Industry Applications

Machine Vision Industry Applications

Table of Contents

Auto Parts & Manufacturing

This is to find car doors in different directions and positions. The first model is defined in the plane area of the car door, and the reference pose is calculated by using the image with the calibration plate in the object plane.

Figure a) Calibration of the standard automobile door
Figure b) Car door with location and 3D dimensions

Electronic Components & Supplies

Figure A) Models of ROIs and dip switches
Figure B) Location of automatically detected dip switch
Inspection of power supply cable

This task is to find the power supply cables and extract and inspect the center of each cable with the width It has been inspected with subpixel accuracy.

Figure) Automatically detect the center of each cable with width

Agriculture, Food

This action recognizes the expiration date printed on the bottle and automatically checks it

Picture) Date Recognition

Medical Services and Life Sciences

Particle analysis
The task is to automatically extract and analyze faint tiny particles from the liquid.

Picture a) Original image
b) Extracted small particles

Angiography

Figure a) X-ray image of the heart
b) Automatically extracted blood vessels

This is to extract blood vessels from the x-ray image of the heart as shown. The blood vessels were highlighted using contrast agents. For diagnosis, it is important to extract the width of the blood vessel in order to detect the local narrowing (stenosis). This work was resolved by extracting lines with subpixel accuracy.

Iron, steel and metal

Casting inspection

The task is to inspect the distance between the holes in the castings after rolling the round grooves.

Figure) Automatically measures the distance between casting holes.

Machine

OCR vision check to recognize characters

Picture a) Original image
Picture b) Read characters
Picture) Automatically extract and check the circular outline of the product

Finding a pipe wrench at multiple locations
This task is to locate pipe wrenches based on predefined ROIs.

Figure A) Register ROI for model image
Figure B) Location of pipe wrench detected at different positions

Packing

Automatic inspection of mixed packaging products
This is to check the contents of manually packed products. The second image inspected with reference to the reference image was identified as defective.

Picture a) Correctly packaged reference image

Automatic inspection of mixed packaging products
This is to check the contents of manually packed products. The second image inspected with reference to the reference image was identified as defective.

Automatic location of cookie boxes

This action can be used to determine whether the reference model is derived from the reference image and whether the box is placed upside down.

Figure) Finding a different direction of the cookie box
Figure) Finding a different direction of the cookie box

Print

Barcode recognition of toner cartridges

This operation is to automatically recognize two barcodes. These Code 39 barcodes are very widely used in packaging and many other product description.

Picture) Recognized barcode

Rubber, synthetic materials, metal foil

Inspecting the borders

This is to automatically inspect the outer boundary of the plastic part. In this case, it shows the part protruding from the part that is required to be clean.

Figure) Overhangs and borders detected at the outer boundary

Semiconductor

Combination of balls
This is to automatically check the diameter of the combined balls.

Figure) Automatic measurement of ball bond diameters

IC measurement and inspection

Figure) Find IC leads appearing at various locations and automatically check lead widths and spacings

Lead measurement of moving IC

This task measures the position of the chip’s leads, which can appear at various positions and angles.

Figure a) Set the ROI on the model image
Figure b) Measuring lead in the changed position

Multiple images are merged (Mosaic) to create one big image

The task is to use a standard image size, acquire images from multiple locations, and then use image matching to create a Mosaic or merged image

Figure) Creating a merged mosaic image from multiple overlapping images

Determine mixed substrate location by color

This task is to automatically find all the components on the printed circuit board.

그림) 감지된 IC, 저항기, 캐패시터

Solar energy, renewable energy, recycling

Photovoltaic cell inspection
This operation extracts defective fingers from the solar cell.

Figure: Shattered finger detection of solar cells

www.nextaoi.com
sales@nextaoi.com

Copyright©NEXTAOI Inc. All rights reserved.