every other row or column) rather than the full pixel array for 3D measurement. This partial sampling approach reduces the data volume significantly while still providing sufficient information for accurate 3D coordinate extraction through the specialized readout circuitry
2Loss of information
If full 2-dimensional intensity images are captured and processed for each illumination line, then complete image data is obtained, but the data processing load and time consumption increase significantly
Solution Approach 1:
The patent segments the image acquisition process into distinct functional components: illumination pattern projection, selective pixel sampling, analog signal aggregation, and digital coordinate extraction. Each segment is optimized for its specific function, with the readout circuitry performing real-time aggregation of pixel signals from selected rows and columns, thereby reducing the data processing burden compared to capturing complete intensity images
Solution Approach 2:
The patent transforms the problem from 2D image space to a different measurement space where coordinates are directly encoded. By using specialized readout circuitry that aggregates signals along specific geometric patterns (rows and columns), the system directly computes 3D coordinates without needing to process full 2D intensity images, effectively changing the dimensionality of the data representation
3Ease of operation
If conventional digital camera readout is used for laser-line 3D imaging, then standard image processing pipelines are available, but the readout speed is insufficient for industrial applications
Solution Approach 1:
The patent replaces the conventional mechanical/image-processing-system approach with a specialized electronic readout system. Instead of using standard digital camera readout circuits designed for capturing complete images, the invention implements custom readout circuitry that directly aggregates pixel signals along geometric patterns and outputs coordinate information, bypassing the need for full image processing pipelines and achieving much higher acquisition speeds
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 approach significantly reduces the time required to capture 3D images, improving throughput by avoiding redundant signal processing and leveraging compressive sensing principles to efficiently encode sparse image data.
Implementation Method 1
accumulating a first pixel signal based on incoming light energy for each of a plurality of pixel elements of a pixel array, the pixel elements each including a light sensor
Data Source
AI summary
A machine vision system to form a digital image that includes information about both (1) relative displacements of segments of an illumination profile within the digital image due to height discontinuities of corresponding illuminated portions of various surfaces in a scene, and (2) relative reflectivity of the illuminated portions of those surfaces.