Intelligent optimization method and system for image processing algorithm of industrial cameras

Abstract

The present invention provides an intelligent optimization method and system for image processing algorithms of industrial cameras, including: based on the industrial camera to collect target images, perform feature image recognition and cropping on the target images to obtain multiple feature images; obtain image processing requirement information of the industrial camera, and match multiple image processing algorithms corresponding to the image processing requirement information in a database; based on each of the image processing algorithms, synchronously process the feature images to obtain the processing effects of each image processing algorithm on the feature images; determine the algorithm with the best processing effect from each image processing algorithm as the target image processing algorithm; obtain the device status of the industrial camera and the lighting status of the current environment; optimize the algorithm parameters of the target image processing algorithm based on the device status and lighting status. In the present invention, the defect that the algorithm cannot be adjusted for different application scenarios and environmental conditions is overcome.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to an intelligent optimization method and system for an image processing algorithm of an industrial camera. Background Art

[0002] Industrial cameras are widely used in various visual inspection, recognition and measurement tasks. With the development of the Internet of Things (IoT), higher requirements are placed on the speed, accuracy and adaptability of industrial camera image processing.

[0003] Traditional industrial cameras usually use fixed image processing algorithms and often cannot select the best image processing algorithm. When the application scenario changes or the environmental conditions fluctuate, the performance of the image processing algorithm will drop significantly. For example, under different lighting conditions, the same algorithm cannot maintain consistent processing effects. Summary of the invention

[0004] The main purpose of the present invention is to provide an intelligent optimization method and system for the image processing algorithm of an industrial camera, aiming to overcome the defect that the performance of the image processing algorithm is significantly reduced when the application scenario or environmental conditions change.

[0005] To achieve the above object, the present invention provides an intelligent optimization method for an image processing algorithm of an industrial camera, comprising the following steps:

[0006] Based on the industrial camera, the target image is collected, and the feature image is recognized and cropped to obtain multiple feature images; the image processing requirement information of the industrial camera is obtained, and multiple image processing algorithms corresponding to the image processing requirement information are matched in the database;

[0007] Based on each of the image processing algorithms, the characteristic images are synchronously processed to obtain processing effects of each of the image processing algorithms on the characteristic images; wherein the processing effects include processing speed and accuracy;

[0008] Determine the algorithm with the best processing effect from various image processing algorithms as the target image processing algorithm; obtain the device status of the industrial camera and the lighting status of the current environment; and optimize the algorithm parameters of the target image processing algorithm based on the device status and lighting status.

[0009] Furthermore, the acquisition of target images based on industrial cameras includes:

[0010] The camera module based on the industrial camera obtains the preview image of the target object and enhances the preview image in real time;

[0011] The enhancement process includes:

[0012] By integrating ray tracing technology, multiple light sources are simulated to illuminate the target object;

[0013] Use optical flow algorithm to estimate the motion trajectory of the target object to capture the target object;

[0014] Virtual fluid dynamics is used to simulate the flow effects on the surface of objects and dynamically enhance image features.

[0015] Furthermore, the image processing requirement information includes image defect detection and object recognition.

[0016] Furthermore, the optimizing the algorithm parameters of the target image processing algorithm based on the device state and the illumination state includes:

[0017] In a database, matching the optimization value corresponding to the device state and the illumination state; wherein the database stores a mapping relationship between the device state, the illumination state and the optimization value;

[0018] The algorithm parameters of the target image processing algorithm are optimized based on the optimization value; wherein the optimization value includes the adjustment amount of each algorithm parameter.

[0019] Further, after optimizing the algorithm parameters of the target image processing algorithm based on the device state and the illumination state, the method further comprises:

[0020] Processing the target image based on the optimized target image processing algorithm to obtain a processed image, and saving the processed image to a preset folder;

[0021] Obtaining identification information of each of the image processing algorithms;

[0022] generating an encrypted password based on identification information of each of the image processing algorithms;

[0023] encrypting the preset folder based on the encryption password;

[0024] Based on the identification information of the target image processing algorithm, an identification character is generated, and the identification character is used as the folder name of the preset folder.

[0025] Furthermore, the generating of an encrypted password based on the identification information of each of the image processing algorithms comprises:

[0026] Sorting each of the image processing algorithms based on a preset order to obtain an algorithm sequence;

[0027] Create a data table; wherein the data table includes cells of multiple rows and columns;

[0028] According to the algorithm sequence, the identification information of each of the image processing algorithms is sequentially added to the data table to obtain a first identification data table; wherein the identification information of one image processing algorithm is added in a row, and one character in the identification information is added in one cell;

[0029] Rearranging the first identification data table to obtain a rearranged data table;

[0030] Each character in the identification information of the target image processing algorithm is used as the target character;

[0031] In the rearrangement data table, all characters identical to the target character are removed to obtain a removed character table;

[0032] In the eliminated character table, the number of cells having characters in each row is obtained;

[0033] The number of cells with characters in each row is combined in sequence to obtain a combined number as the encryption password.

[0034] Furthermore, the first identification data table is rearranged to obtain a rearranged data table, including:

[0035] In the first identification data table, the row where the identification information of the target image processing algorithm is located is obtained as the first target row; each numeric character in the first target row is obtained as the target number;

[0036] In the first identification data table, the row corresponding to the target number is used as the second target row; each of the second target rows is translated to the end of the data table to obtain a rearranged data table.

[0037] Furthermore, the target image is subjected to feature image recognition and cropping to obtain a plurality of feature images, including:

[0038] Identifying a plurality of characteristic targets in the target image based on a target detection algorithm;

[0039] A bounding box is added to each of the feature targets, and the images are cropped according to the edges of the bounding box to obtain multiple feature images.

[0040] Furthermore, the target detection algorithm is used to identify multiple characteristic targets in the target image, including:

[0041] Obtaining an association code of an industrial camera, and obtaining a plurality of target detection algorithms corresponding to the association code from a database; wherein each target detection algorithm detects a different target;

[0042] A plurality of target images are copied, and a target detection algorithm is used to identify characteristic targets for each target image.

[0043] The present invention also provides an intelligent optimization system for an image processing algorithm of an industrial camera, comprising:

[0044] An acquisition unit is used to acquire a target image based on an industrial camera, perform feature image recognition and cropping on the target image to obtain multiple feature images; obtain image processing requirement information of the industrial camera, and match multiple image processing algorithms corresponding to the image processing requirement information in a database;

[0045] A processing unit, used to synchronously process the characteristic images based on the respective image processing algorithms to obtain processing effects of the respective image processing algorithms on the characteristic images; wherein the processing effects include processing speed and accuracy;

[0046] The optimization unit is used to determine the algorithm with the best processing effect from various image processing algorithms as the target image processing algorithm; obtain the device status of the industrial camera and the lighting status of the current environment; and optimize the algorithm parameters of the target image processing algorithm based on the device status and lighting status.

[0047] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above methods when executing the computer program.

[0048] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned methods are implemented.

[0049] The image processing algorithm intelligent optimization method and system of the industrial camera provided by the present invention include: based on the industrial camera to collect the target image, the target image is identified and cropped to obtain multiple feature images; the image processing requirement information of the industrial camera is obtained, and multiple image processing algorithms corresponding to the image processing requirement information are matched in the database; the feature image is synchronously processed based on each of the image processing algorithms to obtain the processing effect of each image processing algorithm on the feature image; wherein the processing effect includes processing speed and accuracy; the algorithm with the best processing effect is determined from each image processing algorithm as the target image processing algorithm; the device status of the industrial camera and the lighting status of the current environment are obtained; the algorithm parameters of the target image processing algorithm are optimized based on the device status and lighting status. In the present invention, the most suitable image processing algorithm is automatically selected, and the changes in the actual device status and lighting conditions are considered at the same time, and the algorithm parameters are adjusted in real time to ensure that the best image processing effect can be achieved under various conditions. The above technical solution overcomes the defect that the algorithm cannot be adjusted for different application scenarios and environmental conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 1 is a schematic diagram of the steps of an intelligent optimization method for an image processing algorithm for an industrial camera according to an embodiment of the present invention;

[0051] Figure 2 It is a structural block diagram of an intelligent optimization system for an image processing algorithm of an industrial camera in one embodiment of the present invention;

[0052] Figure 3 It is a schematic block diagram of the structure of a computer device according to an embodiment of the present invention.

[0053] The implementation, functional features and advantages of the present invention will be further described in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0055] Reference Figure 1 In one embodiment of the present invention, a method for intelligently optimizing an image processing algorithm for an industrial camera is provided, comprising the following steps:

[0056] Step S1, collecting a target image based on an industrial camera, performing feature image recognition and cropping on the target image to obtain multiple feature images; obtaining image processing requirement information of the industrial camera, and matching multiple image processing algorithms corresponding to the image processing requirement information in a database;

[0057] Step S2, synchronously processing the feature images based on the respective image processing algorithms to obtain the processing effects of the respective image processing algorithms on the feature images; wherein the processing effects include processing speed and accuracy;

[0058] Step S3, determining the algorithm with the best processing effect from various image processing algorithms as the target image processing algorithm; obtaining the device status of the industrial camera and the lighting status of the current environment; and optimizing the algorithm parameters of the target image processing algorithm based on the device status and lighting status.

[0059] In this embodiment, it specifically includes:

[0060] Target image acquisition: This stage includes the selection of appropriate industrial cameras. Depending on the application scenario, high-resolution cameras are used to capture subtle defects, or high-speed cameras are used to adapt to fast production lines. When acquiring images, the camera's installation position, angle, and distance from the subject need to be considered to ensure that the image quality meets the requirements of subsequent processing. Use industrial cameras to capture target images, perform feature image recognition and cropping on them, and obtain multiple feature images. This preprocessing step helps reduce the amount of computation for subsequent processing while ensuring that the algorithm can focus on key areas in the image.

[0061] Image processing requirements acquisition: Specific requirements analysis is the key to customized processing strategies. The image processing requirements information mainly includes image defect detection and object recognition. Specifically, image defect detection can include specific tasks such as metal surface scratch detection and product size measurement; object recognition can be the accurate classification of parts of different models. For different image processing requirements, there will be differences in image processing algorithms, and these requirements will directly affect the selection of algorithms and parameter configuration. Specifically, search and match suitable image processing algorithms in the database. The algorithm library usually contains a variety of algorithms, such as defect detection, edge detection, morphological operations, machine learning classifiers, etc. Each algorithm has its specific advantages and applicable scenarios.

[0062] Algorithm synchronization and effect evaluation: Next, all the selected candidate algorithms will be used to synchronously process the feature images. During the processing, key performance indicators such as the processing speed and accuracy of each algorithm are recorded to evaluate the effects of different algorithms.

[0063] Determination of the best algorithm: Based on the above evaluation results, the algorithm with the best processing effect will be selected from the candidate algorithms as the target image processing algorithm. The best here refers to the best comprehensive processing speed and accuracy, or a certain indicator is given more importance according to the specific application scenario.

[0064] Algorithm parameter optimization and environmental adaptation,After determining the target algorithm, it is necessary to further obtain the device status of the industrial camera (such as hardware configuration, operating temperature) and the current environmental lighting conditions. This information is crucial for optimizing algorithm parameters because they directly affect the image quality and algorithm execution efficiency.

[0065] Device status monitoring includes checking camera temperature, CPU load, etc. to ensure that the hardware is running within a safe range. Light status detection involves measuring ambient light intensity and light source color temperature.

[0066] Parameter tuning strategy: Dynamically adjust algorithm parameters according to environmental changes, such as dynamically adjusting the algorithm threshold, filter parameters, model weights, changing the degree of image enhancement, adjusting image preprocessing parameters, etc., to ensure that the algorithm can achieve the best processing effect in a specific environment. This step can be automated using intelligent optimization methods such as genetic algorithms and particle swarm optimization.

[0067] In summary, the above technical solution realizes highly intelligent and adaptive management of industrial camera image processing by deeply integrating image feature processing, algorithm optimization and dynamic parameter adjustment, effectively solving the limitations of traditional methods in complex industrial environments. It automatically selects the most suitable image processing algorithm, takes into account the actual device status and changes in lighting conditions, and adjusts the algorithm parameters in real time to ensure the best image processing effect under various conditions. The above technical solution overcomes the current defect that the algorithm cannot be adjusted for different application scenarios and environmental conditions.

[0068] In one embodiment, the step of acquiring a target image based on an industrial camera includes:

[0069] The camera module based on the industrial camera obtains the preview image of the target object and enhances the preview image in real time;

[0070] The enhancement process includes:

[0071] By integrating ray tracing technology, multiple light sources are simulated to illuminate the target object;

[0072] Use optical flow algorithm to estimate the motion trajectory of the target object to capture the target object;

[0073] Virtual fluid dynamics is used to simulate the flow effects on the surface of objects and dynamically enhance image features.

[0074] In this embodiment, it specifically includes:

[0075] Camera module preview image: Industrial cameras capture real-time images of targets through highly sensitive camera modules. The ability to instantly obtain preview images is critical for rapid response and dynamic monitoring, especially in high-speed production lines or scenarios that require real-time feedback.

[0076] Enhanced processing technologies include: integrated ray tracing technology, multi-light source simulation illumination: This technology simulates a variety of lighting conditions in nature, calculates the propagation path of light in the scene through software algorithms, and can virtually project multiple light sources on the target object, thereby improving the details of the shadow area and enhancing the overall contrast and three-dimensional sense of the image. This is especially important in environments with insufficient or uneven lighting, and can help reveal surface features or defects that were originally hidden in the shadows.

[0077] Motion trajectory estimation: The optical flow algorithm is a computer vision technique used to estimate the direction and speed of movement of pixels between consecutive frames. In industrial camera applications, it is used to capture fast-moving or rotating targets and accurately locate them even when the targets are in dynamic changes. By tracking the motion trajectory of the target in real time, the continuity and stability of image capture can be ensured, avoiding image quality degradation caused by motion blur.

[0078] Virtual fluid dynamics application, dynamic image feature enhancement: This is an innovative method that draws on the principles of fluid mechanics to dynamically adjust the brightness, contrast and color saturation of the image by simulating the flow of virtual fluid on the surface of an object. This technology can enhance the details and texture features of the surface of an object, especially for objects with special surface effects (such as gloss and reflection), which can significantly improve the clarity and realism of the image and facilitate subsequent feature extraction and recognition.

[0079] In this embodiment, the integration of the above-mentioned series of enhanced processing technologies not only enhances the visual effect of the image, but more importantly, improves the robustness and accuracy of the image processing algorithm. Ray tracing technology solves the problem of insufficient and uneven lighting, the optical flow algorithm ensures stable tracking of targets in dynamic scenes, and the application of virtual fluid dynamics further optimizes the image features, so that industrial cameras can still capture high-quality images in complex and changing environments, providing more reliable image data for subsequent defect detection, object recognition and other tasks.

[0080] In one embodiment, the image processing requirement information includes image defect detection and object recognition.

[0081] In this embodiment, the image processing requirement information is an important basis for determining the image processing algorithm selection and parameter configuration, which is directly related to the specific goals of industrial detection and recognition tasks.

[0082] Image defect detection refers to the use of images taken by industrial cameras to identify and locate abnormal or non-standard defects on the surface of products or materials, such as cracks, stains, deformations, missing parts, etc. through algorithm analysis. This process is of great significance to improving product quality control and reducing the outflow of defective products. In order to achieve efficient defect detection, the algorithm needs to be highly sensitive to tiny abnormal changes and have strong anti-interference capabilities to accurately identify defects in different lighting conditions, changes in viewing angles or background clutter. Common technical means include template matching, edge detection, machine learning classifiers, and deep learning models.

[0083] Object recognition refers to the automatic identification and classification of specific objects or subjects in an image. In industrial scenarios, it involves the identification of different parts, raw materials, and packaging to facilitate automated sorting, tracking, or guiding robot operations. Object recognition requires the algorithm to accurately extract object features from complex backgrounds and classify or instance segment them. Traditional methods such as SIFT and SURF feature matching and modern deep learning models such as YOLO and Mask R-CNN are all effective tools for object recognition. The accuracy of object recognition directly affects whether the subsequent automation process can proceed smoothly, so it is one of the core functions of industrial vision systems.

[0084] In practical applications, clear information about image processing requirements not only guides the selection of algorithms, but also affects the adjustment and optimization strategies of algorithm parameters. For example, defect detection may focus more on capturing high resolution and tiny details, while object recognition may focus more on the efficiency of rapid recognition and classification. Therefore, customized image processing processes based on specific requirements are the key to improving the performance of the entire industrial vision system.

[0085] In one embodiment, the optimizing the algorithm parameters of the target image processing algorithm based on the device state and the illumination state includes:

[0086] In a database, matching the optimization value corresponding to the device state and the illumination state; wherein the database stores a mapping relationship between the device state, the illumination state and the optimization value;

[0087] The algorithm parameters of the target image processing algorithm are optimized based on the optimization value; wherein the optimization value includes the adjustment amount of each algorithm parameter.

[0088] In this embodiment, it specifically includes:

[0089] Database construction and mapping relationship: The database stores a large amount of historical data, recording the mapping relationship between different device states (such as camera temperature, lens clarity), lighting conditions (such as light intensity, light source color temperature) and algorithm parameter optimization values. These optimization values ​​are obtained through a large number of experiments, tests and data analysis in the early stage, reflecting the optimal parameter settings under specific conditions.

[0090] Each mapping relationship is essentially a lookup table or model that can quickly find or predict the corresponding algorithm parameter adjustment based on the input device status and light status. For example, when the device is under high temperature and low light conditions, the database indicates that the image brightness, contrast, and exposure time need to be increased to compensate for the negative impact of the environment.

[0091] Dynamic parameter optimization process: Real-time monitoring of the device status of the industrial camera and the lighting status of the current environment, which is usually done through built-in sensors and environmental monitoring modules.

[0092] Query the database: Based on the monitored status information, search the database for the closest or matching device status and lighting status combination to obtain the corresponding optimization value.

[0093] Parameter adjustment: According to the optimized value found, the parameters of the image processing algorithm are automatically adjusted, such as gain, contrast, sharpness, noise reduction level, etc., to ensure that the algorithm can perform at its best under the current conditions. The adjustment amount includes direct numerical adjustment (such as increasing brightness by 10%), or dynamic adjustment strategies based on proportions and rules.

[0094] Effect verification and feedback: The optimized parameters are applied in the image processing process, and the processing effect can be continuously monitored, such as key indicators such as accuracy and processing speed, and the optimization value can be further fine-tuned according to the actual effect to form a closed-loop feedback mechanism, gradually optimize the mapping relationship in the database, and realize self-learning and evolution.

[0095] In this embodiment, the above technical solution has the following advantages:

[0096] Strong adaptability: This technical solution enables the image processing system to dynamically adapt to the changing environmental conditions in the factory without the need for frequent human intervention, thus improving the stability and efficiency of the system.

[0097] Intelligent management: Through intelligent parameter optimization strategies, it reduces dependence on professional technicians and reduces maintenance costs, while accelerating processing speed and improving the quality and accuracy of image processing.

[0098] Continuous optimization: With the enrichment of database content and the iteration of optimization algorithms, it is possible to continuously learn and improve. In the long run, it will enhance the intelligence level of the entire industrial visual inspection and recognition system.

[0099] In one embodiment, after optimizing the algorithm parameters of the target image processing algorithm based on the device state and the illumination state, the method further includes:

[0100] Processing the target image based on the optimized target image processing algorithm to obtain a processed image, and saving the processed image to a preset folder;

[0101] Obtaining identification information of each of the image processing algorithms;

[0102] generating an encrypted password based on identification information of each of the image processing algorithms;

[0103] encrypting the preset folder based on the encryption password;

[0104] Based on the identification information of the target image processing algorithm, an identification character is generated, and the identification character is used as the folder name of the preset folder.

[0105] In this embodiment, it specifically includes:

[0106] Optimization: After optimizing the parameters of the image processing algorithm according to the device status and lighting conditions, these optimized algorithm parameters are immediately applied to process the original target image. This step ensures the maximum image quality and improves the image recognizability and analysis accuracy in a specific environment.

[0107] Save processing results: The processed images are automatically saved to a preset folder. This operation automates the downstream process of image processing, simplifies the workflow and facilitates subsequent analysis or archiving.

[0108] Obtain algorithm identification information: For each image processing algorithm involved in the processing, its unique identification information will be extracted. This identification information may include the algorithm name, version number, parameter configuration, etc., which is a digital representation of the algorithm characteristics.

[0109] Generate Encryption Password: Using this algorithm identification information, a unique encryption password is generated using a predetermined encryption algorithm. This practice directly links the use of the algorithm to the folder's encryption key, increasing security and ensuring that only users who know the specific algorithm details can access the data.

[0110] Folder Encryption: Encrypt the preset folder using the generated encryption password. This step is to prevent unauthorized access and protect sensitive or proprietary image data from being leaked after processing.

[0111] Generate identification characters: Further, extract key elements from the identification information of the target image processing algorithm to generate a short, representative identification character that reflects the characteristics of the algorithm and is easy to identify.

[0112] Name the folder: Use this identifier as the name of the preset folder. This naming method makes the folder structure clear, easy to manage and retrieve, especially when processing a large number of images processed by different algorithms, you can quickly locate a specific processing result.

[0113] In summary, the above technical solution combines intelligent image processing with data security management, ensuring that every step from image processing to data storage is efficient and secure. It not only improves the quality and efficiency of image processing, but also effectively protects the privacy of processed images through dynamic encryption strategies, which is suitable for application scenarios with strict requirements on data security.

[0114] In one embodiment, generating an encrypted password based on identification information of each of the image processing algorithms includes:

[0115] Sorting each of the image processing algorithms based on a preset order to obtain an algorithm sequence;

[0116] Create a data table; wherein the data table includes cells of multiple rows and columns;

[0117] According to the algorithm sequence, the identification information of each of the image processing algorithms is sequentially added to the data table to obtain a first identification data table; wherein the identification information of one image processing algorithm is added in a row, and one character in the identification information is added in one cell;

[0118] Rearranging the first identification data table to obtain a rearranged data table;

[0119] In the first identification data table, the row where the identification information of the target image processing algorithm is located is obtained as the first target row; each numeric character in the first target row is obtained as the target number;

[0120] In the first identification data table, the row corresponding to the target number is used as the second target row; each of the second target rows is translated to the end of the data table to obtain a rearranged data table;

[0121] Each character in the identification information of the target image processing algorithm is used as the target character;

[0122] In the rearrangement data table, all characters identical to the target character are removed to obtain a removed character table;

[0123] In the eliminated character table, the number of cells having characters in each row is obtained;

[0124] The number of cells with characters in each row is combined in sequence to obtain a combined number as the encryption password.

[0125] The first identification data table is rearranged to obtain a rearranged data table, including:

[0126] In the first identification data table, the row where the identification information of the target image processing algorithm is located is obtained as the first target row; each numeric character in the first target row is obtained as the target number;

[0127] In the first identification data table, the row corresponding to the target number is used as the second target row; each of the second target rows is translated to the end of the data table to obtain a rearranged data table.

[0128] In this embodiment, it specifically includes:

[0129] Algorithm sorting: First, sort all participating image processing algorithms according to a preset order (which can be algorithm priority, complexity, or alphabetical order, etc.) to form an ordered algorithm sequence. This step ensures the consistency and predictability of information organization.

[0130] Create a data table: Build a data table structure with multiple rows and columns to prepare for the subsequent steps. This data table is like a matrix, and each cell is used to store a single character in the algorithm identification information.

[0131] Fill the first identification data table: fill the identification information of each algorithm one by one in the data table according to the algorithm sequence, and each row represents the identification information of an algorithm. The purpose of this is to accurately locate and operate the information of a specific algorithm later.

[0132] Determine the first target row: Find the row where the identification information of the target image processing algorithm is located. This row will become the key reference for subsequent operations.

[0133] Rearrange the data table: Identify the numeric characters in the first target row, find the corresponding row (second target row) in the data table, and then move the second target row to the end of the data table. This step uses specific numeric characters as key features to disrupt the layout of the original data table and enhance security and uniqueness.

[0134] Eliminate target characters: All characters that are identical to the characters in the target algorithm identification information are removed from the rearranged data table, leaving a screened table that does not contain direct identification information of the target algorithm, called the eliminated character table.

[0135] Count the number of character cells: Calculate the number of cells with characters remaining in each row after removing the characters from the table. This is a way to indirectly reflect the original information by using the distribution pattern of the remaining characters.

[0136] Combination number: connect the number of character cells in each row in series to form a sequence, which is the final encryption password. Because it is obtained after complex operations based on algorithm identification information and has no direct character association, it is very difficult to crack.

[0137] In summary, the above method converts the algorithm identification information into a seemingly random digital sequence through complex row shifting, character filtering and quantity encoding, which greatly improves the complexity and security of the encryption process. Even if the attacker knows part of the algorithm information, it is difficult to reversely calculate the original encryption password, thereby effectively protecting the security of the stored image data. In addition, the flexibility of this method lies in the fact that the preset order and filtering rules can be adjusted according to different needs, further enhancing the diversity of encryption strategies.

[0138] In one embodiment, generating an encrypted password based on identification information of each of the image processing algorithms includes:

[0139] Sorting each of the image processing algorithms based on a preset order to obtain an algorithm sequence;

[0140] Create a data table; wherein the data table includes cells of multiple rows and columns;

[0141] According to the algorithm sequence, the identification information of each of the image processing algorithms is sequentially added to the data table to obtain a second identification data table; wherein the identification information of one image processing algorithm is added in a row, and one character in the identification information is added in one cell;

[0142] In the second identification data table, obtaining a row where identification information of a target image processing algorithm is located as a third target row;

[0143] Obtaining the digital characters in the third target row, and calculating the average and sum of the digital characters; generating a linear line in a coordinate system based on the average and sum;

[0144] Rotate the linear line clockwise by ninety degrees and then translate it into the second identification data table; wherein the linear line passes through the upper left corner vertex of the second identification data table after translation;

[0145] A cell intersecting the linear line is determined as a target cell, and characters in the target cell are combined in sequence to serve as the encrypted password.

[0146] In this embodiment, it specifically includes:

[0147] Algorithm sorting: First, all image processing algorithms are sorted according to a preset logic or order (such as algorithm complexity, frequency of use, etc.) to establish an orderly algorithm sequence. This step ensures the consistency and predictability of information organization.

[0148] Create and fill a data table: Create a data table with multiple rows and columns. Each row is used to store the identification information of an image processing algorithm, and each character in the identification information is assigned to a separate cell. The purpose of this is to accurately locate and operate the information of a specific algorithm later.

[0149] Target row identification: In the second identification data table, identify the row where the identification information of the target image processing algorithm is located and mark it as the third target row.

[0150] Mathematical operations on numeric characters: extract all numeric characters from the third target row and calculate the mean and sum of these numbers. These two statistics will be used to generate a geometric element - a linear line.

[0151] Generation and rotation of linear straight line: Based on the mean and sum of the digital characters in the third target row, a linear straight line is drawn in the coordinate system. This linear straight line reflects the spatial relationship of the digital features. Subsequently, the straight line is rotated 90 degrees clockwise to simulate a geometric transformation, which increases the complexity of password generation. In this embodiment, a linear straight line is generated in the coordinate system based on the mean and sum, and a variety of common methods in mathematics can be used to generate it. Specifically, in one embodiment, in a two-dimensional coordinate system, the mean and sum are used as the x-axis and y-axis of a coordinate point respectively, and then the coordinate point is connected to the origin of the coordinate system to obtain the linear straight line. In other embodiments, it is also possible to create a first coordinate point and a second coordinate point, the x-axis and y-axis of the first coordinate point are both the mean, and the x-axis and y-axis of the second coordinate point are both the sum, and then the first coordinate point and the second coordinate point are connected to obtain a linear straight line. There are many ways, which will not be listed here.

[0152] Interaction between the linear line and the data table: The rotated line is translated to the upper left corner of the second marked data table so that the line passes through the data table. This process is actually looking for a mapping relationship in the physical space to determine the target cell.

[0153] Target cell selection and character combination: Identify all cells that intersect the translated line, which are considered target cells. Extract characters from these target cells and combine them in the order they appear in the data table, and the encrypted password is finally obtained.

[0154] In summary, the above method combines the principles of informatics and geometry to not only improve the abstraction level of the encryption process, but also introduces nonlinear transformations, making the password more difficult to predict or crack by brute force. Its uniqueness lies in:

[0155] Multi-dimensional transformation: From character information to mathematical statistics, and then to geometric figures, multi-dimensional information encoding is achieved.

[0156] Dynamic adaptability: Even if the algorithm identification information changes, it can still generate valid and unique passwords by adjusting the preset sequence or selecting different statistical parameters.

[0157] Visual password: The introduction of linear lines gives the password generation process certain visualization characteristics, which may help to understand and verify the correctness of the encryption process in some application scenarios.

[0158] In general, the above technical solution provides an innovative and highly customized encryption method, which is suitable for image processing algorithm identification information protection scenarios that require high security.

[0159] In one embodiment, the step of performing feature image recognition and cropping on the target image to obtain a plurality of feature images includes:

[0160] Identifying a plurality of characteristic targets in the target image based on a target detection algorithm;

[0161] A bounding box is added to each of the feature targets, and the images are cropped according to the edges of the bounding box to obtain multiple feature images.

[0162] In this embodiment, it specifically includes:

[0163] The target detection algorithm is used to identify multiple characteristic targets in the target image. The target detection algorithm is a computer vision technology that can identify specific objects in an image and locate their positions. Common target detection algorithms include but are not limited to YOLO (You Only Look Once), SSD (Single Shot MultiBox Detector), Faster R-CNN, etc. These algorithms learn to extract features from images through deep learning networks, such as convolutional neural networks (CNN), and then identify and locate target objects.

[0164] In industrial applications, these feature targets can be product defects, parts, barcodes, text, and any other objects that need to be detected and analyzed. The target detection algorithm identifies these feature targets in the image and adds a bounding box to it, that is, a rectangular box is used to encircle the location of the target, providing guidance for the next step of cropping.

[0165] After identifying the feature target and adding a bounding box to it, the target image will be cropped according to the edge of the bounding box to obtain multiple feature images containing only the specific feature target. This process helps to remove irrelevant background information and reduce the amount of calculation for subsequent processing. It also helps to enhance the clarity and contrast of the target features and improve the accuracy of subsequent image processing.

[0166] The cropped feature image is usually more compact and more focused on the target area, which is very important for improving the efficiency and accuracy of image processing algorithms. Especially in industrial environments, fast and accurate processing of large amounts of image data is essential for real-time monitoring and quality control.

[0167] By combining object detection and bounding box cropping, this technical solution can efficiently extract key features from complex images without sacrificing information integrity, providing high-quality input for subsequent image processing and analysis, thereby achieving more accurate industrial automation detection and control.

[0168] In one embodiment, the identifying multiple characteristic targets in the target image based on the target detection algorithm includes:

[0169] Obtaining an association code of an industrial camera, and obtaining a plurality of target detection algorithms corresponding to the association code from a database; wherein each target detection algorithm detects a different target;

[0170] A plurality of target images are copied, and a target detection algorithm is used to identify characteristic targets for each target image.

[0171] In this embodiment, it specifically includes:

[0172] First, obtain the association code of the industrial camera. This association code is similar to an identifier, which is associated with a specific industrial camera and its preset configuration information, including but not limited to camera model, resolution, application field, etc. Through this association code, the system can call the matching settings and algorithms from the database.

[0173] Next, based on the association code of the industrial camera, a series of various target detection algorithms related to the camera are retrieved from the database. These algorithms cover different detection targets. For example, some algorithms are specifically used to detect barcodes or QR codes on products, while others may focus on detecting the shape or surface defects of parts. The algorithm library in the database is pre-trained and verified to meet the detection needs of industrial cameras in different scenarios.

[0174] Copying multiple target images means that the original image will be copied multiple times, and each image will be processed independently by a different target detection algorithm. This has several advantages:

[0175] Comprehensiveness: By applying multiple target detection algorithms in parallel, it can ensure that no type of feature target is missed, thus improving the comprehensiveness and reliability of detection.

[0176] Efficiency: Although duplicating images in parallel may seem like an increased computational burden, with the support of modern multi-core processors and GPU acceleration, this parallel processing can actually significantly speed up processing, especially when processing high-resolution images or when real-time feedback is required.

[0177] Adaptability: Different algorithms may perform well on different types of feature targets. By applying them in parallel, the system can automatically select the most effective algorithm for a specific target, thereby optimizing the overall detection performance.

[0178] Through the above steps, the above technical solution can ensure efficient, comprehensive and accurate feature target recognition of target images in complex and changing industrial environments, providing a solid foundation for subsequent image analysis and decision-making. This method based on the application of association codes and parallel algorithms reflects the pursuit of intelligence and adaptability in modern industrial automation systems.

[0179] Reference Figure 2 In another embodiment of the present invention, an intelligent optimization system for an image processing algorithm of an industrial camera is provided, comprising:

[0180] An acquisition unit is used to acquire a target image based on an industrial camera, perform feature image recognition and cropping on the target image to obtain multiple feature images; obtain image processing requirement information of the industrial camera, and match multiple image processing algorithms corresponding to the image processing requirement information in a database;

[0181] A processing unit, used to synchronously process the characteristic images based on the respective image processing algorithms to obtain processing effects of the respective image processing algorithms on the characteristic images; wherein the processing effects include processing speed and accuracy;

[0182] The optimization unit is used to determine the algorithm with the best processing effect from various image processing algorithms as the target image processing algorithm; obtain the device status of the industrial camera and the lighting status of the current environment; and optimize the algorithm parameters of the target image processing algorithm based on the device status and lighting status.

[0183] In this embodiment, for the specific implementation of each unit in the above system embodiment, please refer to the above method embodiment, which will not be repeated here.

[0184] Reference Figure 3 In an embodiment of the present invention, a computer device is also provided. The computer device may be a server, and its internal structure may be as follows: Figure 3As shown. The computer device includes a processor, a memory, a display screen, an input device, a network interface and a database connected through a system bus. Among them, the processor designed by the computer is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the corresponding data in this embodiment. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the above method is implemented.

[0185] Those skilled in the art will understand that Figure 3 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied.

[0186] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the above method is implemented. It can be understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0187] In summary, the intelligent optimization method and system of the image processing algorithm of the industrial camera provided in the embodiment of the present invention include: based on the target image collected by the industrial camera, the feature image recognition and cropping of the target image are performed to obtain multiple feature images; the image processing requirement information of the industrial camera is obtained, and multiple image processing algorithms corresponding to the image processing requirement information are matched in the database; the feature image is synchronously processed based on each of the image processing algorithms to obtain the processing effect of each image processing algorithm on the feature image; wherein the processing effect includes processing speed and accuracy; the algorithm with the best processing effect is determined from each image processing algorithm as the target image processing algorithm; the device status of the industrial camera and the lighting status of the current environment are obtained; the algorithm parameters of the target image processing algorithm are optimized based on the device status and lighting status. In the present invention, the most suitable image processing algorithm is automatically selected, and the changes in the actual device status and lighting conditions are considered at the same time, and the algorithm parameters are adjusted in real time to ensure that the best image processing effect can be achieved under various conditions. The above technical solution overcomes the defect that the algorithm cannot be adjusted for different application scenarios and environmental conditions.

[0188] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media provided by the present invention and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double-speed data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM.

[0189] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the presence of other identical elements in the process, device, article or method including the element.

[0190] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. An intelligent optimization method for image processing algorithms of industrial cameras, characterized in that: The following steps are involved: Based on the industrial camera, the target image is collected, and the feature image is recognized and cropped to obtain multiple feature images; the image processing requirement information of the industrial camera is obtained, and multiple image processing algorithms corresponding to the image processing requirement information are matched in the database; Based on each of the image processing algorithms, the characteristic images are synchronously processed to obtain processing effects of each of the image processing algorithms on the characteristic images; wherein the processing effects include processing speed and accuracy; Determine the algorithm with the best processing effect from various image processing algorithms as the target image processing algorithm; obtain the device status of the industrial camera and the lighting status of the current environment; optimize the algorithm parameters of the target image processing algorithm based on the device status and lighting status; Processing the target image based on the optimized target image processing algorithm to obtain a processed image, and saving the processed image to a preset folder; Acquire identification information of each of the image processing algorithms; generate an encryption password based on the identification information of each of the image processing algorithms; encrypt the preset folder based on the encryption password; Based on the identification information of the target image processing algorithm, an identification character is generated, and the identification character is used as the folder name of the preset folder; The step of generating an encrypted password based on the identification information of each of the image processing algorithms comprises: Sorting each of the image processing algorithms based on a preset order to obtain an algorithm sequence; creating a data table; wherein the data table includes cells of multiple rows and columns; According to the algorithm sequence, the identification information of each of the image processing algorithms is sequentially added to the data table to obtain a first identification data table; wherein the identification information of one image processing algorithm is added in a row, and one character in the identification information is added in one cell; Rearranging the first identification data table to obtain a rearranged data table; Each character in the identification information of the target image processing algorithm is used as a target character; in the rearrangement data table, all characters identical to the target character are removed to obtain a removed character table; In the eliminated character table, the number of cells with characters in each row is obtained; the number of cells with characters in each row is combined in sequence to obtain a combined number as the encrypted password.

2. The intelligent optimization method for image processing algorithm of industrial camera according to claim 1, characterized in that: The method of acquiring a target image based on an industrial camera includes: The camera module based on the industrial camera obtains the preview image of the target object and enhances the preview image in real time; The enhancement process includes: By integrating ray tracing technology, multiple light sources are simulated to illuminate the target object; Use optical flow algorithm to estimate the motion trajectory of the target object to capture the target object; Virtual fluid dynamics is used to simulate the flow effects on the surface of objects and dynamically enhance image features.

3. The intelligent optimization method for image processing algorithm of industrial camera according to claim 1, characterized in that: The image processing requirement information includes image defect detection and object recognition.

4. The method for intelligent optimization of image processing algorithms for industrial cameras according to claim 1, characterized in that: The optimizing the algorithm parameters of the target image processing algorithm based on the device state and the illumination state includes: In a database, matching the optimization value corresponding to the device state and the illumination state; wherein the database stores a mapping relationship between the device state, the illumination state and the optimization value; The algorithm parameters of the target image processing algorithm are optimized based on the optimization value; wherein the optimization value includes the adjustment amount of each algorithm parameter.

5. The method for intelligent optimization of image processing algorithms for industrial cameras according to claim 1, characterized in that: The step of performing feature image recognition and cropping on the target image to obtain a plurality of feature images includes: Identifying a plurality of characteristic targets in the target image based on a target detection algorithm; A bounding box is added to each of the feature targets, and the images are cropped according to the edges of the bounding box to obtain multiple feature images.

6. The method for intelligent optimization of image processing algorithms for industrial cameras according to claim 5, characterized in that: The target detection algorithm is used to identify multiple characteristic targets in the target image, including: Obtaining an association code of an industrial camera, and obtaining a plurality of target detection algorithms corresponding to the association code from a database; wherein each target detection algorithm detects a different target; A plurality of target images are copied, and a target detection algorithm is used to identify characteristic targets for each target image.

7. An intelligent optimization system for image processing algorithms of industrial cameras, characterized in that: include: An acquisition unit, used for acquiring a target image based on an industrial camera, performing feature image recognition and cropping on the target image to obtain a plurality of feature images; Obtain image processing requirement information of an industrial camera, and match multiple image processing algorithms corresponding to the image processing requirement information in a database; A processing unit, used to synchronously process the characteristic images based on the respective image processing algorithms to obtain processing effects of the respective image processing algorithms on the characteristic images; wherein the processing effects include processing speed and accuracy; An optimization unit, used to determine the algorithm with the best processing effect from various image processing algorithms as the target image processing algorithm; obtain the device status of the industrial camera and the lighting status of the current environment; and optimize the algorithm parameters of the target image processing algorithm based on the device status and the lighting status; Processing the target image based on the optimized target image processing algorithm to obtain a processed image, and saving the processed image to a preset folder; Acquire identification information of each of the image processing algorithms; generate an encryption password based on the identification information of each of the image processing algorithms; encrypt the preset folder based on the encryption password; Based on the identification information of the target image processing algorithm, an identification character is generated, and the identification character is used as the folder name of the preset folder; The step of generating an encrypted password based on the identification information of each of the image processing algorithms comprises: Sorting each of the image processing algorithms based on a preset order to obtain an algorithm sequence; creating a data table; wherein the data table includes cells of multiple rows and columns; According to the algorithm sequence, the identification information of each of the image processing algorithms is sequentially added to the data table to obtain a first identification data table; wherein the identification information of one image processing algorithm is added in a row, and one character in the identification information is added in one cell; Rearranging the first identification data table to obtain a rearranged data table; Each character in the identification information of the target image processing algorithm is used as a target character; in the rearrangement data table, all characters identical to the target character are removed to obtain a removed character table; In the eliminated character table, the number of cells with characters in each row is obtained; the number of cells with characters in each row is combined in sequence to obtain a combined number as the encrypted password.

Images