TWM605997U - Optical inspection system - Google Patents

Abstract

This creation is about an optical inspection system. The optical detection system includes an automatic control device, a light source, and a control detection module. The automatic control device has an image sensor, and the control detection module includes a path unit, a comparison unit, and a display unit. First, the path information is calculated by the path unit as the movement path of the automatic control device. Then the user excludes those with a large detection angle in the movement path based on the path information to generate the best movement path information. The best moving path information, each light source is turned on or off separately, and then the device is automatically controlled to move along the best moving path, the image information is generated by the light sensor, and finally the defects in the image information are identified by the comparison module Generate comparison image information, and display the comparison image information through the monitor. In this way, the defects existing on the object to be tested are detected, and the purpose of detecting the defects and reducing the detection time is achieved.

Description

Optical inspection system

This creation is about a detection system, especially an optical detection system.

In recent years, in order to comply with the popular trend of vehicle styling design, the rim size has gradually increased. The standard rims equipped on commercial vehicles have changed from 15 inches to 16 inches, and there is a trend towards 17 inches. The larger the rim, the heavier the weight. In the era of high oil prices where energy saving and low fuel consumption are important, light alloy wheels made of aluminum or magnesium alloys have gradually become the mainstream in the market.

Light alloy wheels have many advantages. Taking a diameter of 13 inches as an example, aluminum alloy wheels are about 11% lighter than steel rims, and 14 inches are about 22% lighter. The larger the wheel diameter, the greater the weight difference between the two. The rim is light in weight, and the inertial resistance of the driving rim is relatively small, which will greatly enhance the sensitivity of the user's manipulation and save a little fuel consumption. In addition, taking the aluminum alloy wheel as an example, its heat transfer performance is good and helps to ventilate and dissipate heat, which can increase the life of the tire and reduce the risk of tire blowout.

However, the disadvantage of using light alloy rims is that the elongation of alloy materials is generally lower than that of steel, which results in poor deformability of light alloy rims, and poor quality light alloy rims are prone to the danger of fracture. On the other hand, because the manufacturing process is more complicated, including material composition, casting technology, mold design, heat treatment process, cutting and painting, etc., it is easier to cause problems due to human error. Therefore, how to accurately identify and detect defects on light alloy wheels is one of the problems that R&D personnel should solve.

Therefore, after observing the above-mentioned deficiencies, the creator of this case has produced this creation.

The purpose of this creation is to provide an optical inspection system, which can generate corresponding optimal movement path information according to the structure of the object to be measured, and adjust the light source according to the optimal movement path to detect defects on the surface of the object to be measured, and Significantly reduce the detection time.

To achieve the above objective, this creation provides an optical inspection system, which includes: an automatic control device, which is arranged around an object to be measured, the automatic control module has an image sensor, and the image sensor Shooting the object under test and generating plural image information; plural light sources, which surround the object under test, each of the light sources is turned on or off according to an optimal movement path information of the automatic control device; and a control detection The module is electrically connected to the automatic control equipment and the light sources. The control detection module includes a path unit, a comparison unit and a display unit; wherein, the path unit is directed to the object under test Structure, a path information is calculated, and the path information is used as the moving path of the automatic control device; the comparison unit is electrically connected to the path unit, and the comparison unit stores multiple defect image information, the comparison unit The pairing unit receives the image information generated by the image sensor, and uses the defective image information to identify the defects existing in the image information to generate plural comparison image information; and the display unit, which is It is electrically connected to the comparison unit, and the comparison unit transmits the comparison image information to the display unit, and the display unit displays the comparison image information or the image information.

Preferably, according to the optical inspection system of the present invention, the automatic control equipment can use one of a three-axis robotic arm and a multi-axis robotic arm.

Preferably, according to the optical detection system of the present invention, the control detection module can be one of a computer, a server, and a smart phone.

Preferably, according to the optical inspection system of the present invention, the image sensor can be one of an industrial camera and a high-speed camera.

Preferably, according to the optical inspection system of the present invention, the control inspection module is one of a computer, a server, and a smart phone.

Preferably, according to the optical inspection system of the present invention, the control inspection module further includes a setting unit for setting the structure of the test object.

Preferably, according to the optical inspection system of the present invention, the comparison module filters the noise in the image information through Gaussian Blur and Binarization to improve the recognition of the image information The accuracy of the flaws on the

Preferably, according to the optical inspection system of the present invention, the comparison module is combined with a convolutional neural network to identify defects in the image information.

Preferably, according to the optical inspection system of the present invention, the comparison unit can respectively identify two types of defects: dirt particles and paint scratches.

Preferably, according to the optical inspection system of the present invention, the comparison image information generated by the comparison unit is marked with a first warning pattern on the dirt particles and a mark on the scratched part of the paint The second warning pattern.

To sum up, the optical inspection system provided by this creation mainly uses the optical inspection system of this creation and matches its detection method. It can generate the corresponding optimal movement path information according to the structure of the object to be measured, and can according to the optimal movement path Adjust the light source to detect defects on the surface of the object to be tested and greatly reduce the inspection time. As a result, this creation will not be affected by the structure of the object to be tested, and can quickly and comprehensively detect the surface defects of the object to be tested.

In order to enable those familiar with the art to understand the purpose, features and effects of this creation, the following specific embodiments are used in conjunction with the accompanying drawings to explain this creation in detail as follows.

The following is a more detailed description of the implementation of this creation in conjunction with the drawings and component symbols, so that those with ordinary knowledge in the technical field can implement it after studying this specification.

However, the present creation is not limited to the embodiments disclosed herein, but will be implemented in various forms.

The following embodiments are only provided as examples, so that those with ordinary knowledge in the technical field can fully understand the disclosure of this creation and the scope of this creation.

Therefore, this creation will only be limited by the scope of the attached patent application.

In the drawings used to describe various embodiments of the present creation, the shapes, sizes, ratios, numbers, etc. shown are merely exemplary, and the present creation is not limited thereto.

In this specification, the same reference numerals generally denote the same elements.

Unless expressly stated otherwise, any reference to the singular may include the plural.

Please refer to FIG. 1 to FIG. 4. FIG. 1 is a schematic diagram of the optical detection system according to the first embodiment of the present invention. As shown in FIG. 1, the optical inspection system 100 according to the present creation includes: an automatic control device 10, a control inspection module 20, and a light source 30.

Specifically, in this embodiment, the test object 200 is a wheel rim, and its material can be one or a combination of iron, aluminum, and magnesium, but the invention is not limited to this.

The automatic control device 10 is arranged around the object 200 to be measured, and the automatic control device 10 has an image sensor 11, and the image sensor 11 photographs the object 200 and generates plural image information 111.

Specifically, the automatic control device 10 according to the present creation may be one of a three-axis robotic arm and a multi-axis robotic arm, but the present creation is not limited to this. The automatic control device 10 is electrically connected to the control detection module 20, and receives instructions from the control detection module 20 to control and move, so that the image sensor 11 on the automatic control device 10 can completely photograph the surface of the object 200 under test.

Specifically, the image sensor 11 according to this creation can be one of an industrial camera and a high-speed camera, but the creation is not limited to this. It should be further explained that the advantages of using an industrial camera are higher stability, easy installation, stable structure, not easy to be damaged, and long continuous working time, etc., compared to using a general camera. Therefore, the image sensor 11 uses an industrial camera for better shooting effects. However, the disadvantage is that the operating cost is greatly increased and the operating time is longer.

Please refer to FIG. 2 and FIG. 3. FIG. 2 is a schematic diagram of the installation position of the optical detection system according to the first embodiment of the invention, and FIG. 3 is a schematic diagram of the optical detection system architecture according to the first embodiment of the invention. The control and detection module 20 according to this creation is electrically connected to the automatic control device 10 and the light source 30. The control and detection module 20 includes a path unit 21, a comparison unit 22, and a display unit 23.

Specifically, the control detection module 20 according to this creation can be one of a computer, a server, and a smart phone, but the creation is not limited to this.

The path unit 21 calculates the path information 211 (not shown) as the moving path of the automatic control device 10 according to the structure of the object 200 under test. When the path information 211 is used as the moving path of the automatic control device 10, it can be guaranteed The image sensor 11 on the automatic control device 10 completely photographs the surface of the object 200 under test.

It should be further explained that the path information 211 is not the best moving path of the automatic control device 10, and some of the movements may include a large detection angle, resulting in a longer detection time. Therefore, according to the first embodiment of the present invention, the optical In the detection system 100, after the path unit 21 calculates the path information 211 as the moving path of the automatic control device 10, the user can remove the path information with a large detection angle according to the path information 211, thereby obtaining the best moving path The information 212 serves as a moving path of the automatic control device 10, effectively reducing the inspection time of the optical inspection system 100.

Please refer to FIG. 2 and in conjunction with FIG. 5, the comparison unit 22 is electrically connected to the path unit 21, the comparison unit 22 stores a plurality of defect image information 221 (not shown), the comparison unit 22 The image information 111 generated by the image sensor 11 is received, and the defect existing on the image information 111 is identified by the defect image information 221 to generate a plurality of comparison image information 222.

Specifically, according to the defect image information 221 of the first embodiment of the present creation, the defect may include impurities, bumps, black spots, cotton wool, shrinkage holes, cavities, abrasion marks, scratches, scratches, bruises, and different colors. , Knife marks, lack of meat, hair, cracked paint, flow paint, thin spray, high-gloss silver difference, etc., all of the above are various manufacturing errors that occur in various stages of rim production, but this creation is not limited to this.

The display unit 23 is electrically connected to the comparison unit 22. The comparison unit 22 transmits the comparison image information 222 to the display unit 23, and the display unit 23 displays the comparison image information 222 or the images Information 111. Specifically, the display unit 23 according to the present creation may be a liquid crystal display, a light emitting diode display or an organic light emitting diode display, but the present creation is not limited to this.

Please refer to FIG. 4 in conjunction with FIG. 3. FIG. 4 is a schematic diagram of the light source setting position according to the first embodiment of the present creation. In this embodiment, the light sources 30 are arranged around the object under test 200, and the specific number of the light sources 30 is 7. However, the present creation is not limited to this, and the user can increase or decrease the number of the light sources 30 according to needs, and The light source 30 is electrically connected to the control detection module 20. Specifically, the light source 30 according to the present invention may use one of a light emitting diode and an organic light emitting diode or a combination thereof, but the present invention is not limited to this.

It should be further explained that, as described above, according to the optical inspection system 100 of the present creation, the optimal movement path information 212 generated according to the structure of the object 200 to be tested is used as the movement path of the automatic control device 10. In this implementation In an example, since the object 200 is made of metal, its surface is shiny and easy to reflect light. As a result, when the automatic control device 10 moves and shoots the surface of the object 200 from different angles, the image sensor 11 cannot capture clear images. Therefore, it is difficult for the comparison unit 22 to identify the flaws, resulting in incorrect detection results or increased detection times.

Therefore, in this embodiment, the user can turn on or off each of the light sources 30 separately according to the optimal movement path information 212 as the movement path of the automatic control device 10, so as to reduce the reflection of the surface of the object to be measured. Therefore, it is ensured that the image sensor 11 captures clear image information 111 for the comparison unit 22 to identify defects on the image information 111, so as to improve the accuracy of the optical inspection system 100 according to the present invention to detect defects.

Please refer to FIG. 5, which is a schematic flowchart of the process of identifying image information by the comparison unit according to the first embodiment of the creation. The comparison unit 22 first performs Gaussian filtering on the image information 111, and replaces the weighted average of the pixels in the image information 111 with the value of each pixel in the image information 111, reducing the dramatic changes in the gray level of the image information 111 and reducing the noise. However, as shown in FIG. 5, after Gaussian filtering, the edges and contours of the image information 111 will be blurred. Then the comparison unit 22 binarizes the image information 111, sets the grayscale of the pixels in the image information 111 greater than the preset threshold gray value as the maximum gray value, and sets the pixels less than the preset threshold gray value The grayscale is set to a minimum grayscale value, so as to achieve binarization and filter the noise in the image information 111, thereby improving the accuracy of the comparison unit 22 in identifying the defects existing on the image information 111.

In order to provide a further understanding of the structural features of this creation, the use of technical means and the expected effects, the method of use of this creation is described. I believe that we can have a deeper and specific understanding of this creation from this, as follows:

Please refer to Figures 6 and 7, and in conjunction with Figures 3 to 5, Figure 6 is a schematic diagram of the identification result of the optical detection system according to the first embodiment of the creation, and Figure 7 is a diagram illustrating the implementation of the first embodiment of the creation The flow chart of the inspection method of the optical inspection system. This creation is based on the above-mentioned optical detection system 100, and further provides a detection method of the optical detection system 100, which includes the following steps:

In the simulation step S1, the optical inspection system 100 calculates the path information 211 by the path unit 21 according to the structure of the object 200. The path information 211 can be used as the movement path of the automatic control device 10, and then the correction step S2 is performed.

In the correction step S2, the user removes those with a large detection angle according to the path information 211, thereby obtaining the best moving path information 212 as the moving path of the automatic control device 10 to reduce the detection time, and then execute the correction step S3.

In the calibration step S3, the user individually activates or deactivates each of the plurality of light sources 30 according to the optimal movement path information 212 generated in the calibration step S2, thereby ensuring that the image sensor 11 can capture clear image information 111. Perform shooting step S4.

In the shooting step S4, the automatic control device 10 moves according to the optimal movement path information 212 generated in the correction step S2, and uses the image sensor 11 to photograph the surface of the object 200 to generate clear image information 111, and perform comparison To step S5.

In the comparison step S5, the image information 111 generated by the image sensor 11 is received by the comparison unit 22, and the image information 111 is identified by the defective image information identification 221 stored in the comparison unit 22 Existing defects, and generate a plurality of comparison image information 222, and then execute the display step S6.

In the display step S6, the comparison unit 22 transmits the comparison image information 222 to the display unit 23, and the display unit 23 displays the comparison image information 222.

For example, please refer to FIG. 6, and in conjunction with FIGS. 3 to 7. According to the optical inspection system 100 of the present invention, the simulation step S1 is first performed, and the path information 211 is calculated by the path unit 21; then the correction step S2 is performed, and the user removes those with a large detection angle according to the path information 211, thereby The best moving path information 212 is obtained as the moving path of the automatic control device 10; afterwards, the calibration step S3 is performed, and the user activates or deactivates the light sources 30 according to the best moving path information 212 generated in the correcting step S2. Each one; then perform the shooting step S4, the automatic control device 10 moves according to the best moving path information 212 generated in the correction step S2, and uses the image sensor 11 to shoot the surface of the object 200; after the shooting is completed, it enters the comparison Step S5, the comparison unit 22 is used to identify the defects in the image information 111 and generate a plurality of comparison image information 222; after the identification is completed, the display step S6 is executed, and the comparison unit 22 is used to compare the The image information 222 is transmitted to the display unit 23, and the display unit 23 displays the compared image information 222.

It should be further explained that in this embodiment, in the comparison step S5, the comparison unit 22 first performs Gaussian filtering on the image information 111, and replaces the weighted average of the pixels in the image information 111 in the image information 111. The value of each pixel in the image information 111 reduces the dramatic changes in the grayscale of the image information 111 to reduce noise. However, as shown in Figure 5, Gaussian filtering will inevitably cause the edges and contours of the image information 111 to be blurred. The information 111 is binarized, and the grayscale of pixels in the image information 111 that are greater than the preset critical gray value is set to the maximum gray value, and the grayscale of the pixel less than the preset critical gray value is set to the minimum gray value. In this way, binarization is achieved and the noise in the image information 111 is filtered, and the accuracy of the comparison unit 22 in identifying the defects existing on the image information 111 is improved.

Therefore, it can be known from the above description that the optical inspection system 100 provided by this creation and its inspection method can generate the best moving path information 212 for the structure of the object 200 during inspection, and by The best moving path information 212 is matched with a suitable light source 30 configuration to accurately identify defects on the surface of the object 200 to be measured. In this way, the optical inspection system 100 provided by this creation will not be affected by the structure of the object 200 , It can quickly and comprehensively detect defects on the surface of the test object 200.

Please refer to FIG. 8 and FIG. 9. FIG. 8 is a schematic diagram of the identification result of the optical inspection system according to the second embodiment of the invention, and FIG. 9 is a flowchart illustrating the steps of the inspection method of the optical inspection system according to the second embodiment of the invention. . Compared with the first embodiment, the main difference of the second embodiment is that the comparison unit 22 further combines a convolutional neural network (not shown) with the defect image information 221, thereby improving the comparison. The speed and accuracy of the unit 22 in recognizing the defects in the image information 111. It needs to be further explained that the Artificial Neural Network is a mathematical model or calculation model that imitates the structure and function of biological neural networks, and the convolutional neural network further adds the concept of blocks to Locking the data types with regional properties, in this way, combined with the use of convolutional neural networks can greatly improve the computing performance of the comparison unit 22. In the second embodiment, the comparison unit 22 can further identify the dirt particles and paint scratches respectively. Therefore, the comparison unit 22 can mark the first warning pattern 41 on the dirt particles and the paint scratches. Mark the second warning pattern 42 at the place.

In this way, the second embodiment can not only achieve the effects of the first embodiment, but also can further identify and classify the types of defects and mark obvious marks on the defects. The comparison unit 22 greatly improves by combining with convolutional neural networks. The applicability and convenience of the optical inspection system 100 based on this creation.

Please refer to FIG. 9 in conjunction with FIG. 8. FIG. 9 is a flowchart illustrating the steps of the warning method of the optical detection system according to the second embodiment of the present creation. This creation is based on the optical detection system 100 of the second embodiment, and further provides a detection method of the optical detection system 1, which includes the following steps:

In step S1', the optical inspection system 100 uses the path unit 21 to calculate the path information 211 based on the structure of the test object 200. The path information 211 can be used as the movement path of the automatic control device 10, and then the correction step S2' is performed.

In the correction step S2', the user removes those with a large detection angle according to the path information 211, thereby obtaining the best moving path information 212 as the moving path of the automatic control device 10 to reduce the detection time, and then executes the correction step S3.

In the calibration step S3', the user individually activates or deactivates each of the plurality of light sources 30 according to the optimal movement path information 212 generated in the calibration step S2', thereby ensuring that the image sensor 11 can capture clear image information 111 , And then perform the shooting step S4.

In the shooting step S4', the automatic control device 10 moves according to the optimal moving path information 212 generated in the correction step S2, and uses the image sensor 11 to photograph the surface of the object 200 to generate clear image information 111, and execute Compare step S5'.

In the comparison step S5', the image information 111 generated by the image sensor 11 is received by the comparison unit 22, and the defect image information 221 stored in the comparison unit 22 is identified with the convolutional neural network. In order to identify the defects existing on the image information 111, the comparison unit 22 can identify two types of defects, dirt particles and paint scratches, and generate a plurality of comparison image information 222, and then perform the marking step S7'.

In the marking step S7', the comparison unit 22 combined with the convolutional neural network is used to identify two types of defects: dirty particles and paint scratches. The comparison unit 22 marks the first on the dirty particles on the compared image information 222. The warning pattern 41 and the second warning pattern 42 are marked on the scratched part of the paint, and then the display step S6' is executed.

In the display step S6', the comparison unit 22 transmits the comparison image information 222 to the display unit 23, and the display unit 23 displays the comparison image information 222.

For example, please refer to FIG. 8, and in conjunction with FIG. 3, FIG. 4 and FIG. 9. According to the optical inspection system 100 of the present invention, the pseudo-step S1' is first performed, and the path information 211 is calculated by the path unit 21; then the correction step S2' is performed, and the user removes those with a large detection angle based on the path information 211 , The best moving path information 212 is obtained as the moving path of the automatic control device 10; afterwards, the calibration step S3' is executed, and the user turns on or off the light sources according to the best moving path information 212 generated in the correcting step S2. Each of 30; afterwards, the photographing step S4' is executed, and the automatic control device 10 moves according to the optimal movement path information 212 generated in the correction step S2', and photographs the surface of the object 200 to be measured by the image sensor 11; After completion, proceed to the comparison step S5'. The comparison unit 221 is combined with the convolutional neural network to identify the defects in the image information 111 and generate the plural comparison image information 222; then after the identification is completed The marking step S7' is performed, and the comparison unit 22 combines with the convolutional neural network to respectively identify two kinds of defects: dirty particles and paint scratches. The comparison unit 22 marks the dirty particles on the compared image information 222 The first warning pattern 41, and the second warning pattern 42 is marked on the scratched part of the paint; finally, the display step S6' is executed to include the comparison image information 222 of the first warning pattern 41 and the second warning pattern 42, The comparison unit 22 transmits to the display unit 23, and the display unit 23 displays the compared image information 222.

It is worth mentioning that although the above description is based on the comparison unit 22 identifying dirt particles and paint scratches, this creation is not limited to this, such as impurities, bumps, black spots, cotton wool, shrinkage holes, and cavities. , Wear marks, scratches, scratches, bumps, different colors, knife marks, lack of meat, hairs, chipped paint, flow paint, thin spray, high-gloss silver difference, etc., when there are enough flawed images stored inside the comparison unit 22 When the information 221 is combined with a convolutional neural network, the comparison unit 22 can further separately identify the above-mentioned defects or combinations thereof, and mark different warning patterns one by one to facilitate user identification.

In this way, this creation has the following implementation and technical effects:

First, based on the optical inspection system 100 of this creation, and in conjunction with the inspection method provided by this creation, it can generate the corresponding optimal movement path information 212 according to the structure of the object 200, and can be based on the most The optimal moving path 212 adjusts the light sources 30 to detect defects on the surface of the test object 200, and greatly reduces the inspection time.

Secondly, by combining the comparison unit 22 with the convolutional neural network, this creation can identify and classify the types of defects and mark the defects with obvious marks, which greatly improves the applicability and convenience of the optical inspection system 100 based on this creation. Sex.

The above is to illustrate the implementation of this creation through specific specific embodiments. Those with ordinary knowledge in the art can easily understand the other advantages and effects of this creation from the content disclosed in this specification.

Although the creation is described by referring to the embodiments depicted in the drawings, it is only an embodiment, and those with ordinary knowledge in the art should understand that various changes and modifications can be made to it. However, these changes and deformations should not deviate from the scope of protection of this creation. Therefore, the scope of protection of this creation must be limited to the scope of the attached patent application.

100: Optical inspection system 10: Automatic control equipment 11: Image sensor 111: Image Information 20: Control detection module 21: Path unit 211: Path Information 212: Best moving path information 22: Comparison unit 221: Defective image information 222: Compare image information 23: display unit 30: light source 41: The first warning pattern 42: The second warning pattern 200: DUT S1: Simulation steps S2: Correction steps S3: Calibration steps S4: shooting steps S5: Comparison steps S6: Show steps S1’: Simulation steps S2’: Correction steps S3’: Calibration steps S4’: Shooting steps S5’: Comparison step S6’: Show steps S7’: marking steps

Fig. 1 is a schematic diagram of an optical inspection system according to the first embodiment of the invention; Fig. 2 is a schematic diagram of the position of the optical detection system according to the first embodiment of the present creation; 3 is a schematic diagram of the optical inspection system architecture according to the first embodiment of the present invention; Fig. 4 is a schematic diagram of a light source setting position according to the first embodiment of the creation; FIG. 5 is a schematic flowchart of the process of identifying image information by the comparison unit according to the first embodiment of the creation; 6 is a schematic diagram of the identification result of the optical inspection system according to the first embodiment of the present creation; FIG. 7 is a flowchart illustrating the steps of performing the inspection method of the optical inspection system of the first embodiment of the creation; 8 is a schematic diagram of the identification result of the optical inspection system according to the second embodiment of the present creation; FIG. 9 is a flowchart illustrating the steps of the inspection method of the optical inspection system according to the second embodiment of the present creation.

100: Optical inspection system

10: Automatic control equipment

11: Image sensor

20: Control detection module

21: Path unit

22: Comparison unit

23: display unit

30: light source

Claims (8)

An optical detection system includes: an automatic control device, which is arranged around an object to be measured, the automatic control module has an image sensor, and the image sensor photographs the object to be measured and generates a plurality of Image information; a plurality of light sources, the light sources are arranged around the object to be measured, each of the light sources is turned on or off according to an optimal movement path information of the automatic control device; and a control detection module, which is Electrically connected to the automatic control equipment and the light sources, the control detection module includes a path unit, a comparison unit, and a display unit; wherein the path unit calculates a path unit for the structure of the object to be tested Path information, the path information is used as the movement path of the automatic control device; the comparison unit is electrically connected to the path unit, the comparison unit stores multiple defect image information, and the comparison unit receives the image The image information generated by the sensor, and the defects existing on the image information are identified by the defect image information to generate plural comparison image information; and the display unit is connected to the comparison unit Electrically connected, the comparison unit transmits the comparison image information to the display unit, and the display unit displays the comparison image information and one or a combination of the image information. The optical inspection system according to claim 1, wherein the automatic control device is one of a three-axis robotic arm and a multi-axis robotic arm. The optical inspection system according to claim 1, wherein the image sensor is one of an industrial camera and a high-speed camera. The optical detection system according to claim 1, wherein the control detection module is one of a computer, a server, and a smart phone. The optical inspection system according to claim 1, wherein the comparison unit filters the noise in the image information through Gaussian filtering and binarization, so as to improve the accuracy of identifying defects in the image information. The optical inspection system according to claim 1, wherein the comparison unit is combined with a convolutional neural network to identify defects in the image information. The optical inspection system according to claim 6, wherein the object to be measured is a rim, and the defects on the image information include one or a combination of dirt particles and paint scratches. The optical inspection system according to claim 7, wherein the comparison image information generated by the comparison unit, the comparison image information is to mark a first warning pattern on the dirt particles, and to scratch the paint The wound is marked with a second warning pattern.

Images