Industrial robot visual inspection execution system

Abstract

The invention provides an industrial robot visual inspection execution system, which comprises a visual sensor module, the conveying end of the visual sensor module is in wireless connection with an illumination system module, and an image acquisition module is in wireless connection with the conveying end of the illumination system module. The image processing module is wirelessly connected to the conveying end of the image acquisition module, and the image processing module comprises an image preprocessing unit, an image segmentation unit, a feature extraction unit and a target identification and classification unit. The invention provides a visual inspection execution system of an industrial robot. An image preprocessing unit, an image segmentation unit, a feature extraction unit and a target recognition and classification unit are respectively arranged in an image processing module in the whole system, so that the data volume and complexity can be reduced, the image quality is improved, a target and a background are separated, the edge and area of an object are accurately extracted, the features of the object are quantified, and a basis is provided for recognition and classification; the target object is accurately identified; and object classification automation is realized.

Description

Technical Field

[0001] The present invention relates to the field of industrial robots, and in particular to an industrial robot visual detection execution system. Background Art

[0002] An industrial robot is a programmable machine device that can automatically perform work tasks. It is widely used in industrial production environments, greatly improving production efficiency and product quality.

[0003] At present, different industrial robots are used in factories to assist operations in different operating environments. Visual inspection robots are used for object inspection. When using visual inspection robots, an inspection execution system is required to control the robot.

[0004] However, the inspection execution system set up in the current visual inspection robot can perform image capture, image processing, data transmission and communication operations on objects. However, when capturing images of objects, it is necessary to capture clear pictures of the objects during the process of designing the system. If the object capture is not clear, it will affect the inspection of the object.

[0005] Therefore, it is necessary to provide an industrial robot visual inspection execution system to solve the above technical problems. Summary of the invention

[0006] The present invention provides an industrial robot visual inspection execution system, which solves the problem that the inspection execution system arranged in the existing visual inspection robot can perform image acquisition, image processing and data transmission and communication operations on an object, but when performing image acquisition of the object, it is necessary to clearly capture the image of the object during the process of designing the system, and the inspection of the object will be affected when the image acquisition is not clear.

[0007] In order to solve the above technical problems, the present invention provides an industrial robot visual inspection execution system, comprising:

[0008] A visual sensor module, wherein a delivery end of the visual sensor module is wirelessly connected to a lighting system module;

[0009] An image acquisition module, the image acquisition module is wirelessly connected to the transmission end of the lighting system module;

[0010] An image processing module, the image processing module is wirelessly connected to the transmission end of the image acquisition module, and the image processing module includes an image preprocessing unit, an image segmentation unit, a feature extraction unit, and a target recognition and classification unit;

[0011] A robot module, wherein the robot is wirelessly connected to the conveying end of the image processing module, the conveying end of the robot module is wirelessly connected to a robot control module, and the conveying end of the robot control module is wirelessly connected to a data transmission and communication module.

[0012] Preferably, the image preprocessing module includes a grayscale processing subunit, a noise reduction processing subunit and a contrast enhancement subunit.

[0013] Preferably, the image segmentation unit includes a threshold segmentation subunit and an edge detection subunit.

[0014] Preferably, the feature extraction unit includes a geometric feature extraction subunit, a texture feature extraction subunit and a color feature extraction subunit.

[0015] Preferably, the target recognition and classification unit includes a template matching subunit and a machine learning classification subunit.

[0016] Preferably, the data transmission and communication module includes a physical interface unit, a data link protocol unit, a network layer routing unit and a transport layer protocol unit.

[0017] Preferably, the physical interface unit includes a wired interface subunit and a wireless interface subunit, the wired interface subunit includes an Ethernet interface, a fiber optic interface and an industrial field bus interface, and the wireless interface subunit includes a wireless network interface and a Bluetooth interface.

[0018] Preferably, the data link protocol unit includes an error control subunit and a flow control subunit, the error control subunit includes a cyclic redundancy check and forward correction, and the flow control subunit includes window-based flow control and rate adaptation.

[0019] Preferably, the network layer routing unit includes a static routing subunit and a dynamic control subunit.

[0020] Preferably, the transport layer protocol unit includes a TCP protocol subunit and a UDP protocol subunit.

[0021] Compared with the related art, the industrial robot visual inspection execution system provided by the present invention has the following beneficial effects:

[0022] The present invention provides an industrial robot visual inspection execution system. An image preprocessing unit, an image segmentation unit, a feature extraction unit and a target recognition and classification unit are respectively arranged in an image processing module in the whole system, which can reduce the amount and complexity of data, improve image quality, separate the target from the background, accurately extract the edge and area of ​​the object, quantify the object features, provide a basis for recognition and classification, accurately recognize the target object and realize automatic object classification. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 A schematic structural diagram of a first embodiment of an industrial robot visual inspection execution system provided by the present invention;

[0024] Figure 2 for Figure 1 The framework schematic diagram of the image processing module shown;

[0025] Figure 3 for Figure 1 The flowchart of the visual inspection execution system shown;

[0026] Figure 4 A schematic diagram of the structure of a second embodiment of an industrial robot visual inspection execution system provided by the present invention;

[0027] Figure 5 This is a schematic structural diagram of a third embodiment of an industrial robot visual inspection execution system provided by the present invention. DETAILED DESCRIPTION

[0028] The present invention will be further described below in conjunction with the accompanying drawings and implementation modes.

[0029] First embodiment

[0030] Please refer to Figure 1 , Figure 2 and Figure 3 ,in, Figure 1 A schematic structural diagram of a first embodiment of an industrial robot visual inspection execution system provided by the present invention; Figure 2 for Figure 1 The framework schematic diagram of the image processing module shown; Figure 3 for Figure 1 A process diagram of a visual inspection execution system is shown. An industrial robot visual inspection execution system includes:

[0031] A visual sensor module, wherein a delivery end of the visual sensor module is wirelessly connected to a lighting system module;

[0032] An image acquisition module, the image acquisition module is wirelessly connected to the transmission end of the lighting system module;

[0033] An image processing module, the image processing module is wirelessly connected to the transmission end of the image acquisition module, and the image processing module includes an image preprocessing unit, an image segmentation unit, a feature extraction unit, and a target recognition and classification unit;

[0034] A robot module, wherein the robot is wirelessly connected to the conveying end of the image processing module, the conveying end of the robot module is wirelessly connected to a robot control module, and the conveying end of the robot control module is wirelessly connected to a data transmission and communication module.

[0035] The image preprocessing module includes a grayscale processing subunit, a noise reduction processing subunit and a contrast enhancement subunit.

[0036] Grayscale processing subunit: used to convert color images into grayscale images, based on weighted average method, etc., with color image as input and grayscale image as output, connected to the image segmentation unit.

[0037] Noise reduction processing subunit: It includes three parts: mean filtering, median filtering, and Gaussian filtering. Mean filtering removes noise by calculating the average value of neighborhood pixels; median filtering sorts the neighborhood pixel values ​​and replaces the pixels with the middle value; Gaussian filtering calculates the weighted average value according to the Gaussian function. The input of this subunit is the original image or the grayscale image, and the output is the denoised image, which is connected to the image segmentation unit.

[0038] Contrast enhancement subunit: includes histogram equalization and linear contrast stretching. Histogram equalization enhances contrast by adjusting the image histogram to make the pixel distribution uniform; linear contrast stretching enhances contrast by stretching the pixel value range through linear transformation. The input is an image and the output is a contrast-enhanced image, which is connected to the image segmentation unit.

[0039] The image segmentation unit includes a threshold segmentation subunit and an edge detection subunit.

[0040] Threshold segmentation subunit: According to the comparison between the pixel gray value and the set threshold (fixed threshold or adaptive threshold), the image is divided into the target object and the background area. The input is the preprocessed image and the output is the segmented image, which is connected to the feature extraction unit.

[0041] Edge detection subunit: includes Sobel operator, Canny operator and segmentation based on region growing.

[0042] The Sobel operator detects edges by calculating horizontal and vertical gradients; the Canny operator accurately detects edges through a multi-stage algorithm; the segmentation based on region growing starts from the seed point and merges the surrounding pixels according to the similarity criterion. The input is an image and the output is an image after edge detection or segmentation, which is connected to the feature extraction unit.

[0043] The feature extraction unit includes a geometric feature extraction subunit, a texture feature extraction subunit and a color feature extraction subunit.

[0044] Geometric feature extraction subunit: calculates shape features such as circumference, area, circularity, rectangularity and size features such as length, width and height related to the shape of the object. The input is the segmented image and the output is geometric feature data, which is connected to the target recognition and classification unit.

[0045] Texture feature extraction subunit: calculates contrast, correlation, energy, entropy and other texture features through gray-level co-occurrence matrix and generates binary code using local binary patterns and statistical patterns to describe texture. The input is image and the output is texture feature data, which is connected to the target recognition and classification unit.

[0046] Color feature extraction subunit: uses color histogram to count the number of pixels in different color intervals to extract color features. The input is a color image and the output is color feature data, which is connected to the target recognition and classification unit.

[0047] The target recognition and classification unit includes a template matching subunit and a machine learning classification subunit.

[0048] Template matching subunit: compares the features of the object to be detected with the pre-stored standard template, based on pixel or feature matching, measures the similarity by calculating the correlation coefficient or mean square error, etc. The input is feature data, the output is the matching result, and is connected to the output result.

[0049] Machine learning classification subunit: including supervised learning and unsupervised learning. Supervised learning uses labeled training data to train the classification model; unsupervised learning uses clustering algorithms to classify based on feature similarity. The input is feature data, and the output is the classification result, which is connected to the output result.

[0050] Output results: Receive the results of target recognition and classification units for subsequent processing or display.

[0051] Process steps:

[0052] S1. Image input: Start the industrial robot visual inspection execution system and input the image to be processed;

[0053] S2. Image preprocessing

[0054] S21, grayscale processing: converting the color image into a grayscale image through the grayscale processing subunit, using a weighted average method for conversion;

[0055] S22, noise reduction processing: In the noise reduction processing subunit, a suitable filtering method is selected according to the noise type. If the noise intensity is small, mean filtering can be selected to calculate the average value of the neighboring pixels around the pixel point instead of the pixel value;

[0056] If there is a lot of salt and pepper noise, use median filtering to sort the neighborhood pixel values ​​and replace the pixel points with the median value;

[0057] For Gaussian noise, Gaussian filtering is used to calculate the weighted average of pixel values ​​in the neighborhood according to the Gaussian function;

[0058] S23, contrast enhancement: In the contrast enhancement subunit, histogram equalization or linear contrast stretching can be selected. Histogram equalization adjusts the histogram by calculating the image cumulative distribution function and mapping it to a new pixel value range to make the pixel distribution uniform; linear contrast stretching performs linear transformation by specifying a pixel value range;

[0059] S3. Image segmentation

[0060] S31. Threshold segmentation: In the threshold segmentation subunit, select the fixed threshold method or the adaptive threshold method according to the grayscale difference between the target object and the background of the image. If the grayscale difference is large, you can manually set the fixed threshold and compare the pixel grayscale value with the threshold. If it is greater than the threshold, it is the target object part, and if it is less than the threshold, it is the background part. If the lighting is uneven, use the adaptive threshold method to automatically adjust the threshold according to the grayscale characteristics of the local area of ​​the image.

[0061] S32, edge detection and region segmentation: In the edge detection subunit, the Sobel operator can be used to calculate the gradient amplitude through two 3×3 convolution kernels and image convolution, and the horizontal and vertical gradient amplitudes are combined to determine the edge position; for situations where more precise edges are required, such as high-precision electronic component detection, the Canny operator is used to detect edges through Gaussian filtering noise reduction, calculation of gradient amplitude and direction, non-maximum suppression and double threshold detection; segmentation based on regional growth can also be used, starting from one or more seed points, and merging surrounding pixels according to criteria such as pixel grayscale value or texture similarity until growth cannot be achieved, which is used to segment object regions with similar features;

[0062] S4. Feature extraction

[0063] S41, geometric feature extraction: In the geometric feature extraction subunit, the shape features of the object are calculated, such as by calculating the number of edge pixels or calculating the perimeter according to the edge curve mathematical model, estimating the area by counting the number of pixels in the target area, calculating the circularity of a circular object and the rectangularity of a rectangular object, etc.; at the same time, the size features of the object, such as length, width, height and other information, are extracted for industrial parts size measurement;

[0064] S42, texture feature extraction: In the texture feature extraction subunit, the gray level co-occurrence matrix is ​​used to calculate the probability of two pixels at a certain distance and angle in the image appearing at the same time, and texture features such as contrast, correlation, energy, and entropy are extracted from them to detect the uniformity of fabric surface texture; or a local binary pattern is used to compare each pixel with its neighboring pixels to generate a binary code, and the binary code pattern is statistically described to describe the texture. This method has a certain robustness to changes in illumination and can be used in the fields of industrial product surface texture detection and face recognition;

[0065] S43, color feature extraction: In the color feature extraction subunit, the color feature is extracted by using a color histogram to count the number of pixels in different color intervals. For example, in food sorting, different types of fruits can be distinguished by extracting the red feature of apples and the yellow feature of bananas.

[0066] S5. Target recognition and classification

[0067] S51, template matching: in the template matching subunit, the features of the object to be detected are compared with the pre-stored standard template, which can be pixel-based or feature-based matching, and the matching similarity is measured by calculating indicators such as correlation coefficient or mean square error, which is used for matching chip images with standard templates in electronic component assembly to determine position and direction;

[0068] S52. Machine learning classification: In the machine learning classification subunit, for the case of labeled training data, supervised learning methods are used to train classification models such as support vector machines and neural networks to distinguish defective and non-defective product images in product quality inspection; for the case of unlabeled data, unsupervised learning methods are used;

[0069] S6. Output results: Output the results of target recognition and classification, which can be used for subsequent operation decisions of the robot, generation of quality assessment reports, or displayed to operators on a display device.

[0070] Compared with the related art, the industrial robot visual inspection execution system provided by the present invention has the following beneficial effects:

[0071] The present invention provides an industrial robot visual inspection execution system. An image preprocessing unit, an image segmentation unit, a feature extraction unit and a target recognition and classification unit are respectively arranged in an image processing module in the whole system, which can reduce the amount and complexity of data, improve image quality, separate the target from the background, accurately extract the edge and area of ​​the object, quantify the object features, provide a basis for recognition and classification, accurately recognize the target object and realize automatic object classification.

[0072] Second embodiment

[0073] Please refer to Figure 4Based on the industrial robot visual inspection execution system provided by the first embodiment of the present application, the second embodiment of the present application proposes another industrial robot visual inspection execution system. The second embodiment is only a preferred embodiment of the first embodiment, and the implementation of the second embodiment will not affect the independent implementation of the first embodiment.

[0074] Specifically, the difference of an industrial robot visual inspection execution system provided by the second embodiment of the present application is that, in an industrial robot visual inspection execution system, the data transmission and communication module includes a physical interface unit, a data link protocol unit, a network layer routing unit and a transport layer protocol unit.

[0075] The physical interface unit includes a wired interface subunit and a wireless interface subunit. The wired interface subunit includes an Ethernet interface, an optical fiber interface and an industrial field bus interface. The wireless interface subunit includes a wireless network interface and a Bluetooth interface.

[0076] Ethernet interface: supports Gigabit Ethernet or 10 Gigabit Ethernet interface, cooperates with network protocol for data encapsulation and transmission, is used for large-capacity image data transmission, connects to LAN or Internet, and connects to data link layer protocol unit;

[0077] Industrial fieldbus interface: including Profibus, CAN bus and other interfaces, used for connecting equipment with short distance and high real-time requirements in industrial environments, transmitting control instructions and status information, with a rate of up to 1Mbps and above, strong anti-interference ability, and connected to the data link layer protocol unit;

[0078] Fiber optic interface: used for data transmission over long distances or in scenarios with strong electromagnetic interference. It uses the low loss, high bandwidth, and anti-electromagnetic interference characteristics of optical fiber to achieve long-distance, high-quality data transmission between the visual inspection system and the robot control center, and is connected to the data link layer protocol unit.

[0079] Wireless interface: supports 802.11n, 802.11ac and other standards to achieve wireless data transmission within a certain range, used for flexible arrangement of equipment scenarios, facilitate communication between devices, data sharing and remote monitoring, and connect with the data link layer protocol unit.

[0080] Bluetooth interface module: used for short-distance, small data volume transmission scenarios, such as visual sensor parameter configuration or interaction with handheld terminals. The transmission distance is about 10 meters. The transmission rate varies depending on the Bluetooth version. It is connected to the data link layer protocol unit.

[0081] The data link protocol unit includes an error control subunit and a flow control subunit. The error control subunit includes cyclic redundancy check and forward correction, and the flow control subunit includes window-based flow control and rate adaptation.

[0082] Cyclic redundancy check module: calculates the check code for the transmitted data, and the receiving end compares the check code to detect data transmission errors. It is used for error detection in the transmission of data packets such as image data, and is connected to the transmitting and receiving devices.

[0083] Forward Error Correction Module: Redundant information is added when sending data, so that the receiving end can correct the erroneous data accordingly. It is used in scenarios with high real-time requirements and inconvenient data retransmission, and is connected to the sending and receiving devices.

[0084] Window-based flow control module: coordinates the data transmission rates of the sender and receiver, adjusts the sending rate according to the receiver's confirmation signal and window size, prevents the receiver's buffer from overflowing, and is used in scenarios where the visual system transmits a large amount of image feature data to the robot control system, and is connected to the sender and receiver devices.

[0085] Rate Adaptation Module: Automatically adjusts the transmission rate according to the real-time network bandwidth and quality, optimizes data transmission in a wireless network environment, and connects with the sending and receiving devices.

[0086] The network layer routing unit includes a static routing subunit and a dynamic control subunit.

[0087] Static routing module: In a simple fixed network topology, the data transmission path is determined by pre-configuring the routing table. It is used for data forwarding between subnets within a small factory, reducing network overhead and latency, and is connected to the data link layer protocol unit and the transport layer protocol unit.

[0088] Dynamic routing module: In response to complex and changing network environments, it exchanges information with adjacent routers to automatically adjust the routing table, select the optimal transmission path, ensure communication in the event of network failure, and connect with the data link layer protocol unit and the transport layer protocol unit.

[0089] The transport layer protocol unit includes a TCP protocol subunit and a UDP protocol subunit.

[0090] TCP protocol module: provides reliable, connection-oriented transmission services for transmitting information with high integrity requirements such as image data and complex control instructions. It establishes connections and confirmation and retransmission mechanisms through three-way handshakes to ensure accurate data arrival and connects with the application layer and network layer.

[0091] UDP protocol module: provides connectionless, unreliable but low-latency transmission services, used to transmit information with high real-time requirements but low integrity requirements, such as sensor status signals and robot basic action feedback signals, and connects with the application layer and network layer.

[0092] Process steps:

[0093] S1. Data preparation and transmission interface selection: The visual inspection system or robot control system prepares the data to be transmitted and selects the appropriate physical layer interface based on factors such as data transmission requirements, equipment distance, and electromagnetic environment. If large-capacity image data is transmitted and there is a high-speed and stable network environment, select the Ethernet interface; if it is at a short distance in the industrial field and has high real-time requirements, select the CAN bus and other industrial field bus interfaces; if it is a long distance or strong electromagnetic interference scenario, select the fiber optic interface; if it is wireless and requires flexible equipment layout and medium data volume transmission, select the wireless interface; if it is a short-distance small data volume transmission, select the Bluetooth interface;

[0094] S2. Physical layer data transmission: For wired interfaces, data signals are converted into a form suitable for physical medium transmission and transmitted on the corresponding medium. For wireless interfaces, a wireless connection is established first, and then the data is transmitted wirelessly.

[0095] S3, data link layer processing

[0096] S31, Error control: The CRC module calculates the checksum of the data at the sending end and adds it to the data packet. The receiving end recalculates the checksum after receiving the data. If it is inconsistent with the checksum of the sending end, it is judged that the data is wrong, and retransmission or further error correction can be selected. For situations where real-time requirements are high and retransmission is not suitable, the FEC module adds redundant information when sending data, so that the receiving end can correct the erroneous data to a certain extent;

[0097] S32, flow control: Based on the window flow control module, the sender waits for the receiver to confirm the signal after sending a certain amount of data. If the receiver is slow in processing, the sender is notified to reduce the rate by adjusting the window size to prevent buffer overflow. The rate adaptation module automatically adjusts the transmission rate according to the real-time bandwidth and quality of the network to optimize data transmission in a wireless network environment;

[0098] S4. Network layer routing

[0099] S41, static routing: determine the data transmission path according to the pre-configured routing table, forward the data to the next hop node, and reduce network overhead and transmission delay;

[0100] S42, Dynamic routing: The router exchanges routing information with adjacent routers, dynamically adjusts the routing table according to the real-time status of the network, selects the optimal path to forward data, and ensures that communication is not affected;

[0101] S5, Transport layer protocol processing: If the transmission requires high data integrity, use the TCP protocol. The sender establishes a connection with the receiver through a three-way handshake, and uses the confirmation and retransmission mechanism to ensure that the data arrives accurately during the transmission process. If the transmission requires high real-time performance but relatively low integrity, use the UDP protocol to send data directly to reduce the delay in connection establishment and confirmation.

[0102] S6. Data processing at the receiving end: The receiving end receives data according to the transport layer protocol, performs unpacking and routing processing at the network layer, performs decapsulation and error processing at the data link layer, and then passes the data to the corresponding application or device module for further processing.

[0103] Compared with the related art, the industrial robot visual inspection execution system provided by the present invention has the following beneficial effects:

[0104] The present invention provides an industrial robot visual inspection execution system, in which a physical interface unit, a data link protocol unit, a network layer routing unit and a transport layer protocol unit are respectively arranged in a data transmission and communication module to adapt to different communication media, such as twisted pair, optical fiber, coaxial cable, etc. Different communication media have different characteristics and applicable scenarios. The physical interface unit can select a suitable communication medium according to specific needs and provide corresponding interface support, so that data transmission can be carried out in various environments.

[0105] Third embodiment

[0106] Please refer to Figure 5 Based on the industrial robot visual inspection execution system provided by the first embodiment of the present application, the third embodiment of the present application proposes another industrial robot visual inspection execution system. The third embodiment is only a preferred embodiment of the first embodiment, and the implementation of the third embodiment will not affect the independent implementation of the first embodiment.

[0107] Specifically, the difference of an industrial robot visual inspection execution system provided by the third embodiment of the present application is that, in an industrial robot visual inspection execution system, the robot control module includes a motion control sub-module, a communication sub-module, a task scheduling sub-module and a safety control sub-module, the motion control sub-module includes trajectory planning and joint control, the communication sub-module includes communication with the visual system and communication with external devices, the task scheduling sub-module includes task priority management and task execution sequence planning, and the safety control sub-module includes speed limit and safety area setting and fault detection and recovery strategy.

[0108] Communicating with vision systems

[0109] Data transmission protocol: Establish a stable and efficient data transmission channel, use standard communication protocols to communicate with the visual inspection system, and ensure that the information on the position, size, and shape of the object detected by the visual system can be accurately transmitted to the robot control module;

[0110] Data synchronization mechanism: Realize the synchronization of robot motion and visual inspection data. Since visual inspection and robot motion are performed at different time scales, it is necessary to design a data synchronization mechanism to ensure that the robot receives the latest visual information at the right time.

[0111] Communicating with external devices

[0112] Communication with PLC: Communicate with the programmable logic controller, receive status information from other equipment on the production line, and feed back the robot's working status to the PLC. In the integration of automated production lines, the robot exchanges signals with the PLC to achieve collaborative work with other equipment;

[0113] Communication with the host computer: Communicate with the host computer monitoring system, upload the robot's operating data, and receive task instructions and parameter settings from the host computer. The host computer can centrally monitor and manage multiple robots, adjust the robot's working mode, motion parameters, etc. through the communication interface, and realize remote control and optimized scheduling.

[0114] Task priority management

[0115] Multitasking: In complex production environments, robots may need to handle multiple tasks simultaneously, such as grasping, inspection, and assembly. The task scheduling submodule needs to prioritize tasks based on their urgency and importance.

[0116] Dynamic priority adjustment: Dynamically adjust task priorities based on real-time conditions on the production line.

[0117] Task execution sequence planning

[0118] Process flow arrangement: According to the process flow of product production, reasonably arrange the robot's task execution sequence;

[0119] Conditional judgment and branch execution: Based on the visual detection results and information fed back by other sensors, conditional judgment is performed to determine the robot's next action.

[0120] Hardware security mechanism

[0121] Emergency stop button and safety door switch: Emergency stop buttons and safety door switches are set around the robot's working area. When an emergency occurs (such as personnel entering a dangerous area, abnormal movement of the robot, etc.), the robot can stop moving immediately by pressing the emergency stop button or opening the safety door to ensure the safety of personnel and equipment.

[0122] Limit switches and collision detection sensors: Install limit switches at the extreme positions of the robot's joint motion range to prevent mechanical damage caused by excessive movement of the robot's joints. At the same time, equip it with collision detection sensors. When the robot collides with other objects, it can quickly detect the collision signal and take corresponding safety measures, such as stopping movement and reducing speed.

[0123] Software Security Strategy

[0124] Speed ​​limit and safety zone setting: Set the speed limit of the robot in different working areas in the software.

[0125] Fault detection and recovery strategy: Real-time monitoring of the robot's operating status, timely detection of faults by detecting parameters such as motor current and joint position error. Once a fault occurs, appropriate recovery strategies are adopted according to the type and severity of the fault, such as automatic restart, return to zero position operation, or prompting maintenance personnel to perform repairs.

[0126] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. An industrial robot visual inspection execution system, characterized in that: include: A visual sensor module, wherein a delivery end of the visual sensor module is wirelessly connected to a lighting system module; An image acquisition module, the image acquisition module is wirelessly connected to the transmission end of the lighting system module; An image processing module, the image processing module is wirelessly connected to the transmission end of the image acquisition module, and the image processing module includes an image preprocessing unit, an image segmentation unit, a feature extraction unit, and a target recognition and classification unit; A robot module, wherein the robot is wirelessly connected to the conveying end of the image processing module, the conveying end of the robot module is wirelessly connected to a robot control module, and the conveying end of the robot control module is wirelessly connected to a data transmission and communication module.

2. The industrial robot visual inspection execution system according to claim 1, characterized in that: The image preprocessing module includes a grayscale processing subunit, a noise reduction processing subunit and a contrast enhancement subunit.

3. The industrial robot visual inspection execution system according to claim 1, characterized in that: The image segmentation unit includes a threshold segmentation subunit and an edge detection subunit.

4. The industrial robot visual inspection execution system according to claim 1, characterized in that: The feature extraction unit includes a geometric feature extraction subunit, a texture feature extraction subunit and a color feature extraction subunit.

5. The industrial robot visual inspection execution system according to claim 1, characterized in that: The target recognition and classification unit includes a template matching subunit and a machine learning classification subunit.

6. The industrial robot visual inspection execution system according to claim 1, characterized in that: The data transmission and communication module includes a physical interface unit, a data link protocol unit, a network layer routing unit and a transport layer protocol unit.

7. The industrial robot visual inspection execution system according to claim 6, characterized in that: The physical interface unit includes a wired interface subunit and a wireless interface subunit. The wired interface subunit includes an Ethernet interface, an optical fiber interface and an industrial field bus interface. The wireless interface subunit includes a wireless network interface and a Bluetooth interface.

8. The industrial robot visual inspection execution system according to claim 6, characterized in that: The data link protocol unit includes an error control subunit and a flow control subunit. The error control subunit includes cyclic redundancy check and forward correction, and the flow control subunit includes window-based flow control and rate adaptation.

9. The industrial robot visual inspection execution system according to claim 6, characterized in that: The network layer routing unit includes a static routing subunit and a dynamic control subunit.

10. The industrial robot visual inspection execution system according to claim 6, characterized in that: The transport layer protocol unit includes a TCP protocol subunit and a UDP protocol subunit.

Images