System for identifying specific video output area of industrial control computer

Abstract

The invention relates to the technical field of industrial control, in particular to a system for identifying a specific video output area of an industrial control computer, which comprises the following steps: responding to an area identification instruction received through a network interface based on a video frame to be identified, preprocessing area data used for identification and analysis are obtained; carrying out identification analysis on the preprocessed area data and carrying out structured processing to obtain a standardized identification result; and taking the identification result as a response to the area identification instruction, and returning the identification result to an application system through the network interface. According to the invention, by identifying the video signal of the industrial control computer, the designated area is accurately identified while real-time display is ensured, and the standardized result is returned to the application system, so that automatic intelligent monitoring is realized, the efficiency is effectively improved, and the labor cost is reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of industrial control technology, in particular to a system for identifying specific regions of video output of an industrial computer. BACKGROUND

[0002] In an industrial production environment, industrial control computers (industrial computers) are widely used to control production equipment and monitor processes. These industrial computers are usually equipped with standard video output interfaces such as VGA, HDMI, and DVI, which output real-time data such as device operating status, production parameters, and alarm information to local display devices for observation and monitoring by on-site operators.

[0003] Currently, the monitoring of such video output information mainly relies on continuous human observation, which not only requires a large amount of manpower, but also may fail to detect abnormalities in a timely manner due to operator fatigue or negligence, making it difficult to effectively guarantee production efficiency and safety. With the development of industrial internet and intelligentization, some new production equipment has begun to integrate built-in monitoring and alarm functions. However, such integrated solutions have the following disadvantages: first, they occupy hardware (such as CPU and memory) and software resources that are used for production tasks, which may affect the real-time performance and stability of the production system in resource-constrained or high-load scenarios; second, for many expensive, stable, but old production equipment, their hardware architecture and software system have been finalized, making it difficult to directly modify or upgrade the system, and if they are discarded due to the inability to implement intelligent monitoring, it will cause economic losses to the enterprise. SUMMARY

[0004] The technical problem to be solved by the present application is to overcome the shortcomings of the prior art and provide a system for identifying specific regions of video output of an industrial computer. By identifying the video signal of the industrial computer, the specified region is accurately identified while ensuring real-time display, and the standardized result is returned to the application system, achieving automatic and intelligent monitoring and effectively improving efficiency and reducing labor costs.

[0005] To solve the above technical problems, the basic idea of the technical solution of the present application is as follows: In a first aspect, a system for identifying specific regions of video output of an industrial computer includes: An input / output module for receiving an original video signal from an industrial computer; splitting the original video signal into a video stream for driving a display device for real-time display and a video stream for image recognition, and grabbing a video frame to be identified from the video stream for image recognition; A region identification module is configured to, based on the video frame to be identified, respond to a region identification instruction received through a network interface, wherein the instruction carries at least one key coordinate region information, to obtain preprocessed region data for identification analysis; perform identification analysis on the preprocessed region data and structured processing to obtain a standardized identification result; and return the identification result to an application system through the network interface as a response to the region identification instruction.

[0006] In a second aspect, a control method of a system for identifying a specific region of a video output by an industrial computer comprises the following steps: receiving an original video signal from the industrial computer; splitting the original video signal into a video stream for driving a display device to perform real-time display and a video stream for image identification, and capturing a video frame to be identified from the video stream for image identification; based on the video frame to be identified, responding to a region identification instruction received through a network interface, wherein the instruction carries at least one key coordinate region information, to obtain preprocessed region data for identification analysis; performing identification analysis on the preprocessed region data and structured processing to obtain a standardized identification result; returning the identification result to an application system through the network interface as a response to the region identification instruction.

[0007] After the above technical solution is adopted, the present application has the following beneficial effects compared with the prior art.

[0008] The original video signal output by the industrial computer is received, and the original information of the video data, including pixel details, timing characteristics and clock control signals, can be completely retained, thereby providing undistorted and lossless basic data for subsequent video stream processing for different purposes; the original video signal is split into two independent video streams, which can realize parallel processing of two types of data requirements, real-time display and image recognition, without interference; the video stream used to drive the display device can directly support on-site observation, and the video stream used for image recognition can focus on the recognition task, avoiding processing delay caused by single data link occupation; the video frames to be recognized are captured from the image recognition video stream, which can accurately locate the data objects required for recognition and eliminate redundant frame data that does not need to be recognized, thereby improving the data processing efficiency; in response to the recognition instruction carrying the key coordinate region information, the pre-processing region data is extracted from the video frames to be recognized, which can accurately focus the data processing range and only extract data from the region corresponding to the key coordinates, thereby avoiding indiscriminate processing of the entire video frame; after the pre-processing region data is recognized and analyzed, the recognition information is converted into standardized results through structured processing, which can unify the data format of the results, such as fixed fields of identification region number, identification content, identification timestamp, etc., so that the results have clear logical levels; the standardized format can directly adapt to the reading and analysis requirements of the application system, reducing the format adaptation cost during data connection, while improving the reusability of the recognition results; the standardized recognition results are returned to the application system through the network interface, which can rely on the stability and efficiency of network transmission to ensure that the recognition results respond to the region recognition instruction in time and quickly reach the application system; the state feedback of network transmission confirms whether the recognition results are successfully delivered, thereby ensuring the integrity of the instruction and response; at the same time, the transmission mode of the network interface is adapted to the application system with network function, which reduces the complexity of data interaction between systems and supports the application system to quickly carry out subsequent device monitoring, production scheduling and other operations based on the recognition results.

[0009] The specific embodiments of the present application will be further described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0010] The drawings described herein are intended to provide further understanding of the present application, and form a part of the present application. The schematic embodiments of the present application and their descriptions serve to explain the present application, and do not constitute an improper limitation on the present application. Some specific embodiments of the present application will be described in detail below with reference to the drawings in an exemplary and non-limiting manner. The same reference signs in the drawings indicate the same or similar components or parts, and those skilled in the art should understand that the drawings are not necessarily drawn to scale, and in the drawings: Figure 1 is a system schematic diagram of the present application for recognizing a specific region of the video output of an industrial computer.

[0011] Figure 2This is a schematic diagram of the control method of the system for identifying specific areas of video output from an industrial control computer according to the present invention.

[0012] Figure 3 This is a schematic diagram of the system structure and signal flow of the present invention.

[0013] It should be noted that these accompanying drawings and textual descriptions are not intended to limit the scope of the invention in any way, but rather to illustrate the concept of the invention to those skilled in the art by referring to specific embodiments. The elements in the drawings are schematic and not drawn to scale. Detailed Implementation

[0014] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present application without creative effort should fall within the scope of protection of the present application.

[0015] The following embodiments of this application use a system for identifying specific areas of video output from an industrial control computer as an example to illustrate the solution of this application. However, these embodiments do not limit the scope of protection of this application.

[0016] like Figure 1 As shown, the present invention provides a system for identifying specific areas of video output from an industrial control computer, comprising: The input / output module 11 is used to receive the raw video signal from the industrial control computer; split the raw video signal into a video stream for driving the display device for real-time display and a video stream for image recognition; and capture the video frame to be recognized from the video stream for image recognition. The region identification module 12 is used to, based on the video frame to be identified, respond to a region identification instruction received through a network interface, wherein the instruction carries at least one key coordinate region information, to obtain preprocessed region data for identification analysis; perform identification analysis and structured processing on the preprocessed region data to obtain standardized identification results; and return the identification results as a response to the region identification instruction to the application system through the network interface.

[0017] In this embodiment of the invention, receiving the raw video signal from the industrial control computer allows for the complete retention of the raw video data, providing undistorted basic data support for subsequent video stream processing for different purposes and ensuring the reliability of the raw data in subsequent data processing stages. Dividing the raw video signal into a real-time display video stream for driving the display device and a video stream for image recognition enables the parallel processing of these two types of data, avoiding mutual interference between different data uses and ensuring the smoothness of real-time display and the data independence of image recognition. Extracting the video frame to be recognized from the video stream for image recognition allows for the precise location of the data object required for image recognition, reducing irrelevant video data from entering subsequent recognition stages, lowering the data processing load of the image recognition stage, and improving the data processing efficiency of image recognition.

[0018] Based on the video frame to be identified, a region identification command carrying at least one key coordinate region information is invoked to obtain preprocessed region data. This enables precise focusing of the data processing range, extracting data only from the key coordinate region, reducing interference from non-target region data in the identification and analysis process, and improving the targeting of data processing. The preprocessed region data is then identified, analyzed, and structured to obtain standardized identification results. This transforms the identified information into a unified and standardized data format, facilitating direct reading, parsing, and use by subsequent application systems, reducing the complexity of data adaptation during application system integration. The identification results are returned to the application system via a network interface, enabling efficient transmission of the results, ensuring timely data response, and guaranteeing that the application system can quickly obtain the required identification data, supporting the smooth operation of subsequent related processes.

[0019] In the system for identifying a specific area of ​​video output from an industrial control computer according to an embodiment of the present invention, the input / output module 11 receives the original video signal from the industrial control computer; splits the original video signal into a video stream for driving a display device for real-time display and a video stream for image recognition, and captures the video frame to be recognized from the video stream for image recognition, including: Step 111: Directly connect to and receive the raw video signal from the video output interface of the industrial control computer via the video input interface; wherein, the physical specifications and electrical characteristics of the video input interface are configured to match the video output interface of the industrial control computer, and the video output interface is any one of HDMI, VGA, or DVI interface, specifically including: first determining the type of video output interface currently used by the industrial control computer, i.e., determining its specific interface type from HDMI, VGA, and DVI interfaces; based on the determined interface type, retrieving the corresponding physical specification parameters and electrical characteristic parameters of that type of interface, wherein the physical specification parameters include the number of interface pins, pin arrangement, and interface dimensions, and the electrical characteristic parameters include the signal transmission voltage range and signal transmission rate. First, the signal impedance matching value is set. Then, the video input interface is configured to match its physical specifications with those of the industrial control computer's video output interface. Simultaneously, the electrical characteristics of the input interface are calibrated to match those of the industrial control computer's video output interface, ensuring that both meet the requirements for physical compatibility and electrical compatibility in signal transmission. After configuration, the video input interface is physically connected to the industrial control computer's video output interface to establish a stable signal transmission link. Once established, the input interface's signal receiving unit receives the original video signal output from the industrial control computer's video output interface in real time. During reception, the stability of the signal transmission is continuously monitored to ensure that the original video signal is transmitted without distortion or packet loss.

[0020] Step 112 involves copying the original video signal to generate a first video signal stream and a second video signal stream. Specifically, this includes: firstly, based on the original video signal, initiating the signal preprocessing stage of the synchronous copying algorithm; secondly, using the algorithm's built-in signal parsing submodule to extract features from the original video signal frame by frame and component by component; specifically, extracting pixel data information, line synchronization signals, field synchronization signals, and clock control signals from the original video signal. The pixel data information includes the brightness and color values ​​of each pixel. The line synchronization signal is used to identify the start and end timestamps of each line of pixel data. The field synchronization signal is used to identify the start and end timestamps of each frame of image. The clock control signal is used as a reference signal to regulate the signal transmission rate. Furthermore, the extracted signal components are time-stamped to determine the timestamp information of each component in the original signal transmission link, laying the data foundation for the timing synchronization of the subsequent copying process.

[0021] Next, the timing reference establishment stage of the synchronous copying algorithm is entered. The algorithm uses the clock control signal in the original video signal as the core reference, locks the frequency parameters (such as Hz value) and timing period of the signal, and generates two completely consistent timing synchronization reference signals based on this. The reference signal must be completely matched with the frequency and phase of the original clock control signal to ensure that the transmission rhythm of the subsequent two copied signals is synchronized with the original signal. At the same time, the horizontal synchronization signal and vertical synchronization signal extracted in the previous step are associated and bound with the generated timing synchronization reference signal, so that the triggering time of the horizontal and vertical synchronization signals can accurately correspond to the timing node of the reference signal, avoiding timing misalignment during the copying process.

[0022] Subsequently, synchronous copying and signal distribution operations are performed. The algorithm activates the dual-channel signal generation submodule, which synchronously distributes the time-marked pixel data, line synchronization signal, and field synchronization signal to two independent signal processing channels according to the rhythm of the established timing synchronization reference signal. During the distribution process, the algorithm needs to monitor the transmission link of each signal in real time to ensure that every pixel data point and every synchronization signal trigger command can be transmitted simultaneously in both channels, and that there is no data loss or out-of-order transmission. Specifically, when a frame of pixel data in the original signal begins to be transmitted, the dual-channel signal generation submodule will simultaneously copy the entire frame of pixel data and write it into the temporary buffers of the two channels respectively, while synchronously transmitting the corresponding line and field synchronization signals, so that the signal streams formed in the two channels are completely consistent in terms of data content and timing progress, thus forming the initial signal forms of the first video signal stream and the second video signal stream respectively.

[0023] Finally, a synchronization verification operation is performed on the copied results. The algorithm calls the signal verification submodule to extract the key parameters (including pixel data integrity, horizontal and vertical synchronization signal timing, and clock signal frequency) of the first and second video signal streams generated in the two channels, and compares the parameters of these two signal streams with the corresponding parameters of the original video signal point by point. If the comparison results show that the parameters of the two signal streams are completely consistent with the original signal, and there is no signal crosstalk or data deviation between the two signal streams, then the synchronization copying operation is considered successful, and the generation of the two video signal streams is completed. If there are parameter mismatches or interference, the algorithm will automatically trigger a re-copying mechanism, and re-execute the signal distribution and copying process based on the timing reference signal until the two generated signal streams meet the consistency requirements with the original signal.

[0024] Step 113: The first video signal stream, used as the video stream driving the display device for real-time display, is directly converted into a signal format conforming to the display device interface specification and continuously output to the display device. Specifically, this includes: first, obtaining the interface specification information of the display device connected to the input / output module 11. This information includes the interface type of the display device (such as HDMI interface, VGA interface, etc.) and the corresponding signal format requirements. The signal format requirements specifically cover supported image resolution, image refresh rate, color depth, and color encoding method. Based on the obtained display device interface specification information, the first video signal stream is parsed to determine its current resolution, refresh rate, color depth, and color encoding method. The parsed current format parameters of the first video signal stream are compared with the format parameters required by the display device interface specification to identify the differences. Upon identifying discrepancies, the signal format conversion unit initiates adaptation processing. If resolution differences exist, a pixel scaling algorithm is used to adjust the image resolution of the signal stream to match the resolution supported by the display device. If refresh rate differences exist, a timing calibration algorithm is used to synchronously calibrate the refresh rate of the signal stream to ensure it matches the refresh rate of the display device. If color depth or color encoding methods differ, a color data conversion algorithm is used to re-encode the color information of the signal stream to ensure the color parameters conform to the receiving standards of the display device. After format conversion, the converted first video signal stream is continuously transmitted to the display device through the video output interface. During transmission, the stability and continuity of the signal output are monitored in real time to ensure that the display device can present the original video content of the industrial control computer in real time and clearly, and that the display effect is consistent with that when the industrial control computer is directly connected to the display device.

[0025] Step 114: The second video signal stream is used as the video stream for image recognition and temporarily stored in the frame buffer area to obtain a temporarily stored video signal stream. Based on the temporarily stored video signal stream, single-frame image data is captured according to a preset sampling frequency or in response to a recognition command. Specifically, this includes: firstly, configuring the parameters of the frame buffer area of ​​the input / output module 11, determining the storage capacity of the frame buffer area according to the single-frame data volume of the second video signal stream, ensuring that the storage capacity can accommodate at least one complete frame of video signal data, and configuring the address mapping rules of the frame buffer area to allocate an independent storage address for each frame of video signal for subsequent data reading and management; then, writing the second video signal stream frame by frame into the frame buffer area according to the temporal order of the video frames, recording the corresponding storage address for each frame signal according to the preset address mapping rules during the writing process, and simultaneously performing integrity verification on each frame signal data written by the data verification unit to ensure that the data writing is error-free and without loss, forming a temporarily stored video signal stream; and capturing single-frame image data. When capturing frame image data, there are two triggering methods: The first is to capture at a preset sampling frequency. First, a fixed sampling time interval is set by the control unit, such as sampling once every 50ms or 100ms. The built-in timing module of the system is started to count the time in real time. When the count of the timing module reaches the preset sampling time interval, a data reading command is triggered. The control unit locates the storage address of the latest video signal according to the address mapping rules of the frame buffer area and reads the corresponding video signal data from that address as single frame image data. The second is to capture in response to recognition commands. First, the region recognition command sent by the external application system is received through the network communication unit. The command is parsed, and the capture trigger signal and timing information contained in the command are extracted. The capture timing is determined according to the parsed trigger signal. Then, based on the capture timing, the address mapping record of the frame buffer area is queried to find the video frame storage address corresponding to the timing. The corresponding video signal data is read from that address to form single frame image data.

[0026] Step 115: Standardize the format of the single-frame image data to obtain the video frame to be identified. Specifically, this includes: first, pre-setting standardized format parameters for the video frame to be identified. These parameters determine a uniform image resolution (e.g., 1920×1080, 1280×720), a fixed color space type (e.g., RGB color space), and a standard image file format (e.g., BMP, PNG), while setting the allowable deviation range for each parameter; then, parsing the format parameters of the single-frame image data, extracting the current image resolution value, color space attributes (e.g., YUV color space, CMYK color space), and file format type (e.g., JPEG, TIFF); comparing the parsed current format parameters with the pre-set standardized format parameters item by item to determine the differences in resolution, color space, and file format; and performing targeted format standardization processing based on these differences: for resolution differences, if the current resolution is higher than the standardized resolution, The image is downsized using a pixel downsampling algorithm. If the current resolution is lower than the standardized resolution, a pixel interpolation algorithm is used to enlarge the image, ensuring the processed image resolution accurately matches the standardized resolution. For color space differences, color space conversion algorithms, such as YUV to RGB or CMYK to RGB, are used to recalculate and encode the image's color data, converting it to color data in the standardized color space, based on the conversion rules between the current color space and the standardized color space. For file format differences, the format encoding conversion unit calls the corresponding format conversion protocol, such as JPEG to BMP or TIFF to PNG, to convert the image file's encoding method to a standardized file format. After all format adjustments are completed, the processed image data is standardized and verified to ensure that its resolution, color space, and file format all meet the preset standardized parameter requirements, and that the image content is undistorted and undeformed, ultimately forming a video frame to be identified that meets the requirements of subsequent recognition and analysis.

[0027] In this embodiment of the invention, by configuring a video input interface that matches the physical specifications and electrical characteristics of the industrial control computer's video output interface (HDMI, VGA, or DVI) to receive the original video signal, the stability and lossless transmission of the original video signal can be ensured, avoiding signal distortion or interruption caused by interface incompatibility, and providing a complete and reliable original data foundation for all subsequent data processing stages. The original video signal is copied to generate two independent video signal streams, enabling parallel allocation and utilization of data without altering the essence of the original video signal. This allows the two signal streams to serve two different needs: real-time display and image recognition, respectively, without interference, ensuring the independence and integrity of the two data uses. The first video signal stream is directly converted into a signal format conforming to the display device's interface specifications and output, allowing the display device to directly recognize and receive the signal, ensuring the smoothness and image quality of the real-time display. The system achieves high accuracy, maintaining consistency with the original output of the industrial control computer to meet on-site observation requirements. Simultaneously, it directly adapts to the display interface in data processing, reducing potential delays or signal errors in intermediate conversion stages. The second video signal stream is temporarily stored in a frame buffer area, ensuring stable storage of the video stream used for image recognition and preventing data loss or corruption during transmission. Based on the temporarily stored signal stream, single-frame image data is captured at a preset sampling frequency or in response to recognition commands. This allows for flexible control of the timing and frequency of data sampling, ensuring the validity of the captured single-frame data without consuming additional resources due to oversampling, thus improving the accuracy and flexibility of image recognition data acquisition. Standardizing the format of single-frame image data to obtain the video frame to be recognized unifies the image data format specifications, allowing subsequent region recognition modules to process the data directly without additional format adaptation, reducing the pre-processing burden in the recognition stage and improving data flow efficiency.

[0028] In the system for identifying specific regions of video output from an industrial control computer according to an embodiment of the present invention, the region identification module 12, based on the video frame to be identified, responds to a region identification instruction received through a network interface, wherein the instruction carries at least one key coordinate region information, to obtain preprocessed region data for identification and analysis, including: Step 121: Receive the area identification command transmitted through the network interface. Specifically, this includes: first, configuring the communication parameters of the network interface. This configuration must ensure that the network interface and the application system sending the area identification command are consistent in terms of communication protocol, data transmission rate, and data frame format to establish a stable bidirectional communication link. The communication protocol is such as TCP / IP, the data transmission rate is such as 100Mbps or 1000Mbps, and the data frame format is such as the byte length and arrangement order of the frame header, data segment, and check segment. After the configuration is completed, a connection request signal is sent to the application system through the network interface. After receiving the connection confirmation signal returned by the application system, the communication link is officially established.

[0029] Subsequently, the instruction receiving unit is activated to monitor the data transmission status of the network interface in real time. When an area identification instruction sent by the application system is detected, the instruction data is temporarily stored in the instruction buffer area built into the module. The storage capacity of this buffer area must be preset to be no less than the maximum data amount of a single area identification instruction in order to avoid instruction data overflow.

[0030] After temporary storage is completed, a data integrity verification operation is initiated. This operation extracts the check segment information, such as the CRC check value, from the instruction data frame and compares it with the check value recalculated based on the received instruction data segment. If the two match, it is determined that the instruction data was not lost or tampered with during transmission, the instruction was received validly, and the subsequent parsing stage can proceed. If the two do not match, an instruction retransmission request is sent to the application system until a complete and valid area identification instruction is received.

[0031] Step 122 involves parsing the region identification instruction and extracting at least one key coordinate region information contained therein. Specifically, after receiving and validating the region identification instruction, the instruction parsing stage is initiated. This stage requires determining the preset data structure of the region identification instruction. This structure typically includes an instruction identifier segment (used to distinguish instruction types), an instruction length segment (identifying the total number of bytes in the entire instruction), a key coordinate region information segment (storing the coordinate data of the region to be identified), and a data verification segment (used for integrity verification in step 121). Each segment is arranged sequentially according to a fixed byte length.

[0032] First, the instruction parsing unit reads the instruction identifier segment data from the instruction buffer area to confirm that the instruction is a region identification instruction, eliminating interference from other types of instructions. Then, it reads the instruction length segment data and compares it with the total number of bytes of the instruction actually received to further confirm that the instruction data is not truncated. Subsequently, it determines the start and end byte addresses of the key coordinate region information segment based on the instruction length segment data, and extracts the key coordinate region information data from this address range.

[0033] During the extraction process, the coordinate data format needs to be parsed according to preset rules. Typically, the information of a single key coordinate region includes the horizontal coordinate (x1) and vertical coordinate (y1) of the top-left pixel and the horizontal coordinate (x2) and vertical coordinate (y2) of the bottom-right pixel. Each coordinate value is stored in binary or decimal numerical form, and the parsing unit needs to convert it into a recognizable numerical format. If the instruction carries multiple key coordinate region information, they are extracted sequentially according to the order of their arrangement in the information segment to form multiple sets of independent coordinate data.

[0034] Finally, a preliminary validity check is performed on each set of extracted coordinate data. The check includes whether the coordinate values ​​are non-negative integers, whether x2 is greater than x1 and y2 is greater than y1 in the same region. If there are coordinate data that do not meet the requirements, the set of data is marked as invalid, and the reason for invalidity is recorded. In subsequent steps, only the regions corresponding to valid coordinate data will be processed to ensure that the key coordinate region information obtained from the parsing can be directly used for positioning operations.

[0035] Step 123: Based on the key coordinate region information, locate one or more corresponding regions to be identified from the video frame to be identified. Specifically, after obtaining the structured key coordinate region information parsed in step 122, it is necessary to first obtain the image resolution parameters of the video frame to be identified. These parameters include the number of horizontal pixels (i.e., width W) and the number of vertical pixels (i.e., height H) of the video frame. These parameters can be directly retrieved from the video frame attribute information transmitted from the input / output module 11 to the region identification module 12, or they can be obtained by counting the number of rows and columns of the pixel matrix of the video frame.

[0036] Subsequently, the coordinate values ​​(x1, y1, x2, y2) of each group of valid key coordinate regions are compared with the video frame resolution parameters (W, H) to further verify the rationality of the coordinates. That is, it is confirmed that x1 and x2 are both within the range of 0 to W, and y1 and y2 are both within the range of 0 to H. If a certain set of coordinates exceeds the range, it is determined that the coordinate region exceeds the boundary of the video frame and is excluded. Only the key coordinate regions whose coordinates are completely within the range of the video frame are retained to avoid pixel data interception errors during positioning.

[0037] After successful verification, the region localization operation is initiated. This operation first converts the pixel data of the video frame to be identified into a two-dimensional pixel matrix, where the row number corresponds to the vertical coordinate y and the column number corresponds to the horizontal coordinate x. Then, for each group of valid key coordinate regions, the row range (from row y1 to row y2) and column range (from column x1 to column x2) of the region in the two-dimensional pixel matrix are calculated. Based on the calculated row and column ranges, all pixel data within the corresponding ranges are extracted from the two-dimensional pixel matrix to form an independent pixel dataset of the region to be identified.

[0038] Finally, the pixel datasets of each region to be identified are temporarily stored in a dedicated data cache. Each cache corresponds to a region to be identified and is labeled with a unique region identifier so that subsequent identification can be carried out one by one by region, while avoiding confusion of pixel data from different regions.

[0039] Step 124: Perform image preprocessing on the region to be identified to obtain preprocessed region data for identification analysis. Specifically, after completing the localization and pixel data extraction of the region to be identified, in order to ensure the effectiveness of subsequent AI recognition or OCR recognition, image preprocessing operations need to be performed on the extracted pixel dataset of the region to be identified. The operation is performed step by step according to the preset processing flow, and is performed independently for the pixel dataset of each region to be identified.

[0040] The first step is to perform noise removal. First, the characteristics of noisy pixels in the area to be identified are determined by statistical analysis of pixel grayscale values, such as isolated pixels with abrupt changes in grayscale values. Then, the neighborhood averaging method is used to process each pixel, that is, the average grayscale value of the pixel and its eight neighboring pixels is calculated, and the original pixel's grayscale value is replaced with this average value to smooth the image and remove noise. If the area to be identified is a color image, it is first converted to a grayscale image before noise removal is performed. The conversion process is calculated using a weighted average method, that is, grayscale value = 0.299 × red component + 0.587 × green component + 0.114 × blue component.

[0041] The second step is to perform a contrast adjustment operation. First, the distribution range of gray values ​​in the grayscale image of the region to be identified is calculated, i.e., the minimum gray value G. min With the maximum gray value G max If this range is less than the preset effective contrast range, such as more than 80% of 0 to 255, then it is adjusted using a grayscale stretching algorithm. The new grayscale value is calculated as: (original grayscale value - G) min )×(255 / (G max -G min This stretches the grayscale value distribution to the full range of 0 to 255 to enhance the distinction between the target and the background in the image.

[0042] The third step is to perform image normalization. Based on the preset input size of the subsequent recognition algorithm (AI or OCR), such as 256×256 pixels or 512×512 pixels, the size of the adjusted image of the region to be recognized is scaled. The scaling process uses bilinear interpolation. By calculating the grayscale value of the target pixel at the corresponding position in the original image and the grayscale value of the surrounding pixels, the grayscale value of the target pixel is obtained by interpolation. This ensures that the scaled image has no obvious distortion and that the size completely matches the input requirements of the recognition algorithm.

[0043] Finally, the quality of each preprocessed region image data to be identified is checked. The checks include noise removal effect (whether the proportion of isolated pixels is lower than the preset threshold), contrast within the effective range, and size consistent with the preset input size. If the check passes, the image data is marked as preprocessed region data and stored in the recognition data input area for subsequent recognition and analysis. If the check fails, the image preprocessing operation for that region is re-executed until the preprocessed region data that meets the requirements is obtained.

[0044] In this embodiment of the invention, receiving region recognition commands via a network interface enables remote and rapid transmission of commands, ensuring that the control requirements of external application systems can be promptly transmitted to the region recognition module. The network interface transmission method guarantees the integrity and stability of commands during transmission, preventing command loss or damage, and providing reliable raw command data for subsequent command parsing, laying the command foundation for data processing. Parsing the region recognition commands and extracting key coordinate region information allows for precise filtering of data directly related to the recognition task, eliminating redundant information. Key coordinate region information provides clear target guidance for subsequent region positioning, making data processing more focused and improving its relevance. Locating the region to be recognized from the video frame based on key coordinate region information allows for direct locking of the specific region to be analyzed, avoiding indiscriminate processing of the entire video frame. Image preprocessing of the located region unifies the format of the region data and optimizes image quality (e.g., noise removal, contrast adjustment). The preprocessed region data can be directly adapted to subsequent recognition and analysis processes, reducing the preprocessing burden of the recognition stage, ensuring smooth connection of the data processing link, and improving the overall consistency and efficiency of recognition and analysis.

[0045] In the system for identifying specific areas of video output from industrial control computers as described in this embodiment of the invention, the above-mentioned identification analysis and structured processing of the preprocessed area data to obtain standardized identification results includes: Step 125: Extract features from the preprocessed region data to obtain feature vectors representing the image content features. Specifically, this includes: after acquiring the preprocessed region data, first determining the image format of the preprocessed region data. If the data is in color image format, it is first converted into a grayscale image according to a preset grayscale conversion rule. During the conversion process, the red, green, and blue components of each pixel in the image are weighted according to the correspondence between color components and grayscale values ​​to obtain a single grayscale value, thereby simplifying the computational workload of subsequent feature extraction. If the preprocessed region data itself is already a grayscale image, it directly enters the feature extraction stage.

[0046] Subsequently, the edge feature extraction operation is initiated. This operation requires loading preset edge detection parameters, such as grayscale difference threshold and neighborhood pixel range. Then, each pixel in the grayscale image is traversed in a row-by-row and column-by-column order. The grayscale value of the current pixel is calculated to differ from the grayscale values ​​of other pixels in its preset neighborhood range (such as a 3×3 pixel area). If the grayscale difference in a certain direction exceeds the preset threshold, the pixel is determined to be an edge pixel, and its coordinate position in the image is marked. After the traversal is completed, the distribution density, continuous length, and directional features of all edge pixels are statistically analyzed to form an edge feature set.

[0047] Next, specific features are extracted for the recognition scenarios corresponding to the preprocessed area data. These scenarios include character recognition and graphic symbol recognition. For character recognition scenarios, it is necessary to further locate the character regions in the image and extract the contour features of each character according to the row and column arrangement, including the number of vertices of the contour, the distance between each vertex, and the proportion of blank areas inside the contour. For graphic symbol recognition scenarios, the shape features of the graphic are extracted, such as whether it is a regular geometric shape, the side length ratio of the graphic, and the interior angle.

[0048] Finally, the extracted edge features and specific features are quantized according to preset dimensional rules, and each feature is converted into a corresponding value. For example, edge pixel density is expressed as a percentage and the number of contour vertices is expressed as an integer. Then, all the quantized values ​​are arranged in the preset feature sorting order to form a feature vector representing the image content features. After the vector is constructed, it is necessary to verify whether the dimension of the vector is consistent with the input dimension requirements of the subsequent recognition algorithm. If they are inconsistent, the vector dimension is adjusted by padding with zeros or feature filtering to ensure that the feature vector can be directly used for recognition and analysis.

[0049] Step 126: Based on the feature vector, perform content recognition analysis using a preset recognition algorithm to obtain preliminary recognition results. Specifically, before loading the preset recognition algorithm model, the preliminary preparation and solidification integration of the preset recognition algorithm must be completed. This preset process revolves around the recognition requirements of specific areas of the industrial control computer video output, and specifically includes three core steps: algorithm selection, model and supporting resource preparation, and module integration and solidification. The results of each step directly support subsequent calculations.

[0050] First, based on the recognition scenarios of industrial control computer video output, such as character recognition of equipment operating parameters, graphic recognition of status indicator lights, and recognition of production progress bar icons, the type of preset algorithm is determined: if the recognition requirement focuses on character information in a specific area of ​​the video frame, such as numerical / alphanumeric parameters like equipment temperature and pressure, then an OCR recognition algorithm is selected. This algorithm must have the ability to adapt to low-distortion, low-contrast characters in industrial scenarios. If the recognition requirement focuses on graphic icons in a specific area of ​​the video frame, such as indicator lights for normal / fault status of equipment or progress bar icons for the production process, then an AI image recognition algorithm is selected. This algorithm must have the ability to quickly match fixed graphic icons. After the selection is completed, the core operation logic of the algorithm is recorded, such as the character contour matching logic of OCR and the feature vector distance calculation logic of AI recognition, as the basis for subsequent model preparation.

[0051] For the selected OCR algorithm, a character library matching the industrial control scenario is first constructed. This character library needs to cover the character types commonly used in industrial control equipment, including numbers 0 to 9, letters A to Z / a to z, and industrial-specific symbols such as ℃, MPa, % etc. The standard features of each character are collected and digitized. These standard features, such as the vertex coordinates of the character outline, the gray-scale variation law of the edge, and the width-to-height ratio of the character, form a character standard feature dataset. Then, the OCR algorithm model (such as the Tesseract model) is optimized and trained based on this dataset. The parameters of the model, such as the character outline matching threshold and gray-scale deviation tolerance, are adjusted to ensure that the model has fault tolerance for possible character blurring and slight tilting in industrial control scenarios. After training, an OCR algorithm model adapted for industrial control character recognition is obtained.

[0052] For the selected AI image recognition algorithm, a feature template library is first constructed. This library needs to include common graphic identifier samples from the video output of industrial control equipment, such as red fault indicator lights, green normal indicator lights, and blue progress bars. The core features of each sample are extracted and quantified. These core features, such as graphic shape parameters, color channel value range, and feature point distribution, form a feature template vector set. This template vector set is then embedded into the AI ​​recognition algorithm model. At the same time, the feature vector matching rules (such as Euclidean distance calculation rules and cosine similarity calculation rules) and matching thresholds of the model are set to complete the preparation of the AI ​​image recognition algorithm model.

[0053] The prepared OCR recognition algorithm model, AI image recognition algorithm model, and corresponding character library and feature template library are solidified into the built-in storage unit (such as a Flash memory chip) of the region recognition module through hardware burning or software implantation. Simultaneously, an algorithm call interface is configured in the control unit. This interface needs to preset two call trigger logics: when the received region recognition command carries a character recognition identifier, the OCR algorithm model and character library are automatically called; when the command carries a graphic identifier recognition identifier, the AI ​​recognition algorithm model and feature template library are automatically called. Furthermore, an independent algorithm runtime cache area is allocated in the storage unit to store temporary data during algorithm runtime, such as normalized feature vectors and similarity data during the matching process, ensuring efficient resource utilization during algorithm runtime.

[0054] First, the region recognition command is parsed to extract the recognition type identifier carried in the command, such as character recognition or graphic identifier recognition. Then, based on the identifier, a preset algorithm call interface is triggered to read the corresponding preset model (OCR algorithm model or AI recognition algorithm model) and supporting resources (character library or identifier feature template library) from the built-in storage unit and load them into the preset algorithm running cache area. For example, if the command identifier is character recognition, the preset tesseract algorithm model and industrial control character library are loaded; if the identifier is graphic identifier recognition, the preset AI recognition algorithm model and identifier feature template library are loaded.

[0055] The logic for normalizing feature vectors and initiating feature matching operations is based on the core operational rules of the preset algorithm: In the normalization process, the region recognition module calls the preset normalization range (such as the interval between 0 and 1) and operational logic (such as linear scaling logic) in the preset algorithm to adjust the feature vector values. This process does not require additional calculation of range parameters and directly reuses the numerical range determined in the preset stage, ensuring that the magnitude of the feature vector meets the operational requirements of the preset algorithm. In the feature matching operation stage, if it is an OCR algorithm, it calls the standard features of each character in the preset character library and performs similarity calculation with the normalized feature vector, such as feature overlap and contour deviation calculation. The calculation logic is the preset OCR character matching logic. If it is an AI recognition algorithm, it calls the feature vectors of each template in the preset identifier feature template library and performs distance calculation with the normalized feature vector, such as Euclidean distance and cosine distance calculation. The calculation logic is the preset AI feature matching logic.

[0056] The process of verifying candidate results based on preset similarity or distance thresholds directly uses the threshold parameters set during the pre-set process. These thresholds are determined during the pre-set algorithm preparation stage according to the recognition accuracy requirements of the industrial control scenario. For example, the similarity threshold for the OCR algorithm is set to 85%, and the distance threshold for the AI ​​recognition algorithm is set to 0.3. These thresholds are also stored in the storage unit along with the algorithm model. During verification, the threshold is directly read from the pre-set model and compared with the similarity or distance value of the current candidate result. There is no need to calculate the threshold temporarily, ensuring that the verification standard is compatible with the recognition capability of the pre-set algorithm, while ensuring the consistency and efficiency of the verification process.

[0057] The operation of labeling the preprocessed area numbers corresponding to the preliminary recognition results has its numbering rules set in conjunction with the area positioning logic of the area recognition module during the preset process: during the preset process, the association relationship between the preprocessed area numbers and the recognition algorithm call identifiers has been preset in the module, such as area 1 corresponding to character recognition and area 2 corresponding to graphic identifier recognition. When temporarily storing the preliminary recognition results, the corresponding preprocessed area numbers can be automatically matched and labeled directly according to the preset association relationship, ensuring a one-to-one correspondence between the preliminary recognition results and the preprocessed area data.

[0058] Step 127 involves performing data structuring processing on the preliminary identification results to obtain standardized identification results. Specifically, this includes: after obtaining the preliminary identification results, first determining the preset data structure of the standardized identification results. This structure needs to be configured according to the reading requirements of the application system and typically includes four core fields: identification region number, identification content, identification confidence, and identification timestamp. The identification region number is used to associate the corresponding preprocessed region, the identification content is the core information of the preliminary identification results, the identification confidence is the quantitative indicator corresponding to the similarity or distance of the candidate results in step 126, and the identification timestamp is the system time for generating the preliminary identification results.

[0059] Subsequently, the preliminary recognition results and associated preprocessing region numbers are extracted from the result cache area of ​​step 126. First, the recognition region numbers are directly filled into the corresponding fields of the preset data structure. Then, the core information in the preliminary recognition results is extracted, such as the characters recognized by OCR and the identifier names recognized by AI. If the preliminary recognition result is invalid, the preset invalid identifier is filled into the recognition content field. If there is no valid recognition information, then the similarity value or distance value recorded in step 126 is retrieved and converted into a confidence value in the range of 0 to 100 according to the preset confidence conversion rules, and filled into the recognition confidence field. Finally, the current system time of the region recognition module is read, and a timestamp is generated according to the preset time format (such as year-month-day, hour:minute:second), and filled into the recognition timestamp field.

[0060] Next, the integrity of the preset data structure of the filled fields is checked to see if all four core fields have been filled with valid information. If any field is missing, such as the recognition timestamp being empty due to the system time not being obtained, the completion mechanism is activated. The time is re-obtained through the clock unit or a preset missing identifier is generated to ensure that all fields have corresponding content. If the verification finds that the recognition content field has ambiguous information, such as ambiguous character recognition, it is necessary to return to step 126 to re-perform feature matching calculation until clear recognition content is obtained.

[0061] Finally, the data structure that has been verified and completed is converted into a preset standardized data format, such as JSON or XML. This format must meet the interface parsing requirements of the application system. During the conversion process, it must be ensured that the field names and data types are consistent with the preset rules of the application system. After the conversion is completed, the final standardized recognition result is generated and stored in the result output cache of the region recognition module.

[0062] In this embodiment of the invention, by extracting features from preprocessed region data to obtain feature vectors representing image content features, key information directly related to the recognition task can be accurately extracted from the preprocessed region image data, redundant and non-core image details are removed, and the amount of data processed in subsequent recognition and analysis steps is reduced. Simultaneously, converting preprocessed region data in image form into feature data in vector form allows the image content to be presented in a structured numerical form, enabling subsequent recognition algorithms to directly process it. This avoids additional adaptation work caused by mismatches between image data format and algorithm input requirements, ensuring smooth data processing and improving the overall efficiency of the recognition process. Based on the feature vectors, content recognition analysis is performed using a pre-set recognition algorithm to obtain preliminary recognition results. This leverages the compatibility between the pre-set algorithm and the previously extracted feature vectors, allowing the recognition algorithm to focus on the feature information representing the core content of the image, eliminating the need for indiscriminate analysis of the complete image data, reducing ineffective computation steps, and improving the targeting and efficiency of the recognition analysis. The system boasts high efficiency and reliability. Simultaneously, the pre-built recognition algorithm possesses stable operational logic and processing flow, ensuring consistency in the recognition process for identical or similar feature vectors. This avoids discrepancies in recognition results caused by fluctuations in algorithm parameters, guaranteeing the stability and reliability of the initial recognition results and providing high-quality foundational data for subsequent structured processing. By performing data structuring on the initial recognition results to obtain standardized recognition results, it can organize potentially irregular and non-uniform formatted data, such as scattered text fragments and disordered identification information, into data structures conforming to preset specifications, such as tables with fixed fields or key-value pairs with unified field names. This gives the recognition results a clear logical hierarchy and a unified presentation. This process reduces the difficulty for application systems to read and parse recognition results, reduces data integration and adaptation costs between application systems and regional recognition modules, and ensures that standardized recognition results meet the needs of subsequent data storage, transmission, and reuse. This guarantees the consistency and usability of recognition data in different scenarios and improves the overall efficiency of data utilization.

[0063] In the system for identifying a specific area of ​​video output from an industrial control computer as described in this embodiment of the invention, the above-mentioned method of returning the identification result as a response to the area identification command to the application system through the network interface includes: Step 128, receiving the standardized identification result, specifically includes: after the standardized identification result is generated, the region identification module first initializes the parameters of the built-in result receiving unit. This initialization operation needs to determine the data source of the receiving unit, that is, lock the identification data output area storing the standardized identification result, and configure the cache parameters of the receiving unit to ensure that the storage capacity of the cache area is not less than the maximum data volume of a single standardized identification result. It needs to be combined with the preset total data length of the four core fields included in the standardized result: identification region number, identification content, identification confidence, and identification timestamp, to avoid overflow during data reception.

[0064] After initialization, the result receiving process is initiated. The receiving unit reads the complete data of the standardized recognition result from the recognition data output area through the internal data bus of the module. During the reading process, the data is received field by field in the preset field order, namely recognition area number, recognition content, recognition confidence level, and recognition timestamp. At the same time, the validity of each field is verified: first, the existence of the field is verified, such as whether the recognition confidence level field is missing; second, the data type of the field is verified to conform to the preset specifications, such as whether the recognition confidence level is a value in the range of 0 to 100, and whether the recognition timestamp is in the preset year-month-day, hour:minute:second format.

[0065] If the verification finds that a field is missing or the data type is abnormal, a data retransmission request is sent to the structured processing unit in step 127 until a complete and compliant standardized identification result is received and verified. If the verification is successful, the standardized identification result is temporarily stored in the dedicated buffer of the receiving unit and labeled with the corresponding region identification instruction ID. This ID is extracted from the region identification instruction received in step 121 and is used to associate the instruction with the result to ensure that the subsequent encapsulation and transmission stages can accurately match the original instruction.

[0066] Step 129: Encapsulate the standardized identification result into a data packet conforming to a preset network communication protocol. Specifically, after receiving and temporarily storing the standardized identification result, the area identification module starts the data encapsulation unit. First, it loads the preset network communication protocol parameters and connects to the application system with network functionality through the network interface card. The preset protocol adopts the TCP / IP protocol commonly used in industrial scenarios. At the same time, it determines the fixed structure of the data packet under this protocol: including three parts: frame header, data segment, and check segment. The frame header occupies 20 bytes, including a 4-byte protocol identifier, a 8-byte total message length, and an 8-byte instruction ID. The length of the data segment is dynamically adjusted according to the actual data volume of the standardized identification result. The check segment occupies 4 bytes and adopts the CRC32 check algorithm.

[0067] Next, the construction operations for each part of the message are carried out: The first step is to construct the frame header. Starting from the first byte of the frame header, the TCP / IP protocol identifier, the total message length, and the region identification instruction ID are filled in sequentially. The TCP / IP protocol identifier is preset to 0x01020304. The total message length is calculated as 20 bytes for the frame header + the actual number of bytes in the data segment + 4 bytes for the check segment, and then converted into 8 bytes of binary data. The region identification instruction ID is also converted into 8 bytes of binary data. The second step is to construct the data segment. The temporarily stored standardized identification results are converted into binary data streams conforming to the TCP / IP protocol in the order of the fields of identification region number, identification content, identification confidence level, and identification timestamp. These data are then filled into the corresponding positions in the data segment, and the actual number of bytes in the data segment is recorded for verification of the total message length in the frame header. The third step is to construct the check segment. Based on the CRC32 check algorithm, the check value is calculated on the complete binary data of the constructed frame header and data, and the resulting 4-byte check value is filled into the check segment.

[0068] After the data packet is constructed, a format compliance check is performed on the entire data packet: the protocol identifier in the frame header is checked to see if it is correct, whether the calculated total length of the packet is consistent with the actual number of bytes in the frame header + data segment + check segment, and whether the check value of the check segment matches the recalculated result. If the check passes, the data packet is determined to meet the requirements of the preset network communication protocol and can proceed to the subsequent sending stage. If the check fails, the packet construction operation is re-executed until a compliant data packet is generated.

[0069] Step 1210: Send the data packet to the application system through the network interface to complete the closed loop of the response to the area identification command. Specifically, this includes: After completing the data packet encapsulation, the area identification module first checks the communication status of the built-in network interface: sends a link detection data packet to the application system through the network interface and waits for the application system to return a link normal acknowledgment; if an acknowledgment is received within a preset time (e.g., 500ms), it is determined that the network link is in a connected state and data transmission can be started; if no acknowledgment is received, the link detection data packet is resent every 200ms until an acknowledgment is received or the preset number of retries is reached, such as 3 times; if the retries fail, a link interruption log is recorded and the transmission process is terminated, and an alarm signal is sent to the module alarm unit.

[0070] After the link connection is confirmed, the data transmission process is initiated: the network interface sends data packets byte by byte to the application system in the form of a binary data stream according to the transmission rules of the TCP / IP protocol; during the transmission process, the transmission progress is monitored in real time by calculating the ratio of the number of bytes sent to the total number of bytes in the packet, and the data reception status signal fed back by the application system is also monitored; if a receipt signal indicating that the packet has been received is received, it is determined that the data packet has been successfully delivered; if a receipt signal indicating that the packet is missing or has a check error is received, the data packet is retransmitted until the application system confirms that the packet has been received completely.

[0071] After confirming that the data message has been successfully delivered, the area identification module records information such as the area identification instruction ID, the data message sending time, and the application system's response time to the module's built-in operation log unit. At the same time, it marks the processing status of the area identification instruction as completed, thus formally completing the response closed loop from receiving the area identification instruction to returning the identification result. This ensures that the complete processing flow of the instruction can be traced through the log, and also provides clear evidence of completed response for subsequent business operations of the application system (such as equipment monitoring and production scheduling).

[0072] In this embodiment of the invention, receiving standardized identification results can accurately inherit the complete data after structured processing, ensuring that the information in core fields such as identification area number, identification content, identification confidence level, and identification timestamp in the standardized identification results is not lost or tampered with during the flow, providing a complete and reliable original data foundation for subsequent data encapsulation. Simultaneously, simple field existence verification can confirm that no key information is missing from the standardized identification results, avoiding interruptions in subsequent encapsulation operations or the generation of invalid data packets due to incomplete data, ensuring a smooth connection in the data processing link from structured results to the encapsulation stage. Encapsulating the standardized identification results into data packets conforming to a preset network communication protocol allows the identification results to adapt to the format specifications of network transmission, and through the message structure specified by the preset protocol... Frames with fixed-length headers, clearly defined data segment divisions, and dedicated checksums ensure that data has a consistent basis for parsing during network transmission, avoiding application systems' inability to recognize or parse data due to format incompatibility. Simultaneously, checksum information (such as CRC checksums) incorporated during encapsulation can be used to verify data integrity in subsequent transmission stages, reducing data corruption caused by signal interference or link fluctuations during transmission and improving the reliability of recognition results before transmission. Sending data packets to the application system via the network interface leverages the stable transmission capabilities of the network interface to achieve efficient and real-time transmission of recognition results, ensuring that the application system can promptly obtain response data to previous area recognition commands, supporting subsequent business operations such as equipment monitoring and production scheduling based on the recognition results.

[0073] like Figure 2As shown, a control method for a system that identifies a specific area of ​​video output from an industrial control computer is disclosed. The control method includes: Receive raw video signals from the industrial control computer; The original video signal is split into a video stream for driving the display device for real-time display and a video stream for image recognition, and the video frame to be recognized is captured from the video stream for image recognition. Based on the video frame to be identified, in response to a region identification instruction received through a network interface, wherein the instruction carries at least one key coordinate region information, preprocessed region data for identification analysis is obtained. The preprocessed region data is identified, analyzed, and structured to obtain standardized identification results; The identification result is used as a response to the region identification command and is returned to the application system through the network interface.

[0074] In this embodiment, as Figure 3 As shown, firstly, connect the VGA / HDMI out interface of the production equipment to the VGA / HDMI in interface of this device. This device connects to the monitor through the VGA / HDMI out interface to achieve the original local display output effect. At the same time, it receives recognition instructions, captures video frames, recognizes the specified area, and returns the recognition results to the application system. When specifying the area, one or more key coordinate areas can be specified for simultaneous recognition. The recognition method can adopt an AI artificial intelligence image recognition system or use OCR recognition software such as Tesseract.

[0075] Embodiments of the present invention also provide a computing device, including: a processor and a memory storing a computer program, wherein the computer program, when executed by the processor, performs the system as described above. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0076] Embodiments of the present invention also provide a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the system as described above. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A system for identifying specific areas in the video output of an industrial control computer, characterized in that, include: The input / output module is used to receive raw video signals from an industrial control computer; split the raw video signals into a video stream for driving a display device for real-time display and a video stream for image recognition; and capture the video frames to be recognized from the video stream for image recognition. The region identification module is used to obtain preprocessed region data for identification analysis based on the video frame to be identified, in response to a region identification instruction received through a network interface, wherein the instruction carries at least one key coordinate region information. The preprocessed region data is identified, analyzed, and structured to obtain standardized identification results; the identification results are then returned to the application system via the network interface as a response to the region identification command.

2. The system for identifying specific areas of video output from an industrial control computer according to claim 1, characterized in that, Receive raw video signals from the industrial control computer, including: The video input interface directly connects to and receives raw video signals from the video output interface of the industrial control computer; wherein the physical specifications and electrical characteristics of the video input interface are configured to match the video output interface of the industrial control computer, and the video output interface is any one of HDMI, VGA or DVI interface.

3. The system for identifying specific areas of video output from an industrial control computer according to claim 2, characterized in that, The original video signal is split into a video stream for driving a display device for real-time display and a video stream for image recognition, and the video frame to be recognized is extracted from the video stream for image recognition, including: The original video signal is copied to generate a first video signal stream and a second video signal stream; The first video signal stream is used as the video stream to drive the display device for real-time display. It is directly converted into a signal format that conforms to the display device interface specification and continuously output to the display device. The second video signal stream is used as the video stream for image recognition and sent to the frame buffer area for temporary storage to obtain a temporary video signal stream; based on the temporary video signal stream, single frame image data is captured according to a preset sampling frequency or in response to a recognition command. The single-frame image data is processed to standardize the format to obtain the video frame to be identified.

4. The system for identifying specific areas of video output from an industrial control computer according to claim 3, characterized in that, Based on the video frame to be identified, in response to a region identification command received via a network interface, wherein the command carries at least one key coordinate region information, preprocessed region data for identification analysis is obtained, including: Receive area identification commands transmitted via the network interface; The region identification command is parsed to extract at least one key coordinate region information contained therein; Based on the key coordinate region information, locate one or more corresponding regions to be identified from the video frame to be identified; The region to be identified is preprocessed to obtain preprocessed region data for identification and analysis.

5. The system for identifying specific areas of video output from an industrial control computer according to claim 4, characterized in that, The preprocessed region data is identified, analyzed, and structured to obtain standardized identification results, including: Feature extraction is performed on the preprocessed region data to obtain feature vectors representing the image content features; Based on the feature vector, content recognition analysis is performed using a preset recognition algorithm to obtain preliminary recognition results; The preliminary identification results are then subjected to data structuring processing to obtain standardized identification results.

6. The system for identifying specific areas of video output from an industrial control computer according to claim 5, characterized in that, The identification result is returned to the application system via the network interface as a response to the region identification command, including: Receive the standardized identification result; The standardized identification results are encapsulated into data packets conforming to a preset network communication protocol; The data packet is sent to the application system through the network interface to complete the closed loop of response to the area identification command.

7. A control method for a system that identifies a specific area of ​​video output from an industrial control computer, characterized in that, Applied to the system as described in any one of claims 1 to 6, the method comprises: Receive raw video signals from the industrial control computer; The original video signal is split into a video stream for driving the display device for real-time display and a video stream for image recognition, and the video frame to be recognized is captured from the video stream for image recognition. Based on the video frame to be identified, in response to a region identification instruction received through a network interface, wherein the instruction carries at least one key coordinate region information, preprocessed region data for identification analysis is obtained. The preprocessed region data is identified, analyzed, and structured to obtain standardized identification results; The identification result is used as a response to the region identification command and is returned to the application system through the network interface.

8. A computing device, characterized in that, include: One or more processors; A storage device for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to perform the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program that, when executed by a processor, implements the method as described in any one of claims 1 to 6.

Images