TWI635463B - Method for detecting appearance of articles with continuous image capturing function - Google Patents

Abstract

The invention relates to a method for detecting the appearance of an article having a continuous image capturing function, which mainly moves an object to be tested through an automatic mechanical device for movement, and continuously captures the appearance of the object to be tested by an image processing device. The image is the same as the image of the contiguous area of the two-phase contiguous image; after the image processing device captures the image, the image of the contiguous area of the two-phase contiguous image is compared, and when the determination is correct, the image is The corresponding image in the preset set of sample images in the image processing device is compared and verified. When the determination is correct, the appearance of the object to be tested is normal, and the image is continuously captured to improve the number of the compared images. The situation, thereby achieving the purpose of improving detection efficiency and accuracy.

Description

Appearance detection method for articles with continuous image capture function

The invention relates to an object appearance detecting method, in particular to an object appearance detecting method with a continuous capturing image function.

Today's consumers are not only required to use performance, but also require a distinctive and good-looking appearance. Therefore, the mobile phone industry has spent a lot of effort to strive to design the mobile device as a boutique. Favor.

At present, the design of the mobile device is more focused on the overall shape, button shape and chassis material, color or texture, etc. Among them, the metal casing is supplemented by the metal hairline, which is favored by consumers. Therefore, the mobile phone industry mostly adopts Metal casing; because the metal hairline of the metal casing is more detailed, the casing manufacturer needs to inspect each metal casing after production to check whether the metal hairline is broken or the grain is incorrect. And other abnormal conditions.

In the prior art, an object appearance detecting method is to use an automatic mechanical device such as a mechanical arm to take a metal casing to move, and after each movement of the robot arm, capture an image of a current position of the metal casing for image comparison. However, the current robotic arm is limited by the problem of the accuracy of the movement, so that the moving distance is too large each time, and the area between the front position and the next position of the metal casing cannot capture the image, and the image cannot be imaged. In comparison, the detection is missing. When there is an abnormality in these metal casings, if the shipment to the mobile phone manufacturer may cause commercial compensation problems and cause losses, the prior art has problems to be solved.

In view of the above problems in the prior art, the main object of the present invention is to provide an object appearance detecting method with a continuous image capturing function, which is obtained by taking an object to be tested by an automatic mechanical device and moving through an image processing device. The moving process of the measured item continuously captures the image to improve the detection accuracy, thereby improving the detection efficiency and the detection accuracy.

The technical means for achieving the above object is to perform the method for detecting the appearance of an article having a continuous image capturing function, which is performed by an image processing device electrically connected to an automatic mechanical device. Moving an object to be tested and pre-setting more than one set of position information, the image processing device presets more than one set of sample images, the method comprising the steps of: reading a preset set of position information in the automatic mechanical device, Corresponding to generate a path information, and correspondingly generate a majority of the trigger signals according to the path information; according to the trigger signals, when the robot moves the object to be tested, the appearance of the object to be tested is continuously captured, and most of the appearance images are continuously captured. And the images of the contiguous areas of the two adjacent images are the same; the images of the contiguous areas of the image are compared; the images are compared with the set of sample images to output a comparison information.

According to the above method, when the automatic mechanical device absorbs the movement of the object to be tested, the image processing device continuously captures the appearance image of the object to be tested to provide sufficient appearance images for verification comparison, thereby improving detection efficiency. And the purpose of precision.

Referring to FIG. 1 and FIG. 2, the present invention includes an automatic mechanical device 10, an image processing device 20, and an object to be tested 30. In this embodiment, The item to be tested 30 can be a mobile phone case.

The robot apparatus 10 includes a base 11 and a mechanical suction body 12. The base 11 is provided with a signal processing device 13. The signal processing device 13 is pre-positioned with more than one set of position information, and the set of position information includes an initial position information. And the end position information, the mechanical suction body 12 is disposed on the base 11 and electrically connected to the signal processing device 13, the mechanical suction body 12 is configured to suck the object to be tested 30 for movement and rotation; In the example, the robot 10 is a robot arm.

The image processing device 20 includes an image capture device 21, a display 22, and an image analyzer 23. The image analyzer 23 is electrically connected to the image capture device 21 and the display 22, respectively. The signal processing device 13 of the robot 10 is electrically connected. The image analyzer 23 is pre-configured with a plurality of sample images and is used to process the received image information. In this embodiment, the image capturing device 21 can be a The camera; the image analyzer 23 is a computer.

First, the image analyzer 23 reads the initial position information and the end position information stored in the signal processing device 13, and performs path analysis to generate a path information corresponding to the path information. A plurality of image capture points are generated, and a corresponding trigger signal is generated according to each image capture point, and then the image analyzer 23 sends an activation signal to the signal processing device 13 to cause the mechanical pickup body 12 to absorb the test target. And the object 30 is moved to the corresponding initial position of the image capturing device 21 according to the initial position information, and the image analyzer 23 sends the path information to the signal processing device 13 . The mechanical analyzer 12 drives the object to be tested 30 to move according to the path information. When the object to be tested 30 moves to the corresponding image capturing point, the image analyzer 23 sends a corresponding trigger signal to the object. The image capturing device 21 causes the image capturing device 21 to continuously capture most of the appearance images during the movement of the object to be tested 30 until the mechanical suction The body 12 stops moving the object to be tested 30 to a corresponding end position according to the end position information. In the embodiment, the path information includes a moving distance of the object to be tested 30 from the initial position to the end position. And moving time.

In this embodiment, the manner of generating the image capturing points may include the following two methods. One method is that the moving distances in the path information are equally spaced according to the number of images to be captured and the image resolution. The method is to generate the image capturing points; the other method is to divide the moving time in the path information into the image capturing points in a fixed interval according to the number of images to be captured and the image resolution. The above two methods are merely examples and are not intended to be limiting.

In this embodiment, when the mechanical suction body 12 stops moving, inertial shaking occurs. To prevent the captured image from being unclear, the image capturing device 21 absorbs the object to be tested 30 from the mechanical suction body 12. When the initial position and the end position are respectively, a starting appearance image and an ending appearance image are captured, and the initial appearance image and the ending appearance image respectively include a part of the environment image and a part of the object to be tested. The image of 30 is provided to provide the necessary image capture buffer; therefore, the first image in the continuously captured appearance image is the initial appearance image, and the last image is the end appearance image.

Since the image is captured by the image capturing device 21 while the object to be tested 30 is continuously moved, the image of the connecting region of the two adjacent images is the same, as shown in FIG. For example, the initial appearance image A includes a first connection area A1 and a second connection area A2. The first appearance image B includes a first appearance image B and a second appearance image C. A first connection area B1 and a second connection area B2, the second appearance image C includes a first connection area C1 and a second connection area C2.

The first connection area of the initial appearance image is an image of the environment; and the second connection area of the end appearance image is an image of the environment.

The image analyzer 23 compares the second contiguous area A2 of the initial appearance image A with the first contiguous area B1 of the next adjacent first appearance image B, and when the same, determines the first The appearance image B and the initial appearance image A are consecutive images, and the image is normal. The image analyzer 23 further divides the second connection region B2 of the first appearance image B with the adjacent next second appearance image C. The first contiguous area C1 is compared. When the images of the contiguous areas of the successive appearance images are compared and both are determined to be normal, the first appearance image B and the second appearance image C are associated with the group. Comparing the corresponding images in the sample image, and determining that the first appearance image B and the second appearance image C are normal when the images are the same, that is, the first appearance image B corresponding to the object to be tested 30 is The appearance of the area of the second appearance image C is normal, and the image analyzer 23 compares all the appearance images; in the preferred embodiment, the image analyzer 23 compares the first connection area of the finished appearance image with the adjacent (connected) the previous one A second connecting area of the image for comparison.

In this embodiment, when the image analyzer 23 compares the appearance images and determines that they are normal, the image analyzer 23 correspondingly generates a correct comparison information, and transmits to the display 22 to display the to-be-tested. The appearance of the item 30 is normal.

In this embodiment, when the images of the contiguous regions of the two adjacent images are not the same, the image analyzer 22 determines that the object to be tested 30 is abnormal in appearance, and correspondingly sends a first abnormal comparison information to the display 22 The appearance of the object to be tested 30 is abnormal, and is sent to the signal processing device 13 for processing, so that the mechanical suction body 12 transfers the object to be tested 30 to an abnormal area for detection by the detecting personnel.

In this embodiment, the image analyzer 23 compares the appearance images with the corresponding images in the set of sample images, and when the comparison results of one of the appearance images are different, the image analysis The device 23 determines that the object to be tested 30 is abnormal in appearance, and correspondingly transmits a second abnormality comparison information to the display 22 to display an abnormal appearance of the object to be tested 30, and transmits the same to the signal processing device 13 to enable the mechanical suction body. 12 Transfer the item to be tested 30 to an abnormal area for placement.

As shown in FIG. 4, the application method of the preferred embodiment of the present invention can further summarize an object appearance detecting method, which is mainly connected by an automatic mechanical device 10 and an image processing device 20, the automatic mechanical device. Presetting a set of more than one position information in the image processing device 20, the automaton device 10 picks up an object to be tested 30 for movement, and the image processing device 20 captures the image, and The image processing device 20 performs the following steps: reading a preset set of position information in the automatic mechanical device 10 to generate a path information correspondingly, and generating a majority trigger signal according to the path information (S10); The position information of the group can be set by the user according to the size of the object to be tested 30. According to the trigger signals, when the robot 10 picks up the object to be tested 30, the corresponding item 30 is determined. Appearance, continuously captures most of the appearance image, and the image of the connection area of the continuous appearance image is the same (S20); the connection area of the continuous appearance image The images are compared (S30); the appearance images are compared with the set of sample images to output a comparison information (S40).

Therefore, the image detecting device 20 generates the path information according to the set of position information by the image processing device 20, so that the robot 10 absorbs the object to be tested 30 according to the path information, and therefore, when the step is performed, When a set of position information preset in the robot mechanism 10 is read to correspondingly generate a path information, and a step of generating a plurality of trigger signals (S10) according to the path information, as shown in FIG. 5, the following is further included. Sub-step: performing path analysis according to an initial position information and an end position information in the group position information to generate a path information correspondingly (S11); in this embodiment, the robot mechanism 10 according to the initial position information The object to be tested 30 is moved to an initial position, so that the image processing device 20 can capture an image, and the robot 10 moves the object to be tested 30 to an end position according to the end position information, so that the image is displayed. The processing device 20 stops capturing the image; the image processing module 20 transmits the path information to the automatic mechanical device 10 to make the automatic mechanical device 1 0 is generated according to the path information; a plurality of image capture points are generated according to the path information, and corresponding trigger signals are generated according to the image capture points (S12); in this embodiment, the path information includes the Moving distance and moving time of the measured article 30 from the initial position to the ending position; the image capturing points, according to the number of images to be captured and the image resolution, the moving distance in the path information, etc. The spacing method generates the image capturing points; or the moving time in the path information is generated at a fixed interval time.

The robot mechanism 10 absorbs the object to be tested 30 according to the path information, and when the object to be tested 30 moves to each image capturing point, the image processing device 20 triggers according to each image capturing point. The signal, and when the object to be tested 30 continues to move, continuously captures most of the appearance images of the object to be tested 30, and the images of the contiguous regions of the successive appearance images are the same.

The image processing device 20 compares the contiguous regions of the successive appearance images, and when the images of the contiguous regions of the successive appearance images are the same, compares with the corresponding images in the set of sample images; When the image of the splicing area is different, the image processing apparatus 20 outputs a first abnormality comparison information for the detecting personnel to know that the appearance of the object to be tested 30 is abnormal, and simultaneously transmits to the automatic mechanical device 10 to make the image The robot 10 transfers the article to be tested 30 to an abnormal area for placement.

Therefore, it can be seen from the above that after the images of the contiguous areas of the successive appearance images are the same, in order to further confirm whether the appearance image is abnormal, the image processing module 20 corresponds the appearance images to the set of sample images. When the image is different from the corresponding image in the sample image, the output comparison information corresponds to a second abnormal comparison information for the detecting personnel to know the object to be tested 30. The appearance is abnormal, and is simultaneously transmitted to the robot mechanism 10, so that the robot mechanism 10 transfers the object to be tested 30 to the abnormal area; if the appearance images and the corresponding images in the group of sample images are In the same case, the comparison information output by the image processing device 20 corresponds to a correct comparison information, and the detecting personnel knows that the appearance of the object to be tested 30 is normal.

In summary, the present invention moves the object to be tested 30 by the robot mechanism 10 and moves through the image processing device 20 during the movement of the object to be tested 30. Taking the appearance images, and the images of the successive areas of the successive appearance images in the captured appearance image are the same, and the image device 20 compares the images of the connection areas of the successive appearance images to determine whether the appearance images are the same. After the discontinuity or abnormality, the appearance images are compared with the corresponding images in the set of sample images to further confirm whether the appearance images are normal, and the images are continuously captured to improve the comparison. The situation that the number of images is insufficient, thereby achieving the purpose of improving detection efficiency and accuracy.

10‧‧‧Automatic mechanical devices

11‧‧‧Base

12‧‧‧ mechanical suction body

13‧‧‧Signal processing unit

20‧‧‧Image processing device

21‧‧‧Image capture device

22‧‧‧ display

23‧‧‧Image Analyzer

30‧‧‧ Items to be tested

1 is a perspective view of an automatic mechanical device and an image processing device according to a preferred embodiment of the present invention. 2 is a schematic view showing the operation of the automatic mechanical device and the image processing device according to the preferred embodiment of the present invention. 3 is a schematic view of an appearance image of a preferred embodiment of the present invention. 4 is a flow chart of a first method of a preferred embodiment of the present invention. Figure 5 is a flow chart of a second method of the preferred embodiment of the present invention.

Claims (9)

An object appearance detecting method with a continuous image capturing function is performed by an image processing device, and the image processing device is electrically connected to an automatic mechanical device for moving an object to be tested and pre-set with more than one set. Position information, the image processing device presets a plurality of sample images, wherein the robot mechanism comprises a base, a mechanical suction body and a signal processing device, the mechanical suction body is disposed on the base, and the signal processing device is configured In the base and electrically connected to the mechanical suction body, the mechanical suction body is configured to absorb the object to be tested for movement, and the signal processing device is configured to process the received information; the set of position information is preset to the signal Processing the device, and the method includes the steps of: reading a preset set of position information in the automechanical device to generate a path information correspondingly, and generating a majority of the trigger signals according to the path information; according to the trigger signals, When the automatic mechanical device absorbs the movement of the object to be tested, corresponding to the item to be tested Appearance, continuously capturing most of the appearance images, and the images of the contiguous areas of the successive two appearance images are the same; comparing the images of the contiguous areas of the continuation appearance images; comparing the appearance images with the group of sample images Yes, to output a comparison information. The method for detecting an appearance of an item having the function of continuously capturing images as described in claim 1 is: “reading a set of position information preset in the automatic mechanical device to generate a path information correspondingly, and corresponding to the path information according to the path information. In the step of generating a majority of trigger signals, the method further includes the following steps: performing path analysis according to an initial position information and an end position information in the group position information to generate the path information correspondingly; and generating a majority image according to the path information. The points are taken, and corresponding trigger signals are generated according to the image capturing points. The method for detecting the appearance of an item having the function of continuously capturing images as described in claim 2, wherein each of the appearance images includes a first connection area and a second connection area, and the image processing device transmits the two adjacent appearance images. The first contiguous area of the next appearance image is compared with the second contiguous area of the previous appearance image, when the first contiguous area of the latter appearance image is different from the second contiguous area of the previous appearance image, The image processing device sends a first abnormal comparison information to the automatic mechanical device to cause the automatic mechanical device to suck the object to be tested to an abnormal region for placement. The method for detecting the appearance of an item having the function of continuously capturing images as described in claim 3, wherein the first appearance image of the plurality of consecutive appearance images is a starting appearance image, and the last appearance image is an end image. The first image of the initial appearance image is an image of the environment, and the second connection area of the finished image is an image of the environment, and the image processing device selects the second connection region of the initial appearance image. Comparing the first contiguous regions of the adjacent next appearance images, and when not identical, transmitting the first abnormal comparison information; the image processing device is configured to overlap the first contiguous region of the appearance image with the adjacent The second consecutive areas of the previous appearance image are compared, and when they are not the same, the first abnormal comparison information is correspondingly transmitted. The method for detecting the appearance of an item having the function of continuously capturing images according to claim 4, wherein when the one of the appearance images and the image corresponding to the group of sample images are different, the comparison information output by the image processing device corresponds to one The second abnormality compares the information so that the robot moves the object to be tested and moves to the abnormal area for placement. The method for detecting the appearance of an item having the function of continuously capturing images as described in claim 5, wherein when the image corresponding to the image of the set of sample images is the same, the comparison information output by the image processing device corresponds to A correct comparison of information. The method for detecting an appearance of an item having the function of continuously capturing images according to claim 6, wherein the path information includes moving the item to be tested from the corresponding initial position to the corresponding end position. The moving distance is generated according to the number of images to be captured and the image resolution, and the moving distance of the path information is generated in an equally spaced manner. The method for detecting an appearance of an item having the function of continuously capturing images according to claim 6, wherein the path information includes a moving time of moving the item to be tested from a corresponding initial position to a corresponding ending position, and the required time is as follows. The number of images and the resolution of the image, and the moving time of the path information is generated at a fixed interval. The method for detecting an appearance of an item having a continuous image capturing function according to claim 7 or 8, wherein the image processing device comprises an image capturing device, a display and an image analyzer, wherein the image analyzer separately captures the image The image analyzer is electrically connected to the signal processing device, and the display is configured to display the first abnormality comparison information, the second abnormal comparison information, and the correct comparison information, and the image capture The image is captured by the image to be tested according to the trigger signals, and the image analyzer is configured to process the received image information; the set of sample images is preset in the image analyzer.