Visual monitoring system and method for an electric arc additive manufacturing installation

Abstract

This invention discloses a visual monitoring system and method for arc additive manufacturing equipment, relating to the field of visual monitoring. In this system, the host computer is used by the operator to calibrate the CCD camera, acquire monitoring data, and provides a unified interface for interaction. The PS terminal receives start commands and CCD camera calibration commands from the host computer, and sends program functions to the host computer to obtain the molten pool width, molten pool center point coordinates, molten pool collapse amount, and collapse direction obtained through CCD camera monitoring. It is also used to store the original molten pool image at a fixed location in the off-chip DDR3 memory. The PL terminal receives control commands from the PS terminal to the IP cores in the PL terminal. The CCD camera is used to acquire the original molten pool image. The LCD screen communicates with the PL terminal through a TFTLCD interface. This invention can improve the real-time performance of the monitoring system and reduce costs.

Description

Technical Field

[0001] This invention relates to the field of visual monitoring, and in particular to a visual monitoring system and method for arc additive manufacturing equipment. Background Technology

[0002] Industrial robots / multi-axis machine tools offer advantages such as high processing efficiency and stable formed parts, and have been widely used in shipbuilding, automotive, metallurgy, and aerospace industries. Arc additive manufacturing also relies on industrial robots / machine tools. In actual arc additive manufacturing, operators initiate a process by either offline teaching (using industrial robots) or generating processing files through path planning (using machine tools). Currently, the most widely used method in the actual forming process of printed parts is visual inspection, where processing defects are identified manually to adjust corresponding parameters. This method is only suitable for situations with a small number of arcs (inspection objects) and is not conducive to the prefabrication of equipment. In arc additive manufacturing processes where multiple arcs are printed in parallel to improve manufacturing efficiency, this method is almost entirely unsuitable.

[0003] To address the aforementioned issues, in the DSP-based hardware implementation of molten pool monitoring, the entire system includes a host computer, a CCD camera, an industrial robot / machine tool, and a DSP acquisition module. The host computer monitors and instructs the hardware modules (DSP, machine tool, industrial robot) and enables human-machine interaction. Specifically, it can be implemented in two forms: one is a PLC + industrial touchscreen, and the other is an industrial control computer. The CCD camera acquires raw images of the molten pool during the arc additive manufacturing process. The industrial robot / machine tool receives motion command signals from the host computer and performs corresponding movements. The DSP acquisition module is a hardware circuit board designed around a DSP chip, connected to the CCD camera via Ethernet protocol and RJ45 interface, acquiring raw molten pool images and performing processing such as image filtering, binarization, and edge detection to obtain key molten pool data. In the GPU-based molten pool monitoring system, the entire system includes a host computer, a CCD camera, an industrial robot / machine tool, and a GPU computing module. The host computer controls the machine tool (industrial robot) operation, schedules the GPU to complete data processing tasks, and enables human-machine interaction. A CCD camera is used to acquire raw images of the molten pool during the arc additive manufacturing process and send them to the host computer. An industrial robot / machine tool receives motion command signals from the host computer and performs the corresponding movements. A GPU computing module, a non-independent hardware unit embedded in the host computer, primarily performs the calculations for the molten pool identification algorithm.

[0004] Most existing image processing solutions for electric arc additive manufacturing typically rely on the GPU built into the computer itself, with the CPU handling all image data scheduling. For additive manufacturing systems with multiple electric arcs (multiple monitoring objects), the CPU of the host computer needs to handle a large number of scheduling tasks, which increases system latency. Furthermore, the GPU itself has relatively high power consumption and cost.

[0005] Therefore, it is urgent to design dedicated hardware circuits to achieve high-speed image transmission and rapid calculation of various processing algorithms during the molten pool monitoring process, so as to improve the real-time performance of the monitoring system and reduce costs. Summary of the Invention

[0006] The purpose of this invention is to provide a visual monitoring system and method for arc additive manufacturing equipment, which can improve the real-time performance of the monitoring system and reduce costs.

[0007] To achieve the above objectives, the present invention provides the following solution:

[0008] A visual monitoring system for arc additive manufacturing equipment includes: a host computer, a PS terminal, a PL terminal, an LCD screen, an external DDR3 memory, and a CCD camera; the PS terminal and the PL terminal together form a ZYNQ chip;

[0009] The host computer is used by operators to calibrate CCD cameras, acquire monitoring data, and provide a unified interface for interaction.

[0010] The PS terminal communicates with the host computer, the external memory DDR3, and the CCD camera. The PS terminal is used to receive the start command and CCD camera calibration command sent by the host computer, and to send the program function to the host computer the molten pool width, molten pool center point coordinates, molten pool collapse amount and collapse direction obtained by the CCD camera. It is also used to store the original molten pool image at a fixed location in the external memory DDR3.

[0011] The PL terminal communicates with the PS terminal, and the PL terminal receives control commands from the PS terminal to the IP core in the PL terminal.

[0012] The CCD camera is used to acquire raw molten pool images;

[0013] The LCD screen communicates with the PL terminal via a TFTLCD interface.

[0014] Optionally, the host computer includes: a human-computer interaction interface and a driver program;

[0015] The human-machine interface is used by the operator to calibrate the CCD camera and acquire monitoring data;

[0016] The driver program is used to provide a unified interface for the interactive interface.

[0017] Optionally, the PS terminal is an ARM Cortex-A7.

[0018] Optionally, the PS terminal is connected to the CCD camera via a PHY chip and an RJ45 interface, conforming to the GigEVision protocol.

[0019] Optionally, the PS terminal communicates with the off-chip DDR3 memory via the AXI4 bus.

[0020] Optionally, the PS end includes: ROI capture program, GVSP program, GVCP program, calculation result output program, LCD configuration reading program, clock configuration program, VDMA configuration program, and frame interval N configuration program;

[0021] The ROI cropping program is used to crop the region of interest from the acquired raw melt pool image;

[0022] The GVSP program is used to acquire the raw molten pool image sent from the CCD camera to the PS terminal;

[0023] The GVCP program is used by the PS terminal to send control commands to the CCD camera to calibrate the CCD camera, adjust the exposure, and receive the original molten pool image; it is also used to receive control feedback signals from the CCD camera to the PS terminal.

[0024] The calculation result output program is used to obtain information on melt width, melt pool center point coordinates, melt pool collapse direction and melt pool collapse amount obtained from the PL end by scanning the GP interface;

[0025] The LCD configuration reading program and clock configuration program are used to adapt to LCD screens of different resolutions and configure the corresponding screen display clock.

[0026] The VDMA configuration program is used to dynamically configure the VDMA module on the PL end according to the image resolution supported by the LCD screen to adapt to molten pool images of different resolutions.

[0027] The frame interval N configuration program is used to receive the frame interval N setting from the host computer and send it to the PL terminal via AXIGPIO2IP.

[0028] Optionally, the PL terminal communicates with the PS terminal via the on-chip bus AXI4.

[0029] Optionally, the PL terminal includes: AXI-SmartConnectIP / VDMA, Stream Copy IP, Gaussian Filter IP, Binarization and Melt Width Detection IP, Melt Width and Melt Pool Center Point Recording IP, Melt Pool Defect Detection IP / AXIGPIO2IP, AXI-StreamtoVideoOutIP / rgb2lcdIP, VTCIP, AXIdynclkIP, and AXI-InterConnect / AXIGPIO1IP;

[0030] AXI-SmartConnectIP / VDMA is used to convert raw melt pool images in AXI4 protocol format read from off-chip DDR3 memory into AXI-Stream format;

[0031] The Stream Copy IP is used to copy an AXI-Stream format image data stream, which is then passed to the AXI-StreamtoVideoOut IP and the Gaussian Filter IP for image display and data processing, respectively.

[0032] The Gaussian filter IP is used to convert the molten pool image format from RGB to grayscale and to perform Gaussian filtering on the molten pool image to remove noise caused by arc interference in the image;

[0033] The binarization and weld width detection IP is used to binarize the image pixels and record the weld width data and weld center point coordinates of the current weld pool image, which are then passed to the weld pool defect detection IP and the weld width and weld pool center point recording IP, respectively.

[0034] The melt width and melt pool center point recording IP is used to record the melt width and melt pool center point coordinates of the previous frame image, which are then read by the melt pool defect detection IP.

[0035] The molten pool defect detection IP / AXIGPIO2IP is used to read the current molten pool width and molten pool center point coordinates from the binarization and molten pool width detection IP, and to read the molten pool width and molten pool center point coordinates of the previous frame image from the molten pool width and molten pool center point recording IP. After subtracting the corresponding values, the IP is compared with a set threshold to calculate the molten pool collapse direction and collapse amount. It is also used to receive the frame interval number N sent from the PS end through AXIGPIO2IP. It is also used to record the molten pool collapse direction and collapse amount of the last N frames of images, compare the molten pool collapse direction of the last N frames of images to calculate the molten pool collapse amount of the image, and then determine whether the molten pool status is normal and store the judgment for the PS end to read.

[0036] AXI-StreamtoVideoOutIP / rgb2lcdIP is used to convert the AXI-Stream protocol to a video streaming protocol. The converted melt pool image will be displayed on the LCD screen.

[0037] VTCIP and AXIdynclkIP are used to receive a specified clock frequency sent by the PS end and configure the AXI-StreamtoVideoOutIP and rgb2lcdIP clocks to adapt to LCD screens of different resolutions.

[0038] The AXI-InterConnect / AXIGPIO1IP is used to scan the LCD screen resolution and transmit the data to the PS (Power Supply) terminal, which then determines the clock frequency that needs to be configured for the LCD screen.

[0039] A visual monitoring method for arc additive manufacturing equipment, applied to a visual monitoring system for the aforementioned arc additive manufacturing equipment, the visual monitoring method comprising:

[0040] The exposure value of the CCD camera, the image acquisition frequency f, and the interval frame number N for molten pool defect detection are set using the host computer human-computer interaction interface;

[0041] Configure the IP core according to the LCD screen resolution;

[0042] The acquired raw molten pool image is sequentially processed by image data copying, Gaussian filtering, binarization, molten width, and molten pool coordinate storage;

[0043] Determine whether the difference between the coordinates of the center point of the molten pool in two consecutive frames is greater than the acceptable value;

[0044] If the value is greater than 1, it is determined that the melt pool collapsed during the two-frame interval, and the monitoring result is output to the host computer via the PS terminal.

[0045] If the value is less than or equal to the value, it is determined whether the cumulative offset of the center point of the molten pool in the last N frames of images is greater than the acceptable value. If it is greater, it indicates that the molten pool has collapsed within the minimum resolution time N / f of the molten pool accumulated defect and the monitoring result is output to the host computer through the PS terminal, and the next frame of image processing is performed. If the value is less than or equal to the value, it is determined that the molten pool is in normal condition.

[0046] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0047] This invention provides a visual monitoring system and method for arc additive manufacturing equipment. For arc additive manufacturing equipment, especially those involving multiple arcs (multiple detection objects), high-speed transmission and processing of image data is crucial. A dedicated hardware circuit, independent of the host computer, is built using a host computer, PS terminal, PL terminal, LCD screen, external DDR3 memory, and CCD camera to accelerate the processing of molten pool images. Compared to existing GPU-based algorithm processing systems, this circuit is designed and programmed entirely according to the algorithm and system control function requirements. Image data from each detection object does not need to be scheduled by the CPU, significantly improving system data transmission and processing speed. Furthermore, during the processor's algorithm calculations, addition and subtraction operations typically require several orders of magnitude less time than multiplication and division operations, requiring fewer system resources. In the molten pool detection design of this system, a defect assessment method is adopted by comparing the difference between consecutive frames and a threshold, avoiding large-scale matrix operations (multiplication) on image data. This greatly reduces algorithm processing latency and saves system resource overhead. Attached Figure Description

[0048] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 A schematic diagram of the structure of a visual monitoring system for an arc additive manufacturing equipment provided by the present invention;

[0050] Figure 2 This is a schematic diagram of the PS terminal and the off-chip DDR3 memory structure;

[0051] Figure 3 This is a schematic diagram of the melt pool detection algorithm and LCD screen structure in the PL terminal;

[0052] Figure 4 This is a schematic diagram of a visual monitoring method for an arc additive manufacturing equipment provided by the present invention. Detailed Implementation

[0053] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0054] The purpose of this invention is to provide a visual monitoring system and method for arc additive manufacturing equipment, which can improve the real-time performance of the monitoring system and reduce costs.

[0055] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0056] like Figure 1 As shown, the visual monitoring system for arc additive manufacturing equipment provided by the present invention includes: a host computer, a PS terminal, a PL terminal, an LCD screen, an external DDR3 memory, and a CCD camera; the PS terminal and the PL terminal together form a ZYNQ chip;

[0057] The host computer is used by operators to calibrate CCD cameras, acquire monitoring data, and provide a unified interface for interaction. The host computer includes a human-machine interface and driver programs. The unified interface is written in C language.

[0058] The PS terminal communicates with the host computer, the external memory DDR3, and the CCD camera. The PS terminal is used to receive the start command and CCD camera calibration command sent by the host computer, and to send the program function to the host computer the molten pool width, molten pool center point coordinates, molten pool collapse amount and collapse direction obtained by the CCD camera. It is also used to store the original molten pool image at a fixed location (0x0F00000-0xF4E200) in the external memory DDR3.

[0059] like Figure 2 As shown, the PS terminal is a CPU, specifically an ARM Cortex-A7.

[0060] The PS terminal connects to the CCD camera via a PHY chip and an RJ45 interface, conforming to the GigEVision protocol.

[0061] The PS terminal communicates with the off-chip DDR3 memory via the AXI4 bus.

[0062] The PS terminal includes: ROI capture program, GVSP program, GVCP program, calculation result output program, LCD configuration reading program, clock configuration program, VDMA configuration program, and frame interval N configuration program.

[0063] The ROI (Region of Interest) cropping program is used to crop the acquired raw melt pool image; it converts the acquired raw melt pool image (480x800) into (200x200) and stores it in a specified location on off-chip DDR3 memory. Its purpose is to remove pixels outside the key areas of the melt pool image to compress the data processing scale.

[0064] The GVSP (GigEVision Data Protocol) program is used to acquire the raw molten pool image sent from the CCD camera to the PS.

[0065] The GVCP (GigEVision Control Protocol) program is used by the PS terminal to send control commands to the CCD camera to calibrate the CCD camera, adjust the exposure, and receive the raw molten pool image; it is also used to receive control feedback signals from the CCD camera to the PS terminal.

[0066] The calculation result output program is used to obtain information on melt width, melt pool center point coordinates, melt pool collapse direction and melt pool collapse amount obtained from the PL end by scanning the GP interface.

[0067] The LCD configuration reading program and clock configuration program are used to adapt to LCD screens of different resolutions and configure the corresponding screen display clock.

[0068] The VDMA configuration program is used to dynamically configure the VDMA module on the PL end according to the image resolution supported by the LCD screen to adapt to molten pool images of different resolutions.

[0069] The frame interval N configuration program is used to receive the frame interval N setting from the host computer and send it to the PL terminal via AXIGPIO2IP.

[0070] Communication with the CCD camera based on GVCP and GVSP protocols was implemented in ARM.

[0071] The PL terminal communicates with the PS terminal, and the PL terminal receives control commands from the PS terminal to the IP core in the PL terminal.

[0072] The PL terminal is an FPGA. For example... Figure 3 As shown, the PL terminal is connected to the PS terminal via the on-chip AXI4 bus, following the AXI4 protocol, and is used to receive control commands from the PS terminal to the circuit modules (IP cores) on the PL terminal. The PL terminal is interconnected with the external DDR3 memory via the HP interface of the PS terminal. Their communication follows the AXI4 protocol, with the PL terminal actively reading each frame of image data transmitted from the PS terminal to the external DDR3 memory and performing subsequent processing. The PL terminal is connected to the LCD screen via the TFTLCD interface for transmitting image data to the LCD screen.

[0073] The PL terminal includes: AXI-SmartConnectIP / VDMA, Stream Copy IP, Gaussian Filter IP, Binarization and Melt Width Detection IP, Melt Width and Melt Pool Center Point Recording IP, Melt Pool Defect Detection IP / AXIGPIO2IP, AXI-StreamtoVideoOutIP / rgb2lcdIP, VTC IP, AXIdynclkIP, and AXI-InterConnect / AXIGPIO1IP.

[0074] AXI-SmartConnectIP / VDMA is used to convert raw melt pool images in AXI4 protocol format read from off-chip DDR3 memory into AXI-Stream format.

[0075] The Stream Copy IP is used to copy an AXI-Stream format image data stream, which is then passed to the AXI-StreamtoVideoOut IP and the Gaussian Filter IP for image display and data processing, respectively.

[0076] The Gaussian filter IP is used to convert the molten pool image format from RGB to grayscale and to perform Gaussian filtering on the molten pool image to remove noise caused by arc interference.

[0077] The binarization and weld width detection IP is used to binarize the image pixels and record the weld width data and weld center point coordinates of the current weld pool image, which are then passed to the weld pool defect detection IP and the weld width and weld pool center point recording IP, respectively.

[0078] The melt width and melt pool center point recording IP is used to record the melt width and melt pool center point coordinates of the previous frame image for the melt pool defect detection IP to read.

[0079] The molten pool defect detection IP / AXIGPIO2IP is used to read the current molten pool width and molten pool center point coordinates from the binarization and molten pool width detection IP, and to read the molten pool width and molten pool center point coordinates from the previous frame image from the molten pool width and molten pool center point recording IP. After subtracting the corresponding values, the IP is compared with a set threshold to calculate the molten pool collapse direction and collapse amount. It is also used to receive the frame interval number N sent from the PS end via AXIGPIO2IP. Furthermore, it is used to record the molten pool collapse direction and collapse amount of the last N frames of images, compare the molten pool collapse direction of the last N frames of images to calculate the molten pool collapse amount of the image, and then determine whether the molten pool status is normal and store the determination for the PS end to read.

[0080] AXI-StreamtoVideoOutIP / rgb2lcdIP is used to convert the AXI-Stream protocol to a video streaming protocol, and the converted melt pool image will be displayed on the LCD screen.

[0081] VTCIP and AXIdynclkIP are used to receive a specified clock frequency sent by the PS end and configure the clocks of AXI-StreamtoVideoOutIP and rgb2lcdIP to adapt to LCD screens of different resolutions.

[0082] The AXI-InterConnect / AXIGPIO1IP is used to scan the LCD screen resolution and transmit the data to the PS (Power Supply) terminal, which then determines the clock frequency that needs to be configured for the LCD screen.

[0083] A dedicated module for molten pool width detection and binarization was designed in the FPGA. Image data is transmitted pixel by pixel in a streaming manner. There is no intermediate buffering stage in data processing, which saves buffering and time overhead.

[0084] A dedicated melt pool contour defect detection module was designed in the FPGA. By calculating the difference between the melt pool center point positions of two consecutive frames of image data, the offset of the melt pool center point is obtained. By accumulating and recording the melt pool center point offset of the last X frames of images, it is inferred whether the melt pool has collapsed, its direction, and the amount of collapse.

[0085] The CCD camera is used to acquire raw images of the molten pool.

[0086] The LCD screen communicates with the PL terminal via a TFTLCD interface.

[0087] The off-chip DDR3 memory is a separate memory chip from the ZYNQ chip. It is connected to both the PS and PL ends via the AXI bus and is mainly used to cache the number of frames of melt pool images for each frame.

[0088] like Figure 4 As shown, the present invention also provides a visual monitoring method for arc additive manufacturing equipment, applied to the aforementioned visual monitoring system for arc additive manufacturing equipment, the visual monitoring method comprising:

[0089] Initialize the visual monitoring system of an electric arc additive manufacturing equipment: complete the startup of the host computer program, the startup of the CCD camera, the initialization of the PS terminal, and the initialization of each circuit module (IP core) of the PL terminal.

[0090] The exposure value of the CCD camera, the image acquisition frequency f, and the interval frame number N for molten pool defect detection are set using the host computer human-computer interaction interface.

[0091] Configure the IP core according to the LCD screen resolution.

[0092] The acquired raw molten pool image is sequentially processed by image data copying, Gaussian filtering, binarization, molten width calculation, and molten pool coordinate storage.

[0093] Determine whether the difference between the coordinates of the center point of the molten pool in two consecutive frames is greater than the acceptable value;

[0094] If the value is greater than the threshold, it is determined that the melt pool collapsed during the two-frame interval, and the monitoring result is output to the host computer via the PS terminal.

[0095] If the value is less than or equal to the value, it is determined whether the cumulative offset of the center point of the molten pool in the last N frames of images is greater than the acceptable value. If it is greater, it indicates that the molten pool has collapsed within the minimum resolution time N / f of the molten pool accumulated defect and the monitoring result is output to the host computer through the PS terminal, and the next frame of image processing is performed. If the value is less than or equal to the value, it is determined that the molten pool is in normal condition.

[0096] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0097] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A visual monitoring system for arc additive manufacturing equipment, characterized in that, include: The ZYNQ chip consists of a host computer, a PS terminal, a PL terminal, an LCD screen, external DDR3 memory, and a CCD camera; the PS terminal and the PL terminal together form the ZYNQ chip. The host computer is used by operators to calibrate CCD cameras, acquire monitoring data, and provide a unified interface for interaction. The PS terminal communicates with the host computer, the external memory DDR3, and the CCD camera. The PS terminal is used to receive the start command and CCD camera calibration command sent by the host computer, and to send the program function to the host computer the molten pool width, molten pool center point coordinates, molten pool collapse amount and collapse direction obtained by the CCD camera. It is also used to store the original molten pool image at a fixed location in the external memory DDR3. The PL terminal communicates with the PS terminal, and the PL terminal receives control commands from the PS terminal to the IP core in the PL terminal. The CCD camera is used to acquire raw molten pool images; The LCD screen communicates with the PL terminal via a TFTLCD interface; The PL terminal includes: AXI-SmartConnect IP / VDMA, Stream Copy IP, Gaussian Filter IP, Binarization and Melt Width Detection IP, Melt Width and Melt Pool Center Point Recording IP, Melt Pool Defect Detection IP / AXI GPIO2 IP, AXI-Stream to VideoOut IP / rgb2lcd IP, VTC IP, AXI dynclk IP, and AXI-InterConnect / AXI GPIO1 IP; AXI-SmartConnect IP / VDMA is used to convert raw melt pool images in AXI4 protocol format read from off-chip DDR3 memory into AXI-Stream format; The stream copying IP is used to copy an AXI-Stream format image data stream, which is then passed to the AXI-Stream toVideo Out IP and the Gaussian filter IP for image display and data processing, respectively. The Gaussian filter IP is used to convert the molten pool image format from RGB to grayscale and to perform Gaussian filtering on the molten pool image to remove noise caused by arc interference in the image; The binarization and weld width detection IP is used to binarize the image pixels and record the weld width data and weld center point coordinates of the current weld pool image, which are then passed to the weld pool defect detection IP and the weld width and weld pool center point recording IP, respectively. The melt width and melt pool center point recording IP is used to record the melt width and melt pool center point coordinates of the previous frame image, which are then read by the melt pool defect detection IP. The molten pool defect detection IP / AXI GPIO2 IP is used to read the current molten pool width and molten pool center point coordinates from the binarization and molten pool width detection IP. It also reads the molten pool width and molten pool center point coordinates from the previous frame image from the molten pool width and molten pool center point recording IP, calculates the difference, and compares it with a set threshold to calculate the molten pool collapse direction and collapse amount. It is also used to receive the frame interval number N sent from the PS end through the AXI GPIO2 IP. Furthermore, it is used to record the molten pool collapse direction and collapse amount of the last N frames of images, compare the molten pool collapse direction of the last N frames of images to calculate the molten pool collapse amount of the image, and then determine whether the molten pool status is normal or not, store the determination, and make it available for the PS end to read. The AXI-Stream to Video Out IP / rgb2lcd IP is used to convert the AXI-Stream protocol to a video streaming protocol. The converted melt pool image will be displayed on the LCD screen. VTC IP and AXI dynclk IP are used to receive a specified clock frequency sent by the PS end and to configure the clock of AXI-Stream toVideo Out IP and rgb2lcd IP to adapt to LCD screens of different resolutions; The AXI-InterConnect / AXI GPIO1 IP is used to scan the LCD screen resolution and transmit the data to the PS (Power Supply) terminal, which then determines the clock frequency that needs to be configured for the LCD screen.

2. The visual monitoring system for an arc additive manufacturing equipment according to claim 1, characterized in that, The host computer includes: a human-computer interaction interface and a driver program; The human-machine interface is used by the operator to calibrate the CCD camera and acquire monitoring data; The driver program is used to provide a unified interface for the interactive interface.

3. The visual monitoring system for an arc additive manufacturing equipment according to claim 1, characterized in that, The PS terminal is an ARM Cortex-A7.

4. The visual monitoring system for an arc additive manufacturing equipment according to claim 1, characterized in that, The PS terminal is connected to the CCD camera via a PHY chip and an RJ45 interface, conforming to the GigEVision protocol.

5. The visual monitoring system for an arc additive manufacturing equipment according to claim 1, characterized in that, The PS terminal communicates with the off-chip DDR3 memory via the AXI4 bus.

6. The visual monitoring system for an arc additive manufacturing equipment according to claim 1, characterized in that, The PS terminal includes: ROI capture program, GVSP program, GVCP program, calculation result output program, LCD configuration reading program, clock configuration program, VDMA configuration program, and frame interval N configuration program. The ROI cropping program is used to crop the region of interest from the acquired raw melt pool image; The GVSP program is used to acquire the raw molten pool image sent from the CCD camera to the PS terminal; The GVCP program is used by the PS terminal to send control commands to the CCD camera to calibrate the CCD camera, adjust the exposure, and receive the original molten pool image; it is also used to receive control feedback signals from the CCD camera to the PS terminal. The calculation result output program is used to obtain information on melt width, melt pool center point coordinates, melt pool collapse direction and melt pool collapse amount obtained from the PL end by scanning the GP interface; The LCD configuration reading program and clock configuration program are used to adapt to LCD screens of different resolutions and configure the corresponding screen display clock. The VDMA configuration program is used to dynamically configure the VDMA module on the PL end according to the image resolution supported by the LCD screen to adapt to molten pool images of different resolutions. The frame interval N configuration program is used to receive the frame interval N setting from the host computer and send it to the PL end via AXI GPIO2 IP.

7. The visual monitoring system for an arc additive manufacturing equipment according to claim 1, characterized in that, The PL terminal communicates with the PS terminal via the on-chip bus AXI4.

8. A visual monitoring method for arc additive manufacturing equipment, applied to the visual monitoring system of the arc additive manufacturing equipment as described in any one of claims 1-7, characterized in that, The visual monitoring method includes: The exposure value of the CCD camera, the image acquisition frequency f, and the interval frame number N for molten pool defect detection are set using the host computer human-computer interaction interface; Configure the IP core according to the LCD screen resolution; The acquired raw molten pool image is sequentially processed by image data copying, Gaussian filtering, binarization, molten width, and molten pool coordinate storage; Determine whether the difference between the coordinates of the center point of the molten pool in two consecutive frames is greater than the acceptable value; If the value is greater than 1, it is determined that the melt pool collapsed during the two-frame interval, and the monitoring result is output to the host computer via the PS terminal. If the value is less than or equal to the value, it is determined whether the cumulative offset of the center point of the molten pool in the last N frames of images is greater than the acceptable value. If it is greater, it indicates that the molten pool has collapsed within the minimum resolution time N / f of the molten pool accumulated defect and the monitoring result is output to the host computer through the PS terminal, and the next frame of image processing is performed. If the value is less than or equal to the value, it is determined that the molten pool is in normal condition.

Images