Adaptive Gradation Mapping for One-Shot Workpiece Inspection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image inspection methods face challenges in handling varying workpiece materials and postures, leading to insufficient dynamic range and robustness issues, particularly in one-shot HDR, which affects tact time and ambient light handling.

Innovation Solution

An image processing apparatus that generates gradation conversion conditions tailored to each workpiece image, optimizing gradation values within a predetermined range by separating and converting low-frequency and high-frequency components, and allowing user input for parameter customization.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Illumination intensity

If multi-shot HDR is used to expand dynamic range, then dynamic range is improved, but tact time increases due to multiple imaging of the same workpiece

Engineering Contradiction:
Improvedynamic rangeVSAvoidtact time
Core Design Contradiction:
Illumination intensityVSProductivity

Solution Approach 1:

The patent applies preliminary action by pre-determining the tone mapping curve before imaging. The image generation unit determines the tone mapping curve based on the captured workpiece image and then applies this pre-determined curve to perform gradation conversion, eliminating the need for multiple shots while achieving HDR效果.

Inventive Principle:
Principle #10Preliminary action

2Productivity

If one-shot HDR with pre-determined tone mapping is used to maintain tact time, then productivity is improved, but robustness decreases when workpiece posture or ambient light changes

Engineering Contradiction:
Improvetact timeVSAvoidrobustness
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent applies dynamics by making the tone mapping curve adaptive rather than fixed. The image generation unit dynamically determines the tone mapping curve based on the actual captured workpiece image, allowing the system to adapt to different workpiece postures and ambient light conditions while maintaining one-shot imaging capability.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent applies feedback by using the captured workpiece image itself to determine the tone mapping curve. The image generation unit analyzes the actual image data and adjusts the tone mapping parameters accordingly, creating a feedback loop that enhances robustness against varying conditions.

Inventive Principle:
Principle #23Feedback

3Speed

If pre-determined tone mapping is applied to all workpiece images, then processing speed is improved, but manufacturing precision decreases due to insufficient adaptation to individual images

Engineering Contradiction:
Improveprocessing speedVSAvoidgradation conversion accuracy
Core Design Contradiction:
SpeedVSManufacturing precision

Solution Approach 1:

The patent applies local quality by tailoring the tone mapping curve to each individual workpiece image rather than using a universal curve. The image generation unit determines separate tone mapping curves for different images based on their specific characteristics, ensuring optimal gradation conversion for each case while maintaining efficient one-shot processing.

Inventive Principle:
Principle #3Local quality

Data Source

PatentUS12567188B2Image processing apparatus
Publication Date: 2026.03.03 KEYENCE CORP
  • US12567188B2 patent drawing
  • US12567188B2 patent drawing
  • US12567188B2 patent drawing

AI summary

An image processing apparatus includes: an imaging unit which generates workpiece images obtained by imaging a workpiece; and an image generation unit which generates an output image having a gradation value in a predetermined gradation range by converting the workpiece images sequentially input from the imaging unit. The image generation unit generates, for each of the input workpiece images, a gradation conversion condition for converting a first gradation value in the input workpiece image into a minimum gradation value in the predetermined gradation range and converting a second gradation value in the workpiece image into a maximum gradation value in the predetermined gradation range, and generates, for each of the input workpiece images, the output image having the gradation value in the predetermined gradation range by performing gradation conversion on each of the gradation values of the workpiece image using the gradation conversion condition corresponding to the workpiece image.