JP2671095B2 - Metal material defect inspection method and apparatus - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention picks up an image of defective sparks caused by defects in a metal material such as norokami, pinholes, sand dents, etc. existing in the surface layer when the surface layer of the metal material is ablated by ablation flame. The present invention relates to a metal material defect inspection method and apparatus for determining the quality of a metal material based on image data subjected to image processing.

[0002]

2. Description of the Related Art Conventionally, this type of metal material defect inspection method, that is, the defect of a metal material based on image data obtained by picking up and processing an image of a defect spark generated at the time of ablation of a surface layer portion of a metal material by an ablation flame There have been proposed several methods for determining.

For example, in the defect inspection method and apparatus for a metal material disclosed in Japanese Patent Laid-Open No. 59-154346,
The component related to the defect spark is discriminated and detected from the photoelectric conversion signal of the captured image of the ablation flame. Also, JP-A-4-35
In the method for inspecting the surface defects of a steel piece disclosed in Japanese Patent No. 0547, the welding flame is imaged at minute time intervals, and the current welding image and the welding image before an appropriate predetermined time are compared and calculated. Defect sparks are extracted from the difference in light intensity between images.

[0004]

By the way, generally, the shape and brightness (luminance) of the ablation flame are not constant over the entire metal material and differ depending on the temperature and composition thereof.

Therefore, in the case of the above-described metal material defect inspection method and metal material defect inspection apparatus, for example, Japanese Patent Laid-Open No. 59-1543.
According to the defect inspection method for a metal material disclosed in Japanese Patent Publication No. 46, it is difficult to discriminate the defective spark from the contour of the fusing flame. or,
Even with the surface defect inspection method for a steel piece disclosed in Japanese Patent Laid-Open No. 4-350547, it is difficult to determine an effective light amount difference for determining a defect spark, and moreover, a plurality of defect sparks are included in the compared ablation images. In such a case, there is a drawback in that the determination cannot be performed accurately.

The present invention has been made to solve the above problems, and its technical problem is to inspect a metal material defect capable of stably detecting a defect spark regardless of the shape and brightness of the ablation flame. A method and an apparatus therefor are provided.

[0007]

According to the present invention, a surface layer portion of a metal material is ablated by a smelting flame, and the quality of the metal material is judged based on a defect spark generated according to a defect in the surface layer portion. In the metal material defect inspection method to be determined, a flame imaging step of obtaining a flame image data by capturing an abrasion flame at predetermined intervals, and setting a specific area for the flame image data, and limiting a limited flame image in the specific area An image processing step of extracting a portion having a large luminance change from the data and further clustering the extracted images to obtain defect spark candidate cluster image data, and the defect spark candidate cluster image data is surrounded by a minimum rectangular frame and the minimum rectangle A metal material defect inspection method including a defect determination processing step in which a frame is regarded as a defect spark when the frame is larger than a predetermined size.

According to the present invention, in the above metal material defect inspection method, in the image processing stage, the flame image data is differentiated and binarized to extract a portion having a large luminance change. can get.

Further, according to the present invention, a flame ablation means for ablating a surface layer portion of the metal material by ablation flame is included, and the metal material of the metal material is detected on the basis of a defective spark generated according to a defect of the surface layer portion. In a metal material defect inspection apparatus for judging pass / fail, a flame imaging means for capturing flame cutting flame at predetermined intervals to obtain flame image data, a specific area is set for the flame image data, Differentiate and binarize the flame image data to extract a portion having a large brightness change, and further cluster the extracted images to obtain defect spark candidate cluster image data, and minimize the defect spark candidate cluster image data. An image processing unit which regards as a defective spark when the minimum rectangular frame is surrounded by a rectangular frame and has a predetermined size or more, the flame image data, the limited flame image data, and the defective spark candidate cluster image Metallic material defect inspection apparatus and a display means for displaying image data is obtained.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The metal material defect inspection method and apparatus according to the present invention will be described in detail below with reference to the drawings.

First, the outline of the metal material defect inspection method of the present invention will be briefly described. This metal material defect inspection method is for smelting the surface layer portion of the metal material with a smelting flame, and determining the quality of the metal material based on the defect sparks generated according to the defects in the surface layer portion. Speaking of this, a flame imaging stage in which the ablation flame is imaged at predetermined intervals to obtain flame image data, and a specific area is set for the flame image data, and a large change in brightness occurs from the limited flame image data in this specific area. An image processing step of extracting a part and further clustering the extracted images to obtain defect spark candidate cluster image data, and enclosing the defect spark candidate cluster image data in a minimum rectangular frame, and the minimum rectangular frame has a predetermined size. In the case of the above, it includes a defect determination processing step which is regarded as a defect spark.

FIG. 1 (a) shows a partial side view of the basic construction of a metal material defect inspection apparatus incorporating such a metal material defect inspection method, and FIG. 1 (b) shows the same. It is the thing which showed the metal material detection part in an inspection device by the top view.

This metal material defect inspection apparatus includes a torch (gas burner) 3 as a flame ablation means for ablating a surface layer portion of a metal material 1 movably placed on a conveyor with a ablation flame 3a, Fused portion 1a formed on metal material 1 by fusing
And the camera 2 as flame imaging means for obtaining flame image data by capturing the abrasion flame 3a in the imaging visual field region 2a and setting a specific region for the flame image data input via the branching device 4. Then, the limited flame image data in the specific area is differentiated and binarized to extract a portion having a large luminance change, and the extracted image is clustered to obtain defect spark candidate cluster image data, and the defect spark is also acquired. An image processing device 5 which regards candidate cluster image data as a defective spark when the candidate cluster image data is surrounded by a minimum rectangular frame and the minimum rectangular frame has a predetermined size or more, and flame image data, limited flame image data, and defective spark A display device (monitor) 6 for displaying the candidate cluster image data as an image is provided.

Of these, the operation panel 8 is the image processing apparatus 5.
The branch device 4 is connected to a video device (VTR) 7 for recording and reproducing an abrasion flame image as required. In this inspection apparatus, the metal material 1, the camera 2, and the torch 3, which are inspection samples, are installed at the inspection site, and the other branching devices 4,
The image processing device 5, the display device 6, the video device 7, and the operation panel 8 are installed in a data processing management room or the like. Further, this metal material defect inspection apparatus may be provided with a judgment result output device (not shown) for outputting the defect judgment result to the outside.

Next, referring to FIG. 1A, the inspection processing operation in this inspection apparatus will be briefly described. However, the operation of this inspection apparatus is controlled by a fusing start signal S1 and fusing end signal S2 output from a metal material movement control device (not shown) for controlling the movement of the metal material 1.

First, a metal material 1 as an inspection sample which is movably placed on a carrier is moved at a constant speed to the direction shown in the figure, and its surface layer portion is abraded by the abrading flame 3a from the torch 3. To be done. The camera 2 is integrally fixed to the torch 3, and the ablation flame 3a in the ablation part 1a of the metal material 1 is housed in the imaging visual field area 2a so that the entire ablation area of the metal material 1 is 0.1 seconds. Images are taken continuously at regular intervals.

The imaging cycle defined here is the metal material 1
This is because the emission time of the defect spark generated according to the defect of the surface layer becomes 0.1 seconds or more. Further, the position of the camera 2 is moved up, down, left and right so as to be set at an appropriate position so that the abrasion flame 3a is located substantially at the center in the imaging visual field region 2a.

An abrasion flame image from the camera 2 is transmitted to the image processing device 5 via the branching device 4, and is input as flame image data as shown in FIG.

The image processing apparatus 5 sets a specific area for the flame image data relating to the imaging visual field area 2a including the ablation flame 3a, and differentiates and binarizes the limited flame image data in the specific area. The part with a large brightness change is extracted. At this time, the shape (visible as the size and area) and the brightness of the ablation flame 3a differ depending on the type of the metal material 1 to be inspected, the surface temperature, and the like. The setting of the area 3b can be arbitrarily set by an operation input from the operation panel 8.

Subsequently, the image processing apparatus 5 expands, contracts, and clusters only a portion of the specific region 3b having a large luminance change, and the defect spark candidate cluster image data X and X'which are candidates for defect sparks. To get Further, in the image processing device 5, as shown in FIG. 3, the defect spark candidate cluster image data X, X'in the limited frame 6b corresponding to the specific area.
Is surrounded by the smallest rectangular frames 6c and 6c ', and of these smallest rectangular frames 6c and 6c', for example, the smallest rectangular frame 6c having a predetermined size or more is regarded as a defect spark, and is smaller than the predetermined size. The minimum rectangular frame 6c ′ is regarded as noise and is ignored. Thereby, the cutting flame 3a
It is possible to accurately detect the defect spark after eliminating the noise regardless of the shape and the brightness.

[0021]

As described above, according to the present invention,
The defect spark candidate cluster image data included in the image of the ablation flame is surrounded by the minimum rectangular frame, and if the size exceeds a predetermined size in the minimum rectangular frame, it is considered as a defective spark, so the shape of the ablation flame is It is possible to provide a metal material defect inspection method and apparatus capable of stably detecting a defective spark regardless of the type of metal material regardless of brightness.

[Brief description of the drawings]

FIG. 1 (a) is a partial side view showing a basic configuration of a metal material defect inspection apparatus incorporating a metal material defect inspection method of the present invention, and FIG. 1 (b) is a metal material in the inspection apparatus. It is the thing which showed the detection part by the top view.

FIG. 2 is a view showing a state of a specific area with respect to flame image data displayed on a display screen of a display device provided in the metal material defect inspection apparatus shown in FIG.

FIG. 3 shows a state of a defect spark candidate cluster image included in limited flame image data image-displayed on a display screen of a display device included in the metal material defect inspection device shown in FIG. 1 (a). .

[Explanation of symbols]

 1 Metal Material 2 Imaging Camera 2a Imaging Field of View 3 Torch (Gas Burner) 3a Welding Flame 4 Branching Device 5 Image Processing Device 6 Display Device (Monitor) 7 Video Device (VTR) 8 Operation Panel

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Katsuichi Suzuki 5-1-1, Shimorenjaku, Mitaka City, Tokyo Nihon Radio Co., Ltd. (56) Reference JP-A-5-87745 (JP, A) JP 61-145458 (JP, A) JP 59-138904 (JP, A) JP 4-350547 (JP, A) JP 59-154346 (JP, A) JP 5-232046 (JP , A)

Claims (3)

(57) [Claims] 1. A method of inspecting a metal material defect, wherein the surface layer portion of a metal material is ablated by a smelting flame, and the quality of the metal material is judged based on a defect spark generated according to a defect in the surface layer portion. A flame image capturing step of obtaining a flame image data by imaging the ablation flame at a predetermined cycle, a specific area is set for the flame image data, and a portion having a large brightness change from the limited flame image data in the specific area is set. An image processing step of extracting and further clustering the extracted images to obtain defect spark candidate cluster image data, and enclosing the defect spark candidate cluster image data in a minimum rectangular frame, and the minimum rectangular frame having a predetermined size or more. And a defect determination processing step of determining the defect spark as the defect spark. 2. The metal material defect inspection method according to claim 1, wherein in the image processing step, the flame image data is differentiated and binarized to extract a portion having a large luminance change. Metal material defect inspection method. 3. A metal material, which comprises a flame smelting means for smelting a surface layer portion of a metal material by a smelting flame, and judges the quality of the metal material based on a defect spark generated according to a defect in the surface layer portion. In the defect inspection device, a flame imaging unit that captures the abrasion flame at predetermined intervals to obtain flame imaging data, and sets a specific area for the flame image data, and sets a limited flame image data within the specific area to On the other hand, differential / binarization extracts a portion having a large luminance change, and further clusters the extracted images to obtain defect spark candidate cluster image data, and encloses the defect spark candidate cluster image data in a minimum rectangular frame. In the case where the minimum rectangular frame has a predetermined size or more, the image processing means that considers the defective spark, the flame image data, the limited flame image data, and the defective spark candidate cluster image data A metal material defect inspection apparatus comprising: a display unit for displaying an image.