JP3130460B2 - Internal inspection method for hollow paths - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for inspecting the inside of a hollow path for inspecting the inside of a test object having a bent hollow path from which the inside cannot be observed from the outside.

[0002]

2. Description of the Related Art A water-cooled engine has a cooling water passage (water jacket) for a combustion chamber in a cylinder block and a head cylinder, and circulates and supplies cooling water to this portion to prevent seizure of the engine. 4 (a) and 4 (b) show a bottom view and a partial vertical sectional view of an example of the head cylinder (1), respectively. The head cylinder (1) is shown in FIG.
As shown in (a) and (b), water holes (3) are formed on the lower surface side (cylinder block side) of the head along the circumference of the cylinder tube (2). The above water holes (3) form a circulating cooling water passage by coupling the head cylinder (1) to the cylinder block, and a cooling water passage hollow passage (4) between adjacent water holes which cannot be observed from the outside. To form
A sand hole is formed around the ignition plug and the intake / exhaust valve on the upper surface side of the head, and a hollow passage for a cooling water passage is formed between the sand holes.

The hollow path (4) forms a bent path as shown in FIG. 4 (c). In particular, as shown in FIG. 4 (e), the hollow path (4a) on the upper surface side of the head is meshed. The head cylinder (1) is an example of a test object having a curved hollow path (4) (4a).

When forming the cooling water channel water hole (3) and the hollow channels (4) and (4a), a core is set in a mold, molten metal (for example, a molten aluminum alloy) is poured and solidified. Although the mold and the core are removed to produce a product, the cooling water passage of the head cylinder (1) has a more complicated shape than the cooling water passage of the cylinder block, and core sand is likely to remain without being removed. In particular, since the interior of the hollow passages (3a) and (4a) cannot be observed from the outside, there are many cases where the core sand remains and the process moves to the final process. As a result, in the worst case, the engine is disposed of as a defective product. Sometimes I have to.

[0005] When the core is formed or set in a mold, if the core is partially missing or has a crack, if the core is set in the mold as it is, the poured molten metal loses these defects. As a result of flowing into the places and cracks and remaining as casting burrs, the cooling water passage is blocked, and the cooling function of the engine is deteriorated.

In view of the above, the bent hollow path (4) (4
It is necessary to inspect the inside of a) to check for abnormalities such as abnormal shape and remaining core, and an example of the inspection means is shown below with reference to FIG. The inspection means comprises a pair of first and second photoelectric sensors (5) and (6) for inserting and receiving water holes (3) on the lower surface side of the head cylinder (1) and sand holes on the upper surface side, respectively. Is provided. When the photoelectric sensors (5) and (6) are inserted into the water holes (3) and the light (La) projected from the first photoelectric sensor (5) is received by the second photoelectric sensor (6), for example. , The interior of the hollow path (4) is determined to be normal.

[0007]

The problem to be solved is that the inner surface of the hollow path (4) has a matte-like fine uneven surface, and the path is bent as shown in FIG. 4 (c). As a result, the light is diffusely reflected and does not accurately reach the second sensor (6) on the light receiving side. About 1mm
Because of the diameter of the spot light, only complete occlusion can be detected.
As shown in FIG. 4D, when the core (m) is half-clogged, the spot light sometimes becomes undetectable due to grazing through the half-clogged portion. Further, as shown in FIG. In addition, the hollow path (4a) around the spark plug on the top side of the head is a core (m)
Is that the inspection cannot be carried out not only by visual inspection but also by the photoelectric sensor method.

[0008]

According to the present invention, when inspecting the inside of a test object having a bent hollow path for the presence or absence of an internal abnormality, diffused light is emitted from the entrance of the path and diffused and reflected inside the path. Receiving the captured light at a path exit at a path exit by a predetermined imaging means and capturing an image, binarizing the captured image to extract a binary image related to inspection, To measure the image feature amount, determine the degree of conformity of the measured data to the normal data by fuzzy inference, and inspect the inside of the hollow path for abnormalities.

[0009]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the method for inspecting the inside of a hollow path according to the present invention will be described below with reference to FIGS. First, FIG. 1 (a) is a partial vertical sectional view of the same head cylinder (1) as FIGS. 4 (a) and 4 (b) and an inspection means (7) for implementing the method of the present invention. ) Is a partial horizontal sectional view of the bent hollow path (4) (4a), FIG.
Shows side views of an operation example of the inspection means (7). The inspection means (7) includes a light emitting unit (8) and an imaging unit (9). The light projecting section (8) is composed of an optical fiber connected to a light source of diffused light, and projects diffused light (Lb) from, for example, a water hole (3) into a hollow path (4). The imaging unit (9) is composed of a camera, which enters through a water hole (3) and has a hollow path (4).
Light reflected irregularly in the inside and emitted from another water hole (3) is received and imaged on a two-dimensional plane.

The operation of the present invention based on the above configuration will now be described. First, when the head cylinder (1) is to be inspected and the inside of the hollow path is inspected for abnormality, as shown in FIG. 1A, for example, an optical fiber (8) is inserted into a water hole (3) on the lower surface side of the head. Insert the light emitting end. When the diffused light (Lb) is projected into the hollow path (4), the light diffusely reflects in the path (4) and exits from another water hole (3) again. Therefore, a plurality of water holes (3) ...
The light (Lb) emitted from the water holes (3)... Is received and imaged in a two-dimensional plane. Then, as shown in FIG. 1B, even if the inside of the path (4) is bent, the diffused light (Lb) passes through the path (4) after being irregularly reflected. Light can be accurately received by the camera (9). And since the reflectance of the core (m) is small,
As shown in FIG. 1D, when the core (m) remains in the hollow path (4), the brightness area value of the light (Lb) emitted from the hollow path (4) after being irregularly reflected. And the perimeter is shrinking. Therefore, the brightness area value and the perimeter of each route exit are determined by fuzzy inference to check for an abnormality in the hollow route (4).

At the time of the fuzzy inference, a binarization threshold (H) is automatically set in accordance with an object to be inspected prior to discrimination, and a binarized light having a luminous intensity equal to or higher than a predetermined level from an image captured by the camera (9). A valued image is taken out and inspected based on features such as the area and perimeter of the image. Therefore, in the above threshold setting, first, a distribution of feature amounts such as an area and a perimeter of a plurality of reference images of a test object is measured in advance, and a membership function of fuzzy inference is created for each path exit. deep. Further, as shown in FIGS. 2A and 2B, a plurality of binarization thresholds (Ha) (Ha) ( Hb)
Set (Hc). Then, in the case of a non-defective product, as the level of the threshold value (H) decreases, the area of the binarized image of each level gradually increases. As shown in FIG.

Therefore, the plane image picked up by the camera (9) is first binarized by the maximum threshold (Ha), and a binarized image (Da) shown in FIG. 2 (c) is extracted. The binarized image area (Sa) at that time is determined for the degree of conformity to the corresponding reference area by fuzzy inference or the like.
The threshold value (Ha) is set as the optimum threshold value (Ho), and the process proceeds to the next inspection for the presence or absence of an internal abnormality. When the binarized image area (Sa) is too large than the reference area, the level of the maximum threshold value (Ha) is set to be larger and readjusted.

Next, when the binarized image area (Sa) is smaller than the reference area, the inspected object is defective or the luminance distribution as shown by the dotted line in FIG. This is a case where even if the product is a good product, the inner surface of the hollow path is rough or satin-like, so that the amount of emitted light is reduced. In this case, 1
The step level is lowered, and the area of the binarized image is measured again at the next largest threshold (Hb). When the binarized image area (Sb) at that time is larger than the previous binarized image area (Sa) based on the threshold value (Ha), or when the binarized image area substantially coincides with the corresponding reference area, it is determined to be non-defective. The threshold (Hb) at that time is set as the optimal threshold (Ho). Also, if it gets too big,
The threshold slightly raised from the threshold (Hb) is set as the optimum value. If the binarized image area (Sb) is constant or becomes 0, it is determined to be defective, and the level is lowered by one step and re-measured at the threshold (Hc). The above operation is repeated plural times, for example, about five times, and the binarized image area is constant or
If it becomes 0, it is determined to be defective, and the threshold value is set again. In this manner, the binarized image area is determined, and the threshold level is automatically adjusted as appropriate to set an optimal binarized threshold (Ho).

Although not shown, there is also a means for directly setting the optimum threshold value (Ho) using the center of gravity calculation of fuzzy inference. For example, the membership functions of the binarized image area and the threshold are set as the input / output unit. Next, FIG.
The binarized image area is extracted from the luminance distribution shown in FIG. 7A by a predetermined threshold value, and the degree of conformity is determined from the membership function of the input unit. Then, the optimum degree (Ho) is calculated by substituting the degree of conformity into the membership function of the output unit and calculating the center of gravity of the corresponding combined trapezoid area.

After setting the above-mentioned optimal binarization threshold (Ho), next, as shown in FIG.
Then, the binarized image (Do) is extracted from the image captured by the camera (9) based on, and the feature amount of the image (Do), for example, the area and the perimeter are measured. Then, as shown in FIG. 3B, a membership function (Ma) of fuzzy inference is set in advance for each different hollow path exit, and for example, measured area data is substituted into the membership function (Ma). Therefore, the degree of conformity (α) of the measured data to the normal data is derived from the membership function (Ma) of fuzzy inference.
Therefore, the degree of conformity (α) is compared with the reference value (A), and when α> A, it is determined as a non-defective product and the presence or absence of an abnormality in the hollow path is inspected. At this time, a plurality of hollow path exits may be imaged at once, and the suitability of the plurality of images for each exit may be combined, and for example, the presence or absence of an abnormality may be determined from the multiplied value. In addition, by adding the center of gravity of the image and the like as discriminating factors in addition to the area and the perimeter, the inspection accuracy can be improved.

The membership function (Ma) measures the area, perimeter, etc. of a plurality of non-defective or defective workpieces, and calculates (N
The probability distribution is drawn with a) as a normal region and (Nb) and (Nc) as abnormal regions. The probability distribution is created for each route exit and for each feature amount such as area and perimeter. For example, when the measured area in the area distribution is (Pa), the degree of conformity to the normal data is (Qa), and the presence or absence of an abnormality in the hollow path is determined based on that. The region (Na) indicates an abnormal region caused by a decrease in the amount of light due to the remaining core, and the region (Nb) indicates an abnormal region caused by an increase in the amount of light due to the remaining cast flash.

Alternatively, when there are a plurality of hollow path outlets,
The plurality of outlets are imaged at a time, and a binarized image including a plurality of outlet images in one image is taken out. Then, based on the image data, for example, the presence or absence of an abnormality in the hollow path (4) may be inspected by fuzzy inference using, for example, the total area of a plurality of images or the ratio of each image to the total area.

As shown in FIGS. 1 (c) and 4 (e), when the inside of the hollow path (4a) is complicated and complicated in a bent maze, light projection from only one direction requires a mesh ( Since the abnormality in 4b) may not be reflected at all in the captured image, it is not possible to sufficiently determine the presence or absence of the abnormality in the route (4a). Therefore, for example, the light emitting direction of the light emitting unit (8) is changed, and light is emitted from a plurality of different angles. Therefore, first, light is projected from one direction toward one side surface (4c) in the hollow path (4a), and the path (4
The light passing through a) and the mesh (4b) is imaged.
Similarly, from the other direction, the other side (4
Light is emitted toward d), within the path (4a) and its mesh (4b).
The light passing through the inside is imaged. Then, the hollow path (4
By projecting light on the two different side surfaces (4c) and (4d) of a), two different captured images are obtained, and the brightness area value and the like differ depending on each image. Then, the binarized images of the captured images are combined, and the characteristics such as the brightness area value are compared and determined, and the presence or absence of an abnormality in the path (4a) is determined by fuzzy inference.

Further, the respective positions of the light projecting section (8) and the image pickup section (9) are shifted to obtain two different picked-up images.
The presence or absence of an abnormality in the path (4a) may be determined by combining the binarized images of the captured images, and the light projection direction and position may be changed.

[0020]

According to the present invention, when inspecting for an internal abnormality of a hollow path in a test object having a hollow path such as an engine head cylinder, a diffused light is emitted to the entrance of the hollow path and the inside of the path is inspected. At the exit of the path, the light that has passed through the diffuse reflection is received and imaged on a two-dimensional plane, the image features are measured from the binarized image of the captured image, and the fitness of the measured data to the normal data is determined by fuzzy inference. Since the inside of the hollow path is inspected for abnormalities, the light is received on a two-dimensional plane, so that the light diffusely reflected can be accurately received even in the bent hollow path. In addition, since feature amounts such as area and perimeter are measured from the binarized image and the presence or absence of an internal abnormality is determined from the measured data by fuzzy inference, erroneous determination can be prevented and accurate determination can be made.

[Brief description of the drawings]

FIG. 1A is a vertical sectional view of a main part of a head cylinder as an example of an inspection unit and an object to be inspected, showing an embodiment of a method for inspecting the inside of a hollow path according to the present invention. (B) is a partial horizontal sectional view of the head cylinder. (C) is another partial horizontal sectional view of the head cylinder. (D) is a vertical sectional view of an essential part showing an operation example of the method for inspecting the inside of a hollow path according to the present invention.

FIG. 2A is a graph showing an example of a luminance distribution of a captured image of a test object and several types of binarization thresholds. (B)
Is a graph showing another example of the luminance distribution of the captured image of the test object and several types of binarization thresholds. (C) is FIG.
It is a figure which shows an example of the image binarized by the threshold value of (a) and (b). (D) is a diagram showing a case where an image cannot be taken out with the threshold values of FIGS. 1 (a) and (b).

FIG. 3A is a diagram illustrating an example of a binarized image.
(B) is a waveform diagram showing an example of a fuzzy judgment membership function.

FIG. 4A is a bottom view of a head cylinder which is an example of an inspection object. FIG. 4B is a partial vertical sectional view of the head cylinder shown in FIG. FIG. 4C is a partial horizontal sectional view of the head cylinder shown in FIG. (D) is a partial vertical sectional view of the head cylinder for explaining the problem of the present invention. FIG. 5E is another partial horizontal sectional view of the head cylinder shown in FIG.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Inspection object (head cylinder) 4, 4a Hollow path 7 Inspection means 8 Light emitting part 9 Imaging part Lb Diffusion light

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) G06T 1/00 G06T ⁷ /00-7/60 F02F 1/24-1/42 G01B 11/24-11 / 255 G01N 21/49-21/53 G01N 21/88-21/958

Claims (4)

(57) [Claims] When inspecting for an internal abnormality in a test object having a bent hollow path, diffused light is projected from an entrance of the path, and light that has been diffusely reflected in the path and passed through by a predetermined imaging means. A step of receiving and capturing an image on a two-dimensional plane at a path exit; a step of binarizing the captured image to extract a binarized image for inspection; and performing image processing on the binarized image to obtain an image feature amount And inspecting the inside of the hollow path for abnormalities by determining the degree of conformity of the measured data to the normal data by fuzzy inference. 2. When projecting diffused light at the entrance of a hollow path, the diffused light is projected from a plurality of different directions or positions at different times, and a binarized image is extracted for each diffused light for imaging and inspection. The method for inspecting the inside of a hollow path according to claim 1, wherein: 3. When there are a plurality of exits in a hollow path, the exits are imaged at a time to extract a binarized image including a plurality of exit images in one image, and based on the binarized image data. The method for inspecting the inside of a hollow path according to claim 1 or 2, wherein the inside of the hollow path is inspected for abnormalities. 4. The method according to claim 1, wherein each exit of the plurality of hollow paths is individually imaged, a binarized image for each exit is taken out, and the presence or absence of an internal abnormality in the hollow path is inspected for each exit.
Or the method for inspecting the inside of a hollow path according to 2.