JPH09161042A - Method for inspecting inside of hollow route - Google Patents

Abstract

PROBLEM TO BE SOLVED: To judge the existence of internal abnormality in a hollow route by projecting diffused light front the inlet of the route, receiving light passed through the inside of the route with irregular reflection by a prescribed image pickup means and discriminating the feature of the picked-up image. SOLUTION: A projection unit 8 and an image pickup unit 9 are inserted into a pair of sand removing holes formed on the upper face side of a head cylinder, a plane mirror 11 and a convex mirror 12 are oppositely arranged so as to interpose a hollow route 4a formed between the holes between both the mirrors 11, 12 and diffused light is projected from a light emitting source to the mirror 11 through an optical fiber 10. Thereby, light Lb reflected from the mirror 11 is irregularly reflected and passed in the hollow route 4a of which inside is a rugged state and the passed light is received by the mirror 12 in two-dimensional plane. Reflected light from the mirror 12 is picked up by a camera 14 like two-dimensional plane through a lens 13 to obtain the two-dimensional plane image of the irregularly reflected light passed through the route 4a. Since features such as the light quantity and contour of the picked-up image are changed in accordance with the normality/abnormality of the inside state of the hollow route 4a, the features of the picked-up image and discriminated.

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 an inside of a hollow path for inspecting an inside of an object to be inspected having a bent hollow path whose 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) in a combustion chamber in a cylinder block and a head cylinder, and cooling water is circulated to this portion to prevent engine seizure. 5 (a) (b)
FIG. 3C shows a bottom view, a partial vertical cross-sectional view, and a partial horizontal cross-sectional view of the head upper surface side of an example of the head cylinder (1). The head cylinder (1) is shown in FIG.
As shown in (b), a plurality of water holes (3) are formed on the lower surface side of the head (cylinder block side) along the circumference of the cylinder tube (2), and the spark plug and the suction plug on the upper surface side of the head are formed. Sand removal holes (4) are formed around the exhaust valve.
The water holes (3) form a circulating cooling water passage by the head cylinder (1) being connected to the cylinder block,
In addition, a bent cooling water channel hollow path (3a) that cannot be observed from the outside is formed between adjacent water holes. Further, a hollow passage (4a) for a cooling water passage is formed between the sand removing holes on the upper surface side of the head. The head cylinder (1) is an example of an object to be inspected having curved hollow paths (3a) (4a),
In particular, as shown in FIG. 5C, the hollow path (4a) on the upper surface side of the head forms a mesh-like complicated bending labyrinth.

When forming the water hole (3) for the cooling water passage and the hollow passages (3a) (4a), after setting the core in the mold and pouring molten metal (for example, molten aluminum alloy) and solidifying it Although the mold and the core are removed to obtain a product, the cooling water channel of the head cylinder (1) has a more complicated shape than the cooling water channel of the cylinder block, and the core sand is likely to remain without being removed. In particular, since the inside of the hollow paths (3a) and (4a) cannot be observed from the outside, the core sand often moves to the final process with the core sand remaining. As a result, in the worst case, the engine is disposed of as a defective product. There is something we cannot help but do.

When the core is partially missing or has cracks at the time of molding the core or setting it in the mold, if the core is set as it is, the molten metal poured into the core will be damaged. As a result of casting burrs flowing into spots and cracks and remaining as casting burrs, the cooling water passages are blocked and the cooling effect of the engine deteriorates.

Therefore, the hollow path (3a) (4)
It is necessary to inspect the inside of a) to inspect for any abnormalities such as shape abnormality and core remaining. An example of the inspection means will be described below with reference to FIG. The inspection means includes a pair of first and second photoelectric sensors (5) for projecting and receiving light which are respectively inserted into the water holes (3) on the lower surface side of the head cylinder (1) and the sand removing holes (4) on the upper surface side. ) (6) is provided. Then, for example, when the photoelectric sensors (5) and (6) are inserted into the water holes (3), respectively, and the light (La) projected from the first photoelectric sensor (5) is received by the second photoelectric sensor (6). , The inside of the hollow path (3a) is determined to be normal.

[0006]

The problem to be solved is that the inner surface of the hollow path (3a) has a fine textured surface with a satin finish, so that the light is diffusely reflected and the second sensor (6) on the light receiving side is accurate. The quality of the light from the first photoelectric sensor (5) for light projection is about 1 mm.
Due to the spot light of the diameter, only complete occlusion can be detected.
As shown in (b), when the core (m) is half clogged, the spot light sometimes passes through the half clogged part and becomes undetectable. Furthermore, as shown in FIG. 5 (c). In addition, since the hollow path (4a) around the spark plug on the upper surface side of the head is a complicated mesh-shaped bending maze, the core (m) is inside the mesh (4c).
Even if it is clogged, the inspection cannot be performed not only by visual inspection but also by a photoelectric sensor method.

[0007]

According to the present invention, when inspecting the inside of an object to be inspected having a bent hollow path for abnormality, diffused light is projected from the path entrance and diffusely reflected in the path to pass. It is characterized in that the predetermined light is received by a predetermined imaging means at the exit of the path on a two-dimensional plane, and the feature of the picked-up image is identified to determine the presence or absence of an internal abnormality of the hollow path.

[0008]

BEST MODE FOR CARRYING OUT THE INVENTION 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. The present invention projects diffused light from an entrance of a path in an inspection object having a bent hollow path, and two-dimensionally images the light diffusely reflected in the path and passed through the exit of the path to determine the brightness and the like. The feature is that the feature of the captured image is identified to determine whether or not there is an abnormality inside the route. Therefore, first, FIG. 1 (a)
Is a partial vertical sectional view of the same head cylinder (1) as in FIGS. 5 (b) and 5 (c) and an inspection means for carrying out the method of the present invention. FIG. 1 (b) shows the head upper surface side and the inspection means (7). FIG. 2 shows a partial horizontal sectional view, and FIG. 2 shows a side view of the inspection means (7).

The inspection means (7) is a light projecting unit (8).
And an imaging unit (9), as shown in FIG.
The light projecting unit (8) includes an optical fiber (10) whose upper end is connected to a light emitting element such as a light emitting diode, and a plane mirror (11) attached to the light projecting end at the tip of the optical fiber.
The image pickup unit (9) also includes a convex mirror (12) arranged to face the plane mirror (11) and a lens (1) for reflecting light reflected by the convex mirror (12).
It has a built-in ultra-compact camera (14) such as a CCD camera which receives a light on a two-dimensional plane via 3) and takes an image.

The operation (method) of the present invention based on the above configuration will be described below. First, the light projection unit (8) and the image pickup unit (9) are inserted into the pair of sand removing holes (4) on the upper surface side of the head cylinder (1) to sandwich the hollow path (4a) between the holes. (11) and the convex mirror (12) are arranged so as to face each other, and diffused light is projected from the light emitting source to the plane mirror (11) via the optical fiber (10). Then, the light (Lb) reflected from the plane mirror (11) is irregularly reflected and passes through the hollow path (4a) having a satin inner surface, and the passing light is received by the convex mirror (12) on the two-dimensional plane. Then, the reflected light of the convex mirror (12) is two-dimensionally imaged by the camera (14) through the lens (13), and an image of the irregularly reflected light passing through the hollow path (4a) by the two-dimensional plane is obtained. obtain. Then, since the features such as the light quantity and the outline of the captured image vary depending on whether the inside of the hollow path (4a) is normal or abnormal, the features of the captured image are identified and, for example, light having a predetermined brightness or more has a predetermined area. When detected above, it is determined that the inside of the hollow path (4a) is normal. Further, with respect to the hollow path (3a) of the water hole (3), it is possible to identify the feature of the planar imaged image in the hollow path (3a) in the same manner to determine whether there is an abnormality inside the path.

Here, when the discrimination test is performed by, for example, fuzzy inference, 2 for imaging is performed prior to fuzzy inference.
The threshold value (H) is automatically set according to the hollow path which is the inspection object. Then, a binarized image having a luminous intensity equal to or higher than a predetermined level is taken out from the image picked up by the camera (14), and the image is inspected based on the feature amount such as the area and the perimeter of the image. Therefore, in the above threshold setting, first, the distributions of the feature quantities such as the area and perimeter of the images that serve as a plurality of reference images of the object to be inspected are measured in advance, and the membership function of fuzzy reasoning is created for each hollow path. deep. In addition, as shown in FIGS. 3A and 3B, a plurality of binarization thresholds (Ha) (Ha) (in order from the highest in the luminance distribution map (Ba) (Bb) of the image captured by the camera (14). Hb) (Hc) is set. Then, in the case of a non-defective product, the binarized image area of each level gradually increases as the level of the threshold (H) decreases, and conversely, in the case of a defective product, the binarized image area remains constant. Or FIG. 3 (d)
As shown in, the threshold value (Ha) to (Hc) remains 0.

Therefore, the plane image picked up by the camera (14) is first binarized with the maximum threshold value (Ha), and the binarized image (Da) shown in FIG. 3C is taken out. The degree of conformity of the binarized image area (Sa) at that time to the corresponding reference area is determined by fuzzy inference or the like.
Set the threshold value (Ha) as the optimum threshold value (Ho) and proceed to the next inspection for the presence or absence of an internal abnormality. When the binarized image area (Sa) becomes too large than the reference area, the level of the maximum threshold value (Ha) is set to a larger value and readjustment is performed.

Next, when the binarized image area (Sa) is smaller than the reference area, the object to be inspected is a defective product, or the luminance distribution shown by the dotted line in FIG. Even if the product is a non-defective product, the inner surface of the hollow path is rough or has a satin finish, so that the amount of emitted light is reduced. In this case, 1
The step level is lowered and the binarized image area is remeasured at the next highest threshold value (Hb). When the binarized image area (Sb) at that time is larger than the binarized image area (Sa) by the previous threshold value (Ha), or when the binarized image area substantially matches the corresponding reference area, it is determined as a good product, The threshold value (Hb) at that time is set as the optimum threshold value (Ho). If it gets too big,
The optimum value is a threshold value slightly raised from the threshold value (Hb). And, even if the threshold is lowered, the binarized image area (Sb)
Is constant or 0 from the beginning, it is determined as a defective product. The above operation is repeated a plurality of times, for example, about 5 times in sequence, and if the binarized image area is constant or 0, it is determined as a defective product. In this way, the binarized image area is determined, and the threshold level is appropriately automatically adjusted accordingly to set the optimum binarized threshold value (Ho).

There is also a means for directly setting the optimum threshold value (Ho) by using the centroid calculation of fuzzy inference. For example, as shown in FIG. 4A, the membership functions (Ma) (Mb) of the binarized image area and the threshold are set as the input / output unit. Next, the binarized image area (So) is extracted from the luminance distribution shown in FIG. 3 (a) by a predetermined threshold value, and the fitness (Ta) (Tb) is calculated from the membership function (Ma) of the input section.
Is determined. Then, the conformity (Ta) (Tb) is substituted into the membership function (Mb) of the output part, and as shown in FIG. 4 (b), it is optimized by calculating the center of gravity of the corresponding combined trapezoidal area (hatched part). Calculate the threshold value (Ho).

When the optimum binarization threshold value (Ho) is set, next, as shown in FIG. 4 (c), the threshold value (Ho) is set.
A binarized image (Do) is taken out from the image picked up by the camera (14) based on, and the feature amount of the image (Do), for example, the area or the perimeter is measured. Then, as shown in FIG. 4D, a fuzzy reasoning membership function (Mc) is set in advance for each different hollow path, and, for example, the measured area data is substituted into the membership function (Mc). Therefore,
The goodness of fit (α) of measured data to normal data is derived from the membership function (Mc) of fuzzy reasoning. Therefore, the matching degree (α) is compared and discriminated with the reference value (A), and α
When> A, the product is judged to be non-defective and inspected for any abnormality in the hollow path.

The membership function (Mc) measures the area and perimeter of a plurality of non-defective or defective workpieces, and (N
Probability distribution is drawn with a) as the normal region and (Nb) and (Nc) as the abnormal regions, and it is created for each hollow path and for each feature quantity such as area and perimeter. For example, in the area distribution, if the measured area is (Pa), the goodness of fit for normal data is (Qa), and from that, (Qa) and (A)
And the presence / absence of abnormality in the hollow path is determined based on the magnitude. The inside of the area (Nb) indicates an abnormal area caused by a decrease in the light amount due to the exit side of the path or the core, and the inside of the area (Nc) indicates an abnormality caused by an increase in the light quantity due to a closed casting burr at the middle of the path or the entrance. Indicates the area.

Alternatively, in the case of a plurality of route exits, the plurality of exits are imaged at once and a binarized image including a plurality of each exit image in one image is taken out. Then, based on the image data, the presence / absence of an abnormality in the hollow path may be inspected by fuzzy reasoning using, for example, the total area of a plurality of images or the ratio of each image to the total area. Further, when a plurality of exits are imaged at once, the suitability of each of the plurality of images for each exit may be combined and, for example, the presence or absence of abnormality may be determined from the multiplication value. Further, in addition to the area and the perimeter, the position of the center of gravity of the image is added as a discriminant, and each fitness is multiplied.
It is possible to improve the inspection accuracy for a defective boundary.

As shown in FIG. 1 (b) and FIG. 5 (c), when the hollow path (4a) has a bent labyrinth shape, an abnormality in the mesh (4c) is caused only by projecting light from one direction. May not be reflected in the captured image at all, so it is not possible to sufficiently determine whether or not there is an abnormality in the route (4a). Therefore, for example, a plane mirror (11)
Or a polygonal mirror in which a plurality of plane mirrors having different projection angles are joined and integrated is used instead of the plane mirror (11) to project light from a plurality of different angles. Therefore, first, light is projected from one direction toward one side surface (4b) of the hollow path (4a), and the light passing through the path (4a) and the mesh (4c) thereof is imaged. Similarly, the hollow path (4a) from the other direction
The light is projected toward the other side surface (4d) of the inside, and the light passing through the path (4a) and the mesh (4c) thereof is imaged. Then, two different picked-up images are obtained by projecting light on two different side surfaces (4b) and (4d) of the hollow path (4a), and a difference in brightness area value or the like occurs depending on each image. . Therefore, the respective captured images are combined and the characteristics such as the brightness area value are compared and determined to determine whether or not there is an abnormality in the path (4a).

Further, the positions of the light emitting and receiving units (8) and (9) are reversed to shift the respective positions, and two different picked-up images are obtained.
The presence / absence of abnormality in a) may be determined.

In addition to the head cylinder as the object to be inspected,
The cylinder block can also be inspected for internal abnormality by the device of the present invention, and the invention can be similarly applied to other inspected objects having other hollow paths.

[0021]

According to the present invention, when inspecting an internal abnormality of an object to be inspected having a bent hollow path such as a head cylinder having a cooling water path of an engine combustion chamber, light is projected from the hollow path inlet to make a path. Light that diffuses and passes through the interior is received at the exit of the path on the two-dimensional plane by the imaging means, and the characteristics of the captured image are identified to determine whether there is an abnormality inside the hollow path. Since the received light can be accurately received, and the characteristics of diffusely reflected light such as brightness are identified and discriminated, not only complete occlusion but also half-clogging can be detected, and the inspection accuracy is greatly improved. In addition, at least two diffused lights whose light projecting directions or light projecting positions are shifted are projected at different times, and passing light is two-dimensionally received at each exit of the path at the exit of the path, and the image is captured. Even if the inside of the route bends in a labyrinth and is complicated and complicated, it is possible to inspect for an abnormality inside.

[Brief description of the drawings]

FIG. 1A is a partial vertical cross-sectional view showing a head cylinder and an inspection means for carrying out an internal inspection method for a hollow path according to the present invention. FIG. 1B is a partial horizontal cross-sectional view showing the upper surface of the head cylinder of FIG. 1A and an inspection means for carrying out the internal inspection method of the inspection object according to the present invention.

FIG. 2 is a side view of the inspection means according to the present invention.

FIG. 3A is a graph showing an example of a luminance distribution of a captured image in the method for inspecting the inside of a hollow path according to the present invention, and several kinds of binarized threshold values. (B) is another example of the brightness distribution of the captured image in the method for inspecting the inside of the hollow path according to the present invention and a graph of several kinds of binarized threshold values. (C)
FIG. 4 is a diagram showing an example of an image binarized with the threshold values shown in FIGS. FIG. 3D is a diagram showing a case where an image could not be extracted with the threshold values of FIGS. 3A and 3B.

FIG. 4A is a membership function of an input / output unit for threshold setting by fuzzy inference. (B) is a membership function showing an example of a centroid calculation of fuzzy inference.
(C) is a figure which shows an example of a binarized image. (D) is a waveform diagram showing an example of a fuzzy determination membership function.

FIG. 5A is a bottom view showing an example of a head cylinder. (B) is a partial vertical sectional view showing an example of a conventional internal inspection method for a head cylinder and a conventional device under test. (C)
FIG. 6 is a partial horizontal sectional view of the upper surface side of the head cylinder.

[Explanation of symbols]

 1 Object to be inspected (head cylinder) 3 Hollow path entrance / exit (water hole) 3a Hollow path 4 Hollow path entrance / exit (sand removal hole) 4a Hollow path 7 Inspection means 8 Projection unit 9 Imaging unit

Claims (2)

[Claims] 1. When inspecting the inside of an object to be inspected having a bent hollow path for abnormalities, diffused light is projected from the entrance of the path, and light passing diffusedly reflected in the path is passed by a predetermined imaging means. A method for inspecting the inside of a hollow path, which comprises: receiving light on a two-dimensional plane at the exit of the path and identifying the characteristics of the captured image to determine whether there is an internal abnormality in the hollow path. 2. At least two diffused lights whose light projecting directions or light projecting positions are shifted are projected at the entrance of the curved labyrinth-like hollow path at different times, and the light passing through each diffused light is output at the exit of the hollow path. 2. The internal inspection of the hollow path according to claim 1, wherein the image pickup means receives light on a two-dimensional plane to pick up an image, and whether or not there is an abnormality in the bent labyrinth-like hollow path is determined from the difference in the characteristics of the picked-up images. Method.