3D Machine Vision Part Inspection with Continuous Sampling

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing machine vision detection solutions suffer from inefficiencies and inaccuracies due to structural design limitations and cost constraints, leading to missed detections and suboptimal performance in applications like surface flatness and step height detection of workpieces.

Innovation Solution

A machine vision detection method utilizing three-dimensional image analysis with uniformly distributed sampling positions and height difference calculations to determine the quality of parts, avoiding missed detections by using three height differences and enabling continuous sampling without stopping at each position.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If traditional machine vision detection methods are used with manual measurement and judgment, then the system structure is simple and cost is low, but detection efficiency is low and automation degree is low

Engineering Contradiction:

Improvedetection efficiencyVSAvoidsystem structure complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent replaces manual measurement and judgment with an automated machine vision system that uses image acquisition devices and image processing algorithms to automatically detect workpiece features, eliminating the need for manual mechanical measurement while significantly improving detection efficiency and automation

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent transitions from traditional two-dimensional image detection to three-dimensional depth information detection by introducing depth maps and height difference calculations, enabling more comprehensive workpiece inspection including surface flatness and step height detection

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

2Measurement precision

If existing machine vision detection solutions are used, then cost is controlled, but detection accuracy is insufficient and structural design is limited

Engineering Contradiction:

Improvedetection accuracyVSAvoidstructural design complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the detection process into multiple independent modules including image acquisition, depth map generation, sampling position determination, height difference calculation, and quality judgment, allowing each module to be optimized independently while improving overall detection accuracy

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary actions by pre-determining optimal sampling positions based on workpiece geometry and pre-calculating height difference thresholds before actual detection, which improves detection accuracy without requiring complex real-time processing

Inventive Principle:

Principle #10Preliminary action

3Productivity

If sampling is performed at each position separately, then detection accuracy is maintained, but detection speed is low and continuous sampling cannot be achieved

Engineering Contradiction:

Improvedetection speedVSAvoidsampling accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent enables continuous sampling by processing multiple sampling positions simultaneously through parallel computation of height differences at all predetermined positions, eliminating the need to stop and restart at each position while maintaining comprehensive coverage and detection accuracy

Inventive Principle:

Principle #20Continuity of useful action

4Reliability

If only single height difference detection is used, then detection process is simple, but missed detections occur due to uneven or tilted placement

Engineering Contradiction:

Improvedetection reliabilityVSAvoiddetection criterion complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent applies partial or excessive action by calculating three different height differences (first height difference for one workpiece, second height difference for another workpiece, and third height difference between corresponding positions) instead of a single height difference, providing redundant detection perspectives that eliminate missed detections caused by uneven or tilted placement

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS12437461B2Machine vision detection method, detection device thereof, and detection system thereof

Publication Date: 2025.10.07 CONTEMPORARY AMPEREX TECHNOLOGY (HONG KONG) LIMITED

AI summary

A machine vision detection method includes receiving a three-dimensional image of a part to be detected including a first part and a second part, determining a plurality of sampling positions that satisfy sampling conditions, acquiring first sample data of the three-dimensional image of the first part at sampling positions and second sample data of the three-dimensional image of the second part at the sampling positions, calculating a first height difference between a plurality of first sample data and a second height difference between a plurality of second sample data, calculating a third height difference between the first sample data and the second sample data, and determining that the part to be detected is unqualified when any one of the first height difference, the second height difference, and the third height difference fails to meet a preset detection criterion.