Part recognition system

Abstract

This application relates to a part identification system, comprising: a support platform (100) having a preset area (R) for placing a part to be tested; at least one imaging device (200) positioned in association with the preset area (R) and configured to capture view images of the part to be tested placed on the preset area (R) at a corresponding angle; and a processor (300) coupled to the at least one imaging device to receive the view images captured thereon, the processor having pre-stored an image matching program for comparing the received view images with view images of the corresponding angle of a desired part and generating a conclusion that the part to be tested is a desired part when all received view images match the corresponding view images of the desired part.

Description

Technical Field

[0001] This application relates to a part identification system for identifying or determining whether a tested part matches or is identical to a specific desired part. This application is particularly advantageous in identifying micro-parts that are small in size and difficult to distinguish with the naked eye. A typical example of such a micro-part is a vacuum pick-up tool commonly used in the semiconductor packaging field. Background Technology

[0002] In the semiconductor packaging industry, vacuum nozzles are typically used to pick up chips from wafers or trays. Different types or sizes of chips are usually equipped with various types or sizes of nozzles. Due to the small size of chips and the extremely small size of nozzles, they appear visually similar. Selecting the right nozzle from among many to pick up a specific type or size chip is very difficult. In typical chip packaging plants, the traditional operation of selecting the appropriate nozzle for a specific chip usually relies on visual identification or manually comparing physical nozzles with pictures. This manual identification or comparison process by technicians is not only time-consuming but also unreliable. Furthermore, this manual or visual comparison process is prone to errors in identification or selection: if the wrong nozzle type or size is selected, or the selected nozzle does not match the chip to be picked up, quality problems will occur. For example, the nozzle may collide with the chip, causing damage; or the nozzle may not hold the chip firmly, causing the chip to fall.

[0003] Hopefully, the above-mentioned technical problems can be resolved. Summary of the Invention

[0004] The purpose of this application is to solve the aforementioned technical problems.

[0005] This application provides a part identification system comprising: a support platform having a preset area for placing a part to be tested; at least one imaging device positioned in association with the preset area and configured to capture view images of the part to be tested placed on the preset area at a corresponding angle; and a processor coupled to the at least one imaging device to receive the captured view images, the processor having pre-stored an image matching program for comparing the received view images with view images of the corresponding angle of a desired part and generating a conclusion that the part to be tested is the desired part when all received view images match the corresponding view images of the desired part.

[0006] Optionally, the part identification system includes at least one of the following: A memory integrated with or accessible by the processor, the memory pre-storing: the image matching program; a list of information on multiple candidate parts, including the desired part, wherein the information on each candidate part includes a view image corresponding one-to-one with a view image captured by the at least one imaging device, and the shape and size of key features in each view image; An interactive interface integrated with or coupled to the processor, configured to allow an operator to select the desired part from the alternative parts or input the information of the desired part, and / or allow an operator to start and / or end the identification process for the part under test; An output device integrated with or coupled to the processor, for outputting the conclusion; A detection device integrated with or coupled to the processor is used to detect the presence of a part under test in the preset area. The detection device is connected to the processor to notify the processor when it detects that the part under test is placed in the preset area.

[0007] Optionally, the at least one imaging device includes a first imaging device positioned relative to the preset area to directly receive light from the part under test placed in the preset area to directly image the part under test.

[0008] Optionally, the first imaging device includes: an upper imaging device positioned above the preset area of ​​the support platform for capturing a top view image of the part to be tested placed in the preset area; and a lower imaging device positioned below the support platform for capturing a bottom view image of the part to be tested.

[0009] Optionally, the at least one imaging device includes a second imaging device, and the part identification system includes an optical auxiliary device for the second imaging device, the optical auxiliary device being positioned relative to both the preset area and the second imaging device such that light rays from the part under test corresponding to an angle of the second imaging device are guided to the second imaging device via the optical auxiliary device.

[0010] Optionally, the second imaging device includes: a front imaging device and a right imaging device positioned on the upper side of the support platform and respectively used to capture a front view image and a right view image of the part under test. Optionally, the second imaging device includes: at least one of a front imaging device and a rear imaging device positioned on the upper side of the support platform (100) and respectively used to capture a front view image and a rear view image of the part under test; and / or at least one of a left imaging device and a right imaging device positioned on the upper side of the support platform (100) and respectively used to capture a left view image and a right view image of the part under test.

[0011] Optionally, the optical auxiliary device is a reflector.

[0012] Optionally, the part identification system includes at least one light source positioned relative to the preset area to illuminate the part being tested.

[0013] Optionally, the support platform is transparent at least within the preset area.

[0014] Optionally, the part under test is a vacuum nozzle used to pick up chips.

[0015] The automatic part identification system of this application is configured as follows: a support platform having a preset area for placing the part to be tested, at least one imaging device positioned in association with the preset area (and thus with the part to be tested to be placed in the area), and a processor connected to each imaging device to receive a view image captured by it reflecting a specific angle of the part to be tested, the processor having a pre-stored image matching program for comparing the received view image with a view image of the desired part at the specific angle and concluding that the part to be tested is the desired part when the comparison results of all received view images are consistent.

[0016] Using the part identification system of this application, each imaging device is positioned in association with a preset area or its relative position to the preset area is pre-calibrated. As long as the part to be tested is placed in the preset area of ​​the support platform of this system, each imaging device can capture a view image of the part at a specific angle. These images are automatically transmitted to the processor. The processor uses a pre-stored image processing (or comparison or matching) program known in the art to compare the received images one by one with the view images of the expected part at the corresponding angle. When all view images match the corresponding view images of the expected part, the processor concludes that the part to be tested is the expected part (or is not the expected part).

[0017] Therefore, after the part to be tested is placed in the preset area on the support platform, the entire identification process is fully automated, and a conclusion can be reached within seconds. Compared to manual identification, this process is fast, accurate, requires no human intervention, and will not yield different conclusions due to different operators. Using this method, the accuracy of the identification process can reach 100%. For vacuum nozzles used to handle expensive chips, this effectively eliminates chip loss accidents and the resulting waste of resources caused by incorrect nozzle selection. Attached Figure Description

[0018] Figure 1 This is an exemplary structure of a vacuum nozzle used as the part being tested.

[0019] Figure 2 This is an exemplary configuration of the part identification system of this application. Detailed Implementation

[0020] This application relates to a part identification system for identifying or determining whether a part under test is a desired part intended to be selected by an operator. This application will describe the part under test as a vacuum nozzle used in the semiconductor packaging field for picking up chips. Figure 1 A vacuum nozzle S is schematically shown. Figure 2 This is an exemplary configuration for a part identification system used in this vacuum nozzle S.

[0021] refer to Figure 1 The vacuum nozzle S may include a base 10 and a nozzle 20 extending longitudinally from the base 10, wherein the base 10 and the nozzle 20 define opposing base surfaces 15 and end surfaces 25, respectively. A vacuum channel 35 capable of connecting to an external vacuum source extends within the nozzle S (e.g., penetrating the nozzle S longitudinally) and opens to the end surface 25. During operation, the vacuum nozzle S can be positioned relative to... Figure 1 In the opposite (inverted) orientation, an external vacuum source draws a vacuum within the vacuum channel 35, causing the chip (not shown) to be adsorbed onto the end surface 25 for operations such as transfer. In an orthogonal coordinate system with the longitudinal extension direction as the Z-axis, the vacuum nozzle S of this structure is symmetrical with respect to the X-axis and Y-axis perpendicular to the Z-axis.

[0022] Therefore, determining whether the tested nozzle S matches or is identical to the desired nozzle can be done by comparing the following four view images of the tested nozzle and the desired nozzle: two view images on opposite sides of the Z-axis, a view image on one side of the X-axis, and a view image on one side of the Y-axis. If the tested nozzle S is oriented with its base surface 15 facing downwards (and... Figure 1If the images are placed in the same way, the four images to be compared can be: a top view (top view) and a bottom view (bottom view) on opposite sides of the Z-axis (coordinate with the vertical direction); and on one side of the X-axis (coordinate with the first horizontal direction) (e.g., ...). Figure 2 A front (side) view image; and one side (e.g., on the Y-axis aligned with the second horizontal direction) of the front (side) view image; Figure 2 The right (side) view image.

[0023] Correspondingly, in Figure 2 The part identification system shown includes a support platform 100, four imaging devices, and a processor 300. The four imaging devices include: an imaging device 200a for capturing a top view image of the nozzle S; an imaging device 200b for capturing a bottom view image of the nozzle S; and imaging devices 200c and 200d for capturing a front view image and a right view image of the nozzle S.

[0024] The support platform 100 is used to place or support the part being tested, i.e. Figure 1 The test nozzle S has a preset area R for placing the test nozzle S. All imaging devices are positioned in association with or pre-calibrated relative to the preset area R of the support stage 100 to ensure that they can capture view images of the test nozzle S at the corresponding angle (or side). Specifically, as illustrated, imaging device 200a is located above the preset area R of the test nozzle S, particularly above the test nozzle S placed thereon. Imaging device 200b is located below the support stage 100, at least the preset area R of the support stage 100 is formed of a transparent material, so that imaging device 200b can image the test nozzle S in the preset area R of the support stage 100 from below the support stage 100. Imaging devices 200c and 200d are also disposed above the support stage 100. To ensure that imaging devices 200c and 200d receive light from the front and right sides of the nozzle under test S, this system is equipped with two optical auxiliary devices 400a and 400b. These devices are positioned or relatively calibrated in association with the corresponding imaging devices 200c and 200d and with a preset area R of the support stage 100, respectively, to ensure that light from the corresponding front and right sides of the nozzle under test S can be guided to the corresponding imaging devices 200c and 200d. The figure also shows auxiliary light sources 500a and 500b, respectively positioned above and below the support stage 100, for emitting light towards the nozzle under test S to enhance the light brightness at the preset area R and the nozzle under test S, thereby improving the image clarity of the imaging devices. Directional terms used herein refer to… Figure 2Understanding is key. Specifically, the terms "upper" and "lower" correspond to two opposite directions extending along the Z-axis, with the nozzle S being tested positioned above the support platform 100 (indicated by the arrow on the Z-axis); the terms "front" and "back" correspond to two opposite directions extending along the X-axis, with the optical auxiliary device marked 400a located in front of the nozzle S being tested (indicated by the arrow on the X-axis); and the directional terms "left" and "right" correspond to two opposite directions extending along the Y-axis, with the optical auxiliary device marked 400b located to the right of the nozzle S being tested (indicated by the arrow on the Y-axis). The understanding of these directional terms changes depending on the angle of the view.

[0025] Processor 300 is coupled to all imaging devices to receive images from each imaging device. Processor 300 may pre-store image processing (or matching or comparison) programs known in the art, configured to perform an operation on each image received from each imaging device that compares the image with an image at the same angle as the desired nozzle, and concludes that the nozzle under test on the current support stage 100 is the desired nozzle if the comparison results for each received image are consistent (or matched or identical). Examples of known image comparison, matching, or processing methods that can be used in this application may include, but are not limited to: pixel-level comparison methods, perceptual hashing algorithms, feature point matching methods, deep learning algorithms utilizing trained neural networks, etc. This application employs image comparison methods known in the art, and this application does not involve improvements to the image matching program. In this document, the terms "comparison," "matching," and "processing" have the same meaning. The terms "consistent," "fitted," "matched," and "identical" for two images also have the same meaning.

[0026] In the illustrated example, processor 300 employs a pre-stored image matching method known in the art. First, it compares the top and bottom view images from the upper and lower imaging devices 200a and 200b with the corresponding top and bottom view images of the desired suction nozzle. Then, it compares the front and right view images from the front and right imaging devices 200c and 200d with the corresponding front and right view images of the desired suction nozzle. These comparison operations are performed sequentially, and the next image comparison step is only performed if the comparison result of the previous image comparison step is a match or identical between the two images. The comparison between a view image of the suction nozzle under test and its corresponding view image of the desired suction nozzle can be understood as including: first, identifying the shape and size of key features in the view image of the suction nozzle under test, and then comparing the shape and size of the identified key features in the view image with the shape and size of corresponding key features in the corresponding view image of the desired suction nozzle. For any key feature in any view image, the shape can be compared first, followed by the size. Finally, if all the comparison results of the images are the same or match, the processor 300 can conclude that the selected nozzle S is the desired nozzle.

[0027] In some examples, if the comparison results of any view image are different or mismatched, the processor 300 can directly conclude that the selected nozzle S1 is not the desired nozzle, or it can be set to perform an optional operation. This optional operation can be set to restart all comparison operations in response to operator input. For example, in this optional operation, the processor 300 provides a prompt to the operator, informing them of the specific view image where the current comparison result is a mismatch. The operator may realize that the placement or orientation of the selected nozzle S1 is incorrect, and then choose to invert or otherwise rotate the selected nozzle S1 placed on the support platform 100 (e.g., 90 degrees or other angles) and re-execute the entire identification process, i.e., restart from comparing the top view image. As another example, in the optional operation, the operator can choose to replace the placed nozzle S1 and re-execute the identification process. For example, the operator can also input an instruction to terminate the identification process, or the processor 300 can be set to conclude that the tested part is not the desired part and directly end the identification process if no instruction is received from the operator after a preset time period during this optional operation.

[0028] In one implementation, for Figure 1The processor 300 can be configured such that: when an optional operation occurs during the comparison operation of the top view and bottom view images (i.e., the comparison result of either of the two view images is mismatched), the operator selects the selected nozzle S1 placed upside down on the support table 100 and re-executes the identification process; when an optional operation occurs during the comparison operation of the front and right view images, the operator rotates the selected nozzle S1 on the support table 100 90 degrees clockwise or counterclockwise and re-executes the identification process.

[0029] The processor 300, capable of performing image comparison functions and drawing conclusions, can have any configuration known in the art, for example, it can be embodied in software, hardware, or a combination thereof. One feasible exemplary configuration of the processor 300 is an embedded computer with software installed therein capable of performing the methods described above; this application does not involve improvements to the software. The processor 300 can be configured in any way to store executable instructions that, when executed, can run the comparison methods described above or perform the comparison functions described in this application. Another feasible exemplary configuration of the processor 300 is a portable device such as a mobile phone.

[0030] This identification system may also include a memory integrated with or coupled to the processor 300 so that it can be accessed by the processor 300. The memory may store the aforementioned executable instructions, software, or programs that can be executed by the processor 300 to implement the identification process of this application. The memory may store information about one or more alternative suction nozzles, the information of each nozzle may include: a view image at at least one angle (e.g., four different angles of the part being measured reflected in the four view images described above), the shape and size of key features of each view image, etc. This information may be stored in the form of images, text, symbols, or combinations thereof.

[0031] This identification system may also include an interactive device integrated with or coupled to the processor 300. The interactive device is configured to allow the operator to select one from a pre-stored list of candidate nozzles as the desired nozzle to match the chip to be picked up, retrieving all information of the desired nozzle as standard information for the comparison operation to be performed. Optionally, the interactive device may also be configured to allow the operator to input various information about the desired nozzle. The interactive device may also be equipped with "start" and / or "end" controls for manual operation by the operator to begin the identification process after the nozzle under test is placed on the support platform 100, and / or to allow the operator to manually end the identification process.

[0032] This identification system may also include any known output device in the art, integrated with or coupled to the processor 300. The output device is used to output to the operator a conclusion regarding whether the tested nozzle placed on the support platform 100 matches the desired nozzle. The output device can be any known output device in the art, capable of notifying the operator of the above conclusion information in any form, such as visual, auditory, or visual. The output device can be integrated with the interactive device or provided separately. The output device and the interactive device can be the same display.

[0033] This identification system may also include any known detection device in the art integrated with or coupled to the processor 300, for detecting the presence of the nozzle to be tested on a preset area R of the support stage 100 and notifying the processor 300 to automatically initiate the identification process of this application. The detection device can be any known detector or sensor in the art. Including this detection device enables a more complete automated identification process.

[0034] As described above, using the identification system of this application, after selecting or inputting the information of the desired nozzle, the operator only needs to place the selected or tested nozzle S on the support platform 100. The system's processor can automatically, quickly, accurately and reliably identify whether the tested nozzle matches or is the same as the desired nozzle by using its internally stored image matching or comparison program and give the corresponding conclusion.

[0035] In the exemplary embodiment shown in the accompanying drawings, when the nozzle S under test is placed in the preset area R of the support stage 100 such that its base surface 15 (or end surface 25) contacts (or moves away from) the support stage 100, the nozzle S under test is symmetrical about the X-axis and Y-axis. Therefore, the identification system of this application can only be equipped with two imaging devices for capturing front view and right view images. Those skilled in the art should understand that if the part under test is asymmetrical about the X-axis and Y-axis, the identification system may also include additional imaging devices for capturing rear view and / or left view images; if the part under test has a uniform cross-section along the longitudinal extension direction, the identification system may only be equipped with the upper imaging device 200a or the lower imaging device 200b, in which case the preset area R of the support stage 100 does not necessarily have to be made of transparent material.

[0036] Depending on the specific structure of the part under test and the key features of interest, the discrimination system may, as an addition or alternative, include imaging devices for imaging the part under test from any other angle. Correspondingly, it is desirable that the part have view images corresponding to the angles of each view image captured by this discrimination system for comparison.

[0037] As described above, this identification system compares one or more view images of the tested part from one or more angles with view images of the desired part from corresponding angles. The view images of the desired part can be input by the operator via, for example, the aforementioned interactive device. In some preferred embodiments, a list of numerous candidate parts can be pre-stored (in a memory such as those described above). The list contains view images of each candidate part, and the stored view images correspond to the configuration of the imaging device of the part identification system, specifically, they strictly correspond one-to-one with the angle at which the imaging device images the tested part. Before performing the identification process of the tested part, the operator first (e.g., via the aforementioned interactive device) selects the desired part from the pre-stored list of candidate parts, i.e., automatically retrieves the view images of the desired part, so that the processor 300 can compare them with the view images captured by the imaging device of this identification system.

[0038] Furthermore, this application does not limit the specific structure of the support platform 100 of the system. In particular, in embodiments requiring comparison of bottom-view images, i.e., embodiments requiring imaging of the lower side of the measured part, at least a predetermined area R of the support platform 100 is made of a transparent material. For example, in Figure 2 In this application, the support platform 100 can be configured to be made entirely of transparent glass and designed to be supported by a frame K. This application does not limit any details of the frame K.

[0039] Although optical auxiliary devices 400a and 400b are configured only for imaging devices 200c and 200d used to capture front view and right view images in the illustrated embodiments, optical auxiliary devices can be configured for any imaging device as needed. The optical auxiliary devices are not limited to reflective optics, nor are they limited to the illustrated plane mirror. Similarly, this application does not limit the details of the imaging devices and light sources, provided that the functions described herein can be achieved.

[0040] Furthermore, this application does not limit the coupling method between the processor 300 and the imaging device 200, such as wireless or wired connection, as long as the processor 300 can obtain images from the imaging device 200.

[0041] As mentioned above, this application only uses existing image comparison or matching programs or methods to identify part images, and does not involve any improvement to the image comparison method. By linking a support stage, at least one imaging device, and a processor pre-stored with an image matching program, and by calibrating or correlating the relative positions of a preset area of ​​the support stage and the imaging device, a fast, automatic, and accurate identification of whether the part under test is the desired part is achieved, eliminating the defects of human identification that may lead to misjudgment and low efficiency. The conclusion of this identification process is objective and accurate, and will not differ due to different operators. This system is particularly advantageous when the part under test is small and difficult to distinguish with the naked eye; furthermore, the advantages of this part identification system are even more obvious when incorrect identification or selection of the part under test would lead to significant waste of resources (e.g., incorrect selection of a vacuum chuck would lead to damage or waste of expensive chips), and when frequent part identification operations are required (e.g., in a chip manufacturing workshop).

[0042] Although the illustrated embodiment is described using a vacuum chuck as an example, those skilled in the art should understand that the part identification system of this application is not only applicable to the identification of vacuum chucks, but also applicable to any similar application. This identification system does not necessarily require four imaging devices, nor does it necessarily require the configuration of optical auxiliary devices and auxiliary light sources. The configuration details and arrangement of the imaging devices, optical auxiliary devices, and auxiliary light sources are not limited. All these details can be modified according to actual conditions, provided that it is possible to identify whether the part being tested is the desired part.

[0043] Although only some possible variations have been described above, those skilled in the art will recognize that changes in form and detail may be made without departing from the spirit and scope of the invention.

Claims

1. A parts identification system, characterized in that, include: A support platform (100) has a preset area (R) for placing the part to be measured. At least one imaging device (200) is positioned and configured in association with the preset area (R) to capture a view image of a part under test placed on the preset area (R) at a corresponding angle; and A processor (300) coupled to the at least one imaging device to receive view images captured therefrom, the processor having pre-stored an image matching program for comparing the received view images with view images of the corresponding angles of the desired part and generating a conclusion that the tested part is the desired part when all the received view images match the corresponding view images of the desired part.

2. The parts identification system according to claim 1, characterized in that, The part identification system includes at least one of the following: A memory integrated with or accessible by the processor, the memory pre-storing the image matching program; A list of information for multiple candidate parts, including the desired part, wherein the information for each candidate part includes a view image corresponding one-to-one with a view image captured by the at least one imaging device, and the shape and size of key features in each view image; An interactive interface integrated with or coupled to the processor, configured to allow an operator to select the desired part from the alternative parts or input the information of the desired part, and / or allow an operator to start and / or end the identification process for the part under test; An output device integrated with or coupled to the processor, for outputting the conclusion; A detection device integrated with or coupled to the processor is used to detect the presence of a part under test in the preset area (R). The detection device is connected to the processor to notify the processor (300) when it detects that the part under test is placed in the preset area (R).

3. The parts identification system according to claim 2, characterized in that, The at least one imaging device includes a first imaging device positioned relative to the preset area (R) to directly receive light from the part under test placed in the preset area (R) to directly image the part under test.

4. The parts identification system according to claim 3, characterized in that, The first imaging device includes: an upper imaging device positioned above the preset area (R) of the support platform (100) for capturing a top view image of the part to be tested placed in the preset area (R); and a lower imaging device positioned below the support platform (100) for capturing a bottom view image of the part to be tested.

5. The parts identification system according to claim 3, characterized in that, The at least one imaging device includes a second imaging device, and the part identification system includes an optical auxiliary device for the second imaging device, the optical auxiliary device being positioned relative to both the preset area (R) and the second imaging device such that light rays from the part under test corresponding to an angle of the second imaging device are guided to the second imaging device via the optical auxiliary device.

6. The parts identification system according to claim 5, characterized in that, The second imaging device includes: at least one of a front imaging device and a rear imaging device positioned on the upper side of the support platform (100) and used to capture front view images and rear view images of the part under test, respectively; and / or at least one of a left imaging device and a right imaging device positioned on the upper side of the support platform (100) and used to capture left view images and right view images of the part under test, respectively.

7. The parts identification system according to claim 5, characterized in that, The optical auxiliary device is a reflector.

8. The parts identification system according to claim 1, characterized in that, The part identification system includes at least one light source positioned relative to the preset area (R) to illuminate the part being tested.

9. The part identification system according to any one of claims 1-8, characterized in that, The support platform (100) is transparent at least within the preset area (R).

10. The part identification system according to any one of claims 1-8, characterized in that, The part under test is a vacuum nozzle (S) used to pick up chips.

Images