IP Library › Granted Patent US 11,132,786
Granted Patent B2
US 11,132,786 · App. 16/134,986 · Granted Sep 28, 2021

Board defect filtering method based on defect list and circuit layout image and device thereof and computer-readable recording medium

Inventors: Ming-Kaan Liang (Hsinchu, TW); An-Chun Luo (Taichung, TW); Yu-Shan Deng (Hsinchu County, TW); Chih-Ming Shen (New Taipei, TW); Ming-Ji Dai (Hsinchu, TW)
Assignee: Industrial Technology Research Institute
G06T7/001G06K9/6267G06N3/08G06T2207/10024G06T2207/20084G06T2207/20132G06T2207/30141
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 11,132,786
App. No.
16/134,986
Granted
Sep 28, 2021
Kind
B2
Abstract

A board defect filtering method is provided. The method includes: receiving a defect list; obtaining a plurality of defect images of a plurality of defect records on the defect list; receiving a circuit layout image; analyzing a defect location of a first defect image of the plurality of defect images according to the circuit layout image; cropping the first defect image to obtain a first cropped defect image according to the defect location; inputting the first cropping defect image to a defect classifying model; and determining whether the first defect image is a qualified product image or not according to an output result of the defect classifying model.

Claims (50)

1. A board defect filtering method, comprising:

receiving a defect list;

obtaining a plurality of defect images of a plurality of defect records on the defect list;

receiving a circuit layout image;

analyzing a defect location of a first defect image of the defect images according to the circuit layout image;

cropping the first defect image according to the defect location of the first defect image to obtain a first cropped defect image;

inputting the first cropped defect image to a defect classifying model;

determining whether the first defect image is a qualified product image or an unqualified product image according to an output result of the defect classifying model; and

transmitting records belonging to the unqualified product image in the defect list to a verified repaired system,

wherein if the output result corresponding to the first defect image is overcame in a subsequent industrial manufacturing process, the first defect image is marked as a qualified product and is not transmitted to the verified repaired system,

wherein analyzing the defect location of the first defect image of the defect images according to the circuit layout image comprises:

adjusting a proportion of the first defect image according to a proportion of the circuit layout image;

comparing the circuit layout image and the first defect image to obtain a location of the first defect image relative to the circuit layout image; and

correcting the location of the first defect image relative to the circuit layout image to obtain the defect location.

2. The board defect filtering method as claimed in claim 1 , wherein inputting the first cropped defect image to the defect classifying model comprises:

inputting the first defect image to a convolutional neural network model to extract a defect feature of the first defect image; and

classifying the defect feature according to a set of reference image features to generate the output result.

3. The board defect filtering method as claimed in claim 1 , wherein inputting the first cropped defect image to the defect classifying model comprises:

extracting a defect profile of the first cropped defect image;

analyzing the first cropped defect image according to the defect profile, so as to generate a defect quantifying feature; and

classifying the defect quantifying feature according to a set of reference image features to generate the output result.

4. The board defect filtering method as claimed in claim 1 , further comprising:

analyzing a defect color level of the first defect image;

determining whether the defect color level is the same as a color of a board material or a color of a wiring material of the circuit layout image; and

determining that the first defect image is the unqualified product image when the defect color level is the same as the color of the board material or the color of the wiring material.

5. A non-transitory computer readable recording medium, storing at least one program code, wherein when an electronic device loads and executes the program code, the board defect filtering method as claimed in claim 1 may be completed.

6. A board defect filtering device, comprising:

an input circuit, receiving a defect list and a circuit layout image, wherein the defect list comprises a plurality of defect records, the input circuit further obtains a plurality of defect images of the defect records;

a memory; and

a central processing unit, connected to the input circuit and the memory, wherein the central processing unit configured to:

analyze a defect location of a first defect image of the defect images according to the circuit layout image, and cropping the first defect image according to the defect location of the first defect image to obtain a first cropped defect image;

input the first cropped defect image to a defect classifying model;

determine whether the first defect image is a qualified product image or an unqualified product image according to an output result of the defect classifying model; and

transmit records belonging to the unqualified product image in the defect list to a verified repaired system,

wherein if the output result corresponding to the first defect image is overcame in a subsequent industrial manufacturing process, the first defect image is marked as a qualified product and is not transmitted to the verified repaired system,

wherein the central processing unit is further configured to:

adjust a proportion of the first defect image according to a proportion of the circuit layout image,

compare the circuit layout image and the first defect image to obtain a location of the first defect image relative to the circuit layout image, and

correct the location of the first defect image relative to the circuit layout image to obtain the defect location.

7. The board defect filtering device as claimed in claim 6 , wherein the central processing unit is further configured to:

input the first defect image to a convolutional neural network model to extract a defect feature of the first defect image, and

classify the defect feature according to a set of reference image features to generate the output result.

8. The board defect filtering device as claimed in claim 6 , wherein the central processing unit is further configured to:

extract a defect profile of the first cropped defect image,

analyze the first cropped defect image according to the defect profile, so as to generate a defect quantifying feature, and

classify the defect quantifying feature according to a set of reference image features to generate the output result.

9. The board defect filtering device as claimed in claim 6 , wherein the central processing unit is further configured to:

analyze a defect color level of the first defect image,

determine whether the defect color level is the same as a color of a board material or a color of a wiring material of the circuit layout image, and

determine that the first defect image is the unqualified product image when the defect color level is the same as the color of the board material or the color of the wiring material.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Sep 20, 2018
From: LIANG, MING-KAAN; LUO, AN-CHUN; DENG, YU-SHAN; SHEN, CHIH-MING; DAI, MING-JI
To: INDUSTRIAL TECHNOLOGY RESEARCH INSTITUTE
Reel/Frame 046917/0647 →
Priority Claims (1)
TW 107122979 · Jul 3, 2018 · national
Continuity (2)
Provisional Application 62613778 · Jan 5, 2018
Related Publication 20190213725A1 · Jul 11, 2019