# US 2010/0163732 A1 · Method for the automatic inspection of a welding seam by means of heat flow thermography

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2010/0163732 A1
- Document ID: US20100163732A1
- Status: Published application
- Kind: A1
- Application number: 12/667,683
- Filed: 2008-07-03
- Published: 2010-07-01
- Earliest claimed: 2007-07-04
- Assignees: Thermosensorik GmbH
- Inventors: Roman Louban; Christoph Döttinger; Peter Stolz; Manfred Böhm
- Web page: https://patentgrep.com/patent/US20100163732A1

## CPC

- G01N 25/72: Investigating presence of flaws
- G01N: INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES (measuring or testing apparatus or processes other than immunoassay, involving enzymes or microorganisms C12M, C12Q)

## Abstract

In a method for the automatic inspection of a welding seam using heat flow thermography, a feature vector is established which represents a time course of a detected heat flow. The feature vector is used to determine, from a series of thermal images, a first characteristic thermal image, which corresponds to a minimum heat flow through an object to be examined, and a second characteristic thermal image, which corresponds to a maximum heat flow through the object, wherein a heat flow directly from an excitation source has already dissipated. A suitable thermal image is used from the series of thermal images to detect and evaluate the welding seam in relation to defects of various defect types, for each defect type, the characteristic thermal images being used as references to determine the respective suitable thermal image.

## Claims

### Claim at position 1

1-15. (canceled)

### Claim 16

A method for the automatic inspection of a welding seam using heat flow thermography,
a. in which an object (1), which is to be examined, with a welding seam (4) is provided,
b. wherein the object (1) is excited by means of at least one excitation source (9), and
c. the heat flow (11) produced by the excitation is detected in a series of thermal images using at least one infrared sensor (10),
wherein
d. a feature vector (W(N)) is produced, which is a time course of the detected heat flow (11),
e. wherein a first characteristic thermal image (T1) is dynamically determined from the series of thermal images using the feature vector (W(N)), and the first characteristic thermal image (T1) has a minimum heat flow (Wmin) through the object (1), and
f. a second characteristic thermal image (T2) is dynamically determined from the series of thermal images using the feature vector (W(N)), and the second characteristic thermal image (T2) has a maximum heat flow (Wmax) through the object (1), and
g. wherein a heat flow (13) directly from the excitation source (9) to the characteristic thermal images (T1, T2) has already dissipated, if it is apparent from the series of thermal images, and
h. a suitable thermal image (TF1, TF2, TF3, TF4) from the series of thermal images is used to detect and evaluate the welding seam (4) with respect to defects (5, 6, 7, 8) of various defect types, for each defect type, the characteristic thermal images (T1, T2) being used as references to determine the respective suitable thermal image (TF1, TF2, TF3, TF4).

### Claim 17 (depends on 16)

A method according to claim 16, wherein the second characteristic thermal image (T2) is used for detecting defects (5) of a first defect type as the suitable thermal image (TF1), the defects (5) of the first defect type being geometric defects of the welding seam (4).

### Claim 18 (depends on 16)

A method according to claim 16, wherein a start image (TS) being used for detecting defects (5) of a first defect type is dynamically detected in such a way,
a. that the start image (TS) is located between the first and the second characteristic thermal image (T1, T2) and has a heat flow (WS), which exceeds a first threshold value (Wstart—thd) in the time course of the detected heat flow (11) for the first time, wherein for the first threshold value (Wstart—thd):
Wstart—thd=Wmin+(Wmax−Wmin)·θ,
applies, wherein
Wstart—thd means first threshold value,
Wmin means minimum heat flow, having the first characteristic thermal image (T1),
Wmax means maximum heat flow, having the second characteristic thermal image (T2),
θ means separation factor,
b. wherein a part series of suitable thermal images from the series is selected from the start image (TS) and a series length (NS),
c. wherein a result image is obtained from the part series, and
d. wherein the defects (5) of the first defect type are geometric defects of the welding seam (4) and the result image is used to detect and evaluate them.

### Claim 19 (depends on 18)

A method according to claim 18, wherein the separation factor (θ) is determined in the range between 0 and 1.

### Claim 20 (depends on 18)

A method according to claim 18, wherein the separation factor (θ) is determined in the range between 0.1 and 0.15.

### Claim 21 (depends on 18)

A method according to claim 18, wherein at least two images from an image quantity, which contains the first characteristic thermal image (T1), the second characteristic thermal image (T2) and the start image (TS), are used as references to dynamically determine the series length (NS).

### Claim 22 (depends on 16)

A method according to claim 16, wherein a thermal image, which was detected before the first characteristic thermal image (T1) and has a maximum heat flow (WAmax) of the excitation source (9) is used for detecting defects (6) of a second defect type as a suitable thermal image (TF2), the defects (5) of the second defect type being continuous defects of the welding seam (4).

### Claim 23 (depends on 16)

A method according to claim 16, wherein a thermal image, which is located between the first and the second characteristic thermal image (T1, T2) and has a heat flow (WF3), which exceeds a second threshold value (Winside—thd) in the time course of the detected heat flow (11) for the first time, is used for detecting defects (7) of a third defect type as the suitable thermal image (TF3), wherein for the second threshold value (Winside—thd):
Winside—thd=Wmin+(Wmax−Wmin)·ζ,
applies, wherein
Winside—thd means second threshold value,
Wmin means minimum heat flow, having the first characteristic thermal image (T1),
Wmax means maximum heat flow, having the second characteristic thermal image (T2),
ζ means first heat flow factor,
the defects (7) of the third defect type being defects located in the interior of the welding seam (4).

### Claim 24 (depends on 23)

A method according to claim 23, wherein the first heat flow factor (ζ) is determined in the range between 0 and 1.

### Claim 25 (depends on 23)

A method according to claim 23, wherein the first heat flow factor (ζ) is determined in the range between 0.6 and 0.9.

### Claim 26 (depends on 16)

A method according to claim 16, wherein a thermal image, which is located between the first and the second characteristic thermal image (T1, T2) and has a heat flow (WF4), which exceeds a third threshold value (Wsurf—thd) in the time course of the detected heat flow (11) for the first time, is used for detecting defects (8) of a fourth defect type as the suitable thermal image (TF4), wherein for the third threshold value (Wsurf—thd):
Wsurf—thd=Wmin+(Wmax−Wmin)·δ,
applies, wherein
Wsurf—thd means third threshold value,
Wmin means minimum heat flow, having the first characteristic thermal image (T1),
Wmax means maximum heat flow, having the second characteristic thermal image (T2), and
δ means second heat flow factor,
the defects (8) of the fourth defect type being surface defects of the welding seam (4).

### Claim 27 (depends on 26)

A method according to claim 26, wherein the second heat flow factor (δ) is determined in the range between 0 and 1.

### Claim 28 (depends on 26)

A method according to claim 26, wherein the second heat flow factor (δ) is determined in the range between 0.1 and 0.4.

### Claim 29 (depends on 16)

A method according to claim 16, wherein at least one of the images (TF1, TF2, TF3, TF4), which are used to detect and evaluate the welding seam (4) in relation to defects of various defect types, is converted into an 8 bit image, wherein
a. a histogram (n(I)) of the image ((TF1, TF2, TF3, TF4) is produced,
b. a first portion (ΔIobj—rel) relevant to information of a first intensity region (ΔIobj), which corresponds to an image object (16), is determined from the histogram (n(I)),
i. wherein a first lower limit value (Iobj—bot—thd) and a first upper limit value (Iobj—top—thd) are dynamically determined, and
ii. wherein the first limit values (Iobj—bot—thd, Iobj—top—thd) delimit first characteristic area portions (Fobj—bot—thd, Fobj—top—thd) of an area (Fobj) of the histogram (n(I)) corresponding to the first intensity region (ΔIobj),
c. a second portion (ΔIbgrd—rel) relevant to information of a second intensity region (ΔIbgrd), which corresponds to an image background (17), is determined from the histogram (n(I)),
i. wherein a second lower limit value (Ibgrd—bot—thd) and a second upper limit value (Ibgrd—top—thd) are dynamically determined, and
ii. wherein the second limit values (Ibgrd—bot—thd, Ibgrd—top—thd) delimit second characteristic area portions (Fbgrd—bot—thd, Fbgrd—top—thd) of an area (Fbgrd) of the histogram (n(I)) corresponding to the second intensity region (ΔIbgrd), and
d. the portions (ΔIobj—rel, ΔIbgrd—rel) relevant to information are used as a basis for the conversion of the image (TF1, TF2, TF3, TF4) to be examined into an 8 bit image.

### Claim 30 (depends on 29)

A method according to claim 29, wherein the first characteristic area portions (Fobj—bot—thd, Fobj—top—thd) in relation to the area (Fobj) corresponding to the first intensity region (ΔIobj) are determined in the range between 0 and 0.5.

### Claim 31 (depends on 29)

A method according to claim 29, wherein the first characteristic area portions (Fobj—bot—thd, Fobj—top—thd) in relation to the area (Fobj) corresponding to the first intensity region (ΔIobj) are determined in the range between 0.05 and 0.4.

### Claim 32 (depends on 29)

A method according to claim 29, wherein the second characteristic area portions (Fbgrd—bot—thd, Fbgrd—top—thd) in relation to the area (Fbgrd) corresponding to the second intensity region (ΔIbgrd) are determined in the range between 0 and 0.5.

### Claim 33 (depends on 29)

A method according to claim 29, wherein the second characteristic area portions (Fbgrd—bot—thd, Fbgrd—top—thd) in relation to the area (Fbgrd) corresponding to the second intensity region (ΔIbgrd) are determined in the range between 0.05 and 0.4.

### Claim 34 (depends on 29)

A method according to claim 29, wherein the characteristic area portions (Fobj—bot—thd, Fobj—top—thd, Fbgrd—bot—thd, Fbgrd—top—thd) are determined independently of one another.

### Claim 35 (depends on 16)

A method according to claim 16, wherein at least one of the feature vector (W(N)) and the histogram (n(I)) are processed with the aid of at least one morphological filter.
