Method and measuring device for checking at least one component made of an at least partially transparent material

Abstract

The invention relates to a method and to a device for checking a component (10) made of an at least partially transparent material for defects (11) present in or on the component, in which method the component (10) is transilluminated from one side by means of light emitted from an IR light source (1) in such a way that the light exiting the component (10) on the opposite side is recorded in the form of an image by an IR camera (2), the image information from which is analyzed by an electronic analysis unit (4) via image processing with respect to defects (11) which differ from the immediate image environment with respect to shape, size, and / or degree of transparency.

Description

[0001] Description

[0002] Title:

[0003] Method and measuring device for testing at least one component made of at least partially transparent material

[0004] The present invention relates to a method and a measuring device for testing at least one component made of an at least partially transparent material, preferably plastic material, for transparency-reducing defects present therein or on it.

[0005] The application of the invention extends to quality control during the manufacturing of components made of transparent or at least partially transparent materials, such as plastics or the like. It is usually necessary to inspect the components to be processed before a manufacturing step with regard to material composition and properties, particularly to detect, for example, foreign inclusions that would impair light transmission. Furthermore, quality control is also desirable during a manufacturing step, such as laser welding several such components together or a different type of component to a partially transparent component of interest here, in order to check, for example, the weld quality.

[0006] State of the art

[0007] German patent DE 36 20 129 A1 discloses a generic device for the optical inspection of components made of a transparent material for surface defects and inclusions. The component is scanned by means of a moving light beam to create a cross-sectional image of the component. This cross-sectional image is observed by a device that is inclined to the direction of incidence of the light beam. This device contains an imaging optical system in whose image plane an interchangeable mask is arranged for separating the image in a plane of the test piece. The light beams passing through this mask strike a receiver, to which an arrangement for evaluating the receiver signals is connected. Using such a testing device, surface defects of the component under test can be detected separately for the front and back of the component.

[0008] DE 100 04 049 A1 describes a method for testing non-transparent plastic objects. A weld seam between plastic parts is tested using thermography. This is done by capturing the thermal radiation of the plastic parts in the weld seam area with at least one camera, for example, a digital infrared camera, and comparing it to at least one reference value using a computer.

[0009] In contrast, the object of the present invention is to provide a method and a measuring device for testing at least one component made of at least partially transparent material, which enables reliable quality testing before and during a manufacturing step, in particular laser welding, by assessing transparency parameters.

[0010] Disclosure of the invention

[0011] The problem is solved process-technically by claim 1. Dependent claim 7 specifies a measuring device for carrying out the method according to the invention. The dependent claims are directed to advantageous embodiments of the invention.

[0012] The invention includes the process teaching that, in order to test at least one component made of at least partially transparent material for defects present therein or on it, the component is illuminated from one side with an IR light source in order to record the light emerging from the component on the opposite side as an image by an IR camera, the image information of which is analyzed by a downstream evaluation unit by means of image processing with regard to defects which differ from the immediate image environment by a lower transparency in terms of shape, size and / or degree of transparency.

[0013] The image environment is understood as the immediate surroundings of a coherent structure of relatively lower transparency than its surroundings. This environment exhibits a homogeneous transparency corresponding to that of the material and is detectable by this transparency. Differences in light intensity within the image information are used to distinguish image areas with respect to transparency.

[0014] The image processing algorithm detects anomalies in the image information, which are interpreted as defects or missing elements, in a manner that is generally known. This is usually achieved through a combination of different techniques. The main methods, which are at least partially combined, are presented below:

[0015] Image preprocessing can include noise reduction using filters such as median or Gaussian filters to reduce image noise that could interfere with anomaly detection. Normalization adjusts brightness and contrast to ensure uniform illumination and improved comparability. Segmentation uses thresholding to separate objects from the background by setting binary thresholds. Region growth merges neighboring pixels with similar properties. Edge detection of the component can be performed using algorithms such as Sobel, Canny, or Prewitt, allowing for differentiation from defects. Feature extraction analyzes the component's shape features, such as contours and geometric forms.In this context, it is also possible to identify texture features by analyzing the texture using methods such as the grayscale coexistence matrix (GLCM) or local binary patterns (LBP) to distinguish them from defects. Furthermore, comparison with reference models is also conceivable, such as template matching, in which the image is compared with predefined patterns or templates. By subtracting the reference image from the test image, differences can be identified that may represent a potential defect. Within the framework of machine learning and artificial intelligence, supervised learning can also be applied by training models with labeled examples of normal and defective images. In addition, unsupervised learning is also conceivable, where algorithms such as k-means or autoencoders are used, which detect anomalies without labeled data.Deep learning enables the use of convolutional neural networks (CNNs) for the automatic detection and classification of defects. Statistical methods for anomaly detection can be used to identify statistical deviations in image data. The evaluation of such image analyses can be enhanced through morphological operations, particularly techniques like erosion, dilation, opening, and closing, to refine the segmentation results. The extracted features can then be evaluated in the context of the expected defects or anomalies.

[0016] These methods and techniques can be combined, depending on the application and type of defects to be detected, in order to maximize detection accuracy.

[0017] In the context of this application, a defective image area is understood to be a region of the illuminated image which, taking into account the contour profile determined by the component geometry, has a relatively lower transparency than its surroundings. Since the transparent component exhibits largely homogeneous transparency, such image areas appear as darker spots or streaks against the surroundings. If the image information is inverted, these areas appear brighter than the surroundings. This is because the IR rays cannot penetrate these image areas, or cannot do so completely. If defects consisting of a non-transparent foreign material are embedded in the component, they stand out with strong contrast against the surroundings.If, on the other hand, the clouding of material areas is due to, for example, a mechanically damaged surface, such defects usually stand out less contrastingly from the surroundings.

[0018] Based on this image information, which can be analyzed and categorized using image processing technology, as well as additional geometric information about the defects, the electronic evaluation unit can assign them to a specific defect type based on shape, size, and / or transparency within the framework of the image analysis described above. This includes material inclusions, incorrect material type, defective laser welding, and the like. For example, a material inclusion can be identified based on the characteristic size and / or shape of the defect. An incorrect material type leads to a homogeneously different transparency of the component than expected. Defective laser welding is evident, for example, through jagged weld edges. This offers the advantage that a decision can be automatically made regarding necessary rework or the classification of the inspected component as a reject part.

[0019] Preferably, the wavelength of the light emitted by the IR light source corresponds to the wavelength of a laser weld on the component. In the case where laser welding is performed on the component, this allows for quality control to be carried out before the welding process, since the measuring device, due to the matching wavelengths, is suitable both for checking the material composition and properties of at least one component before the laser welding step and simultaneously for checking the weld quality during the laser welding process, without requiring two separate testing devices.

[0020] For this purpose, the wavelength of the IR light source is preferably in the range between 900 nm and 1000 nm, and most preferably between 930 nm and 950 nm. The selection of the specific wavelength within the specified range depends on the material of the at least partially transparent component to be welded. For example, a wavelength of approximately 940 nm is chosen for the material polybutylene terephthate (PBT). A suitable measuring device for carrying out the method described above comprises an IR camera, which is preferably arranged above an IR light source, which, according to a preferred embodiment, is designed as a light panel. This offers the advantage that the component to be tested can be placed on it and then illuminated from below.The IR chamber located above can preferably be equipped with an interchangeable band filter to easily adjust the desired wavelength to be recorded.

[0021] Detailed description based on drawing

[0022] Further measures improving the invention are described in more detail below, together with a description of a preferred embodiment of the invention, with reference to the figures. The figures show:

[0023] Fig. 1 shows a schematic perspective side view of a measuring device for testing a component of the type of interest here.

[0024] Fig. 2 shows a schematic flow chart of the test procedure to be carried out here,

[0025] Fig. 3 shows a pictorial representation of image information recorded by the IR camera of a component in a first layer, and

[0026] Fig. 4 shows a pictorial representation of image information recorded by the IR camera of a component in a second layer.

[0027] As shown in Fig. 1, a measuring device comprises an IR light source 1, which is designed as a light table and emits infrared light with a wavelength of 940 nm. A transparent component 10 made of PBT is arranged on this IR light source 1 for inspection for defects. The component 10 is illuminated from below by the IR light source 1. Opposite, i.e., above the IR light source 1 with the component 10 arranged on it, is an IR camera 2 for capturing the emitted image information. The image information consists of components emitted directly by the IR light source 1, namely those areas not covered by the transparent component 10, as well as image areas derived from the light components that have passed through the component. The IR camera 2 is equipped with an IR filter 3 for setting the desired wavelength to be recorded, here 940 nm.The IR camera 2 delivers the captured image information to an electronic evaluation unit 4, which performs an image analysis with regard to defects contained in the transparent component 10.

[0028] The procedure carried out with the test device described above consists, according to Fig. 2, in step I in illuminating the component 10 with an IR light source 1, which is then recorded as an image by an IR camera 2 in step II, in order to subsequently be evaluated for defects in step III by means of an electronic evaluation unit 4.

[0029] According to Fig. 3, an exemplary transparent component 10 has a frame-shaped form and exhibits a defect 11 on one side, which indicates a material inclusion. This defect 11 stands out from the rest of the bright image due to darker, i.e., less transparent, image areas. The image is processed such that the areas that do not pass through the component 10, i.e., those directly captured by the IR camera, are displayed in inverted black.

[0030] In the side view of the frame-shaped component 10 described above, as shown in Fig. 4, it can be seen that the defect 11 extends over the entire width of the frame. This allows the defect to be classified as a defect type which, in this case, indicates a material inclusion.

[0031] The invention is not limited to the preferred embodiment described above. Rather, variations thereof are also conceivable, which are also covered by the scope of protection of the following claims. For example, it is also possible to detect other types of defects, such as surface defects on a transparent component or inclusions, which negatively affect the component properties and therefore allow an assessment as to whether the tested component should be rejected as a scrap part or whether it falls within suitable tolerances for further manufacturing, for example, because a detected material inclusion at a weld edge of the component would be covered by a much wider weld seam.

Claims

Claims 1. Method for inspecting a component (10) made of at least partially transparent material for defects (11) present therein or on it, characterized in that the component (10) is illuminated from one side with light emitted by an IR light source (1) such that the light emerging from the component (10) on the opposite side is recorded as an image by an IR camera (2), the image information of which is analyzed by an electronic evaluation unit (4) by means of image processing with regard to defects (11) which differ from the immediate image environment by a lower transparency with respect to shape, size and / or degree of transparency.

2. Method according to claim 1, characterized in that a range is selected for the wavelength of the light emitted by the IR light source (1) which includes the wavelength of a laser weld on the component (10).

3. Method according to claim 2, characterized in that a range between 900 nm to 1000 nm, preferably a range between 930 and 950 nm, is selected for the wavelength.

4. Method according to claim 1, characterized in that the defect (11) is assigned to a defect type as a result of image processing based on the shape, size and / or degree of transparency, comprising material inclusion, incorrect material type, defective laser welding.

5. Method according to claim 1, characterized in that the inspection for defects (11) is carried out before a subsequent laser welding in order to check the weldability.

6. Method according to claim 5, characterized in that the inspection for defects (11) is additionally carried out during laser welding to check the quality of the laser welding.

7. Measuring device for testing a component (10) made of at least partially transparent material for defects (11) present therein or on it, characterized in that an IR light source (1) is provided for illuminating the component from one side, wherein an IR camera (2) records the light emerging from the component on the opposite side, the image information of which is analyzed by a downstream electronic evaluation unit (4) by means of image processing with regard to defects (11) which differ from the immediate image environment by a relatively lower transparency with regard to shape, size and / or degree of transparency.

8. Measuring device according to claim 7, characterized in that the IR camera (2) is equipped with an IR filter (3) for adjusting the desired wavelength to be recorded.

9. Measuring device according to claim 7, characterized in that the IR light source (1) is designed in the form of a luminous field or light table which provides a support for the component (11) to be tested.

10. Measuring device according to one of the preceding claims, characterized in that the IR camera (2) is arranged above the IR light source (1).

Images