A cloud platform-based image data bag intelligent identification method, system and medium

Abstract

The application relates to a cloud platform-based image data bag intelligent identification method, a cloud platform-based image data bag intelligent identification system and a medium. The cloud platform-based image data bag intelligent identification method comprises the following steps: before capturing an image of an identification code of an image data bag, acquiring instantaneous illumination information of a current lighting environment; a cloud platform acquires the instantaneous illumination information, and quantifies light strobe characteristics and spectral characteristics of the current lighting environment based on the instantaneous illumination information; the cloud platform calculates imaging parameters of an industrial camera based on the quantified light strobe characteristics and the quantified spectral characteristics, wherein the imaging parameters comprise an exposure time, a gain and a white balance; the camera settings of the industrial camera are adjusted according to the imaging parameter instructions, and the industrial camera after the adjustment is used to capture the image of the identification code of the image data bag. The application has high adaptability, high identification accuracy and excellent operation efficiency.

Description

Technical Field

[0001] This application relates to the field of image document bag recognition, and in particular to an intelligent image document bag recognition method, system and medium based on a cloud platform. Background Technology

[0002] In modern industrial warehousing scenarios, automated intelligent identification and rapid extraction systems for medical imaging data bags are becoming increasingly common. These systems need to process thousands or even tens of thousands of physical data bags daily, placing stringent requirements on their ability to operate stably for extended periods at high frequencies. Industrial cameras play a crucial role in rapidly acquiring high-definition images of the data bags' identification codes / tags. The accuracy of their continuous, high-speed acquisition directly impacts the efficiency of subsequent robotic arm sorting and extraction; any delays or misjudgments could potentially affect the progress of medical procedures and data integrity.

[0003] Upgrading and retrofitting warehousing infrastructure has become standard practice in the industry to improve energy efficiency and reduce operating costs. While upgrading the existing lighting system in a medical imaging data bag storage center to use new high-intensity discharge lamps or LED arrays can significantly optimize energy savings and extend lifespan, it may also introduce unforeseen technical challenges: some new lamps (including LEDs using pulse width modulation (PWM) technology) may produce flickering at specific frequencies during startup and stable operation (these flickering frequencies are beyond the range of human perception and are imperceptible), and their spectral composition differs significantly from the original lighting system.

[0004] Such changes to the lighting system can have a critical impact on the imaging performance of industrial cameras. To reduce motion blur in data bags, industrial cameras typically use a high-speed global shutter mode. If the illumination flicker frequency is an integer or near-integer multiple of the camera's frame rate / shutter frequency, or if the camera's exposure time is not precisely synchronized with the light cycle, it can easily induce "stripes" and "flicker artifacts," resulting in uneven stripes in the acquired image. At the same time, changes in the spectral characteristics of new light sources may alter the optical reflectivity of the label material, turning the original diffuse reflection into strong specular reflection, or damaging color reproduction, reducing the local contrast of printed characters / barcode edges, and increasing the difficulty of distinguishing the foreground (label information) from the background (label substrate).

[0005] Image quality degradation directly causes local recognition algorithms to fail. These algorithms typically rely on preprocessing steps such as binarization and edge extraction with fixed parameters. However, bright and dark stripes in the image can cause global threshold failure, leading to information loss or noise interference. Decreased local contrast and flicker artifacts can interfere with edge detection accuracy, resulting in misidentification of the identification code. If the label itself has printing defects or is worn due to long-term circulation, these problems will be further exacerbated, ultimately leading to a significant decline in overall recognition accuracy.

[0006] The decline in recognition accuracy directly leads to a significant decrease in extraction efficiency: after the first recognition failure, the system will initiate a retry process, including camera repositioning, image reacquisition, and repeated grasping by the robotic arm, which will extend the extraction cycle of a single data bag. In high-throughput warehousing scenarios that process thousands or tens of thousands of data bags daily, the cumulative effect of such delays can easily create operational bottlenecks and make it difficult to achieve the preset service level agreement (SLA). In addition, repeated operations will accelerate the wear of the robotic arm's end effector. The continuous driving, grasping, and releasing actions will apply additional stress to the motor, gears, and grasping contact surfaces; in extreme cases, multiple grasping failures may also cause creases, tears, or even displacement of internal imaging films in the data bag, compromising the integrity of the medical data.

[0007] There is room for further improvement in the existing technological system to address the above issues. Summary of the Invention

[0008] This application provides a cloud-based intelligent image data bag recognition method, system, and medium to at least solve the problem in the prior art where the replacement of industrial warehouse lighting systems may lead to a deterioration in the imaging quality of industrial cameras, thereby affecting the accuracy and extraction efficiency of image data bag recognition.

[0009] In a first aspect, this application provides a cloud-based intelligent image portfolio recognition method, comprising the following steps:

[0010] Before capturing the image of the image data bag identification code, obtain the instantaneous illumination information of the current lighting environment;

[0011] The cloud platform acquires the instantaneous illumination information and, based on the instantaneous illumination information, quantifies the light flicker characteristics and spectral characteristics of the current lighting environment;

[0012] The cloud platform calculates the imaging parameters of the industrial camera based on quantized light flicker characteristics and quantized spectral characteristics, wherein the imaging parameters include exposure time, gain, and white balance.

[0013] The camera settings of the industrial camera are adjusted according to the imaging parameter instructions, and the adjusted industrial camera is used to capture the identification code image of the image data bag.

[0014] Optionally, the cloud platform acquires the instantaneous illumination information and, based on the instantaneous illumination information, quantifies the light flicker characteristics and spectral characteristics of the current lighting environment, including:

[0015] The cloud platform receives the instantaneous illumination information, divides the instantaneous illumination information into regions, and obtains the partitioned instantaneous illumination information of multiple illumination regions;

[0016] The instantaneous illumination information of each partition is quantized for flicker characteristics and spectral characteristics to obtain quantization results for multiple illumination regions.

[0017] Based on the degree of influence of each local illumination region on the image data bag identification code imaging, the quantization results of the multiple local illumination regions are weighted and fused to obtain the dominant light flicker characteristics and dominant spectral characteristics.

[0018] The dominant light flicker characteristics and the dominant spectral characteristics are used as the light flicker characteristics and spectral characteristics of the current lighting environment.

[0019] Optionally, the cloud platform calculates the imaging parameters of the industrial camera based on quantized light flicker characteristics and quantized spectral characteristics, including:

[0020] Identify the tag characteristics of the image data bag identification code;

[0021] Based on the characteristics of the tag, select an imaging parameter adjustment strategy;

[0022] By combining the quantized light flicker characteristics, the quantized spectral characteristics, and the imaging parameter adjustment strategy, the exposure time, gain, and white balance parameters of the industrial camera are calculated.

[0023] Optionally, adjusting the camera settings of the industrial camera according to the imaging parameter instructions, and capturing the identification code image of the image data bag with the adjusted industrial camera, includes:

[0024] Based on real-time motion data of the image data bag, the exposure path of the image data bag within the exposure window period of the industrial camera's recognition area is predicted;

[0025] Based on the dynamically updated local light field spatial distribution map, the instantaneous changes in light intensity, flicker frequency, and spectral composition that the image data bag may experience along the exposure path are predicted.

[0026] Based on the predicted results, calculate the industrial camera parameter sequence to dynamically compensate for the instantaneous changes in illumination;

[0027] The industrial camera parameter sequence is preloaded into the local cache of the industrial camera;

[0028] According to preset triggering conditions, when the image data bag moves through the exposure path, the industrial camera automatically switches to the corresponding industrial camera parameters in the industrial camera parameter sequence to capture the image data bag identification code image.

[0029] Optionally, the step of using the dominant ray flicker characteristics and the dominant spectral characteristics as the ray flicker characteristics and spectral characteristics of the current lighting environment to quantify the current lighting environment includes:

[0030] The cloud platform receives the instantaneous light information, identifies the target area of ​​the instantaneous light information, and determines the main identification area where the image data bag identification code label is located;

[0031] The instantaneous illumination information of the main identification area is quantized using stroboscopic and spectral characteristics to obtain the quantization result of the main identification area.

[0032] The quantization results of the main identifier region are corrected to obtain the corrected dominant spectral flicker characteristics and dominant spectral characteristics.

[0033] Optionally, the step of using the dominant ray flicker characteristics and the dominant spectral characteristics as the ray flicker characteristics and spectral characteristics of the current lighting environment further includes:

[0034] The cloud platform receives the instantaneous illumination information, performs high-frequency sampling on the instantaneous illumination information, and obtains a continuous instantaneous illumination data stream;

[0035] The instantaneous illumination data stream is analyzed in real time to identify the start and end points of changes in illumination characteristics;

[0036] The length and position of the quantization window are dynamically adjusted based on the start and end points.

[0037] Within the dynamically adjusted quantization window, the instantaneous illumination information is quantized for flicker and spectral characteristics to obtain the adjusted flicker and spectral characteristics of the current lighting environment.

[0038] Optionally, the cloud platform acquires the instantaneous illumination information and, based on the instantaneous illumination information, quantifies the light flicker characteristics and spectral characteristics of the current lighting environment, including:

[0039] The cloud platform receives the instantaneous illumination information and obtains the lighting device identifier contained in the instantaneous illumination information;

[0040] Based on the lighting equipment identifier, query the operating time, ambient temperature, and ambient humidity data of the lighting equipment, and evaluate the luminous efficiency and spectral output drift of the lighting equipment based on the operating time, ambient temperature, and ambient humidity data;

[0041] The instantaneous illumination information is quantified based on its stroboscopic and spectral characteristics.

[0042] The quantization results are compensated based on the luminous efficiency and the spectral output drift to obtain the light flicker characteristics and spectral characteristics of the current lighting environment.

[0043] Optionally, the cloud platform acquires the instantaneous illumination information and, based on the instantaneous illumination information, quantifies the light flicker characteristics and spectral characteristics of the current lighting environment, including:

[0044] The cloud platform receives the instantaneous illumination information, performs frequency domain analysis on the instantaneous illumination information, and identifies the light flicker components and spectral components of multiple frequency bands as sample components;

[0045] Sample components from multiple frequency bands are imported into a pre-defined non-lighting equipment interference feature library to identify and obtain non-associated flicker components and non-associated spectral components related to non-lighting equipment.

[0046] From the sample components of multiple frequency bands, the unrelated stroboscopic components and the unrelated spectral components are stripped to obtain the pure sample light components of multiple frequency bands;

[0047] The light components of pure samples in multiple frequency bands are quantified to obtain the flicker characteristics and spectral characteristics of the current lighting environment.

[0048] Secondly, this application provides an intelligent recognition system for image document bags, the system comprising:

[0049] The illumination information acquisition module is used to acquire instantaneous illumination information of the current lighting environment through an industrial camera before capturing the image of the image data bag identification code.

[0050] The illumination characteristic quantization module is used to acquire the instantaneous illumination information and, based on the instantaneous illumination information, quantify the light flicker characteristics and spectral characteristics of the current lighting environment;

[0051] An imaging parameter calculation module is used to calculate the imaging parameters of the industrial camera based on quantized light flicker characteristics and quantized spectral characteristics, wherein the imaging parameters include exposure time, gain, and white balance.

[0052] The imaging parameter adjustment module is used to adjust the camera settings of the industrial camera according to the imaging parameter instructions, and to capture the identification code image of the image data bag with the adjusted industrial camera.

[0053] Thirdly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method provided in the first aspect above.

[0054] Compared with related technologies, the cloud-based image portfolio intelligent recognition method, system, and medium provided in this application have at least the following technical advantages:

[0055] The method acquires instantaneous illumination information of the current lighting environment and transmits this information to a cloud platform. Based on the received instantaneous illumination information, the cloud platform can accurately quantify the flicker characteristics and spectral characteristics of the current lighting environment. Furthermore, the cloud platform calculates the imaging parameters of the industrial camera and adjusts the camera settings according to the calculated imaging parameters, ultimately capturing the identification code image of the image data bag using the adjusted industrial camera. The method provided in this application can dynamically sense and adapt to complex changes in the lighting environment. By adjusting the exposure time, gain, and white balance of the industrial camera in real time, it ensures the capture of high-quality, artifact-free, and high-contrast identification code images under different lighting conditions, significantly improving the overall recognition accuracy of the image data bag. Simultaneously, the improved recognition accuracy shortens the extraction cycle of a single image data bag, thereby improving the overall rapid extraction efficiency, ultimately forming a highly adaptable, highly accurate, and efficient intelligent recognition solution.

[0056] Details of one or more embodiments of this application are set forth in the following drawings and description to make other features, objects and advantages of this application more readily apparent. Attached Figure Description

[0057] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0058] Figure 1 This is a flowchart illustrating a cloud-based image data bag intelligent recognition method according to an exemplary embodiment.

[0059] Figure 2 This is a flowchart illustrating step S2 according to an exemplary embodiment.

[0060] Figure 3 This is a flowchart illustrating step S3 according to an exemplary embodiment.

[0061] Figure 4 This is a flowchart illustrating step S4 according to an exemplary embodiment.

[0062] Figure 5 This is a flowchart illustrating step S2 according to another exemplary embodiment.

[0063] Figure 6 This is a block diagram illustrating a cloud-based image portfolio intelligent recognition system according to an exemplary embodiment. Detailed Implementation

[0064] To make the objectives, technical solutions, and advantages of this application clearer, the application is described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments provided in this application without inventive effort are within the scope of protection of this application.

[0065] In modern industrial warehousing environments, particularly in the management of medical imaging data bags, automated intelligent identification and rapid extraction systems face significant challenges. Traditional industrial cameras, when capturing images of identification codes on imaging data bags, suffer from image quality degradation due to the complexity and variability of the lighting environment, especially the flicker and spectral variations introduced by new energy-saving lighting systems. This degradation affects identification accuracy and system efficiency. Failure to address these issues will lead to increased error rates, more repetitive operations, reduced overall processing efficiency, accelerated equipment wear, and even physical damage to the imaging data bags themselves, compromising the integrity of medical data.

[0066] Based on the above, embodiments of the present invention provide a cloud-based intelligent identification method, system, and medium for image data bags, which will be described in detail below with reference to specific embodiments and accompanying drawings.

[0067] Example 1

[0068] This invention provides a cloud-based intelligent identification method for image data bags. Figure 1 This is a flowchart illustrating a cloud-based image portfolio intelligent recognition method according to an exemplary embodiment. Figure 1 As shown, the method includes the following steps:

[0069] S1. Before capturing the image of the image data bag identification code, obtain the instantaneous illumination information of the current lighting environment;

[0070] In this embodiment, the instantaneous illumination information of the current lighting environment refers to real-time data about the lighting environment captured by an industrial camera or other light sensor at a specific point in time. Instantaneous illumination information includes at least light intensity, color temperature, and spectral distribution, used to reflect the current lighting conditions in real time. In this embodiment, the industrial camera has a built-in light sensor. Before each image capture, the light sensor performs a rapid illumination sampling, converting the collected raw light signal into digitized instantaneous illumination information. For example, the light sensor is a broadband photodiode array capable of simultaneously measuring light intensity within different wavelength ranges, thus providing relatively comprehensive spectral information. It is understood that the industrial camera can also be configured with an independent auxiliary illumination sensor module. This module works synchronously with the industrial camera. When the industrial camera is preparing to capture an image, the auxiliary illumination sensor module actively collects ambient light data and transmits this data to the industrial camera or directly to a cloud platform. For example, this auxiliary illumination sensor module is an integrated ambient light sensor capable of outputting light intensity (lux), color temperature (Kelvin), and preliminary flicker detection results.

[0071] S2. The cloud platform acquires instantaneous illumination information and, based on this information, quantifies the light flicker characteristics and spectral characteristics of the current lighting environment.

[0072] In this embodiment, light flicker characteristics refer to the brightness fluctuation characteristics of a lighting source over time, typically quantified by parameters such as flicker frequency and modulation depth. Spectral characteristics refer to the energy distribution of a light source at different wavelengths, i.e., the color composition of light, usually described by spectral power distribution curves or parameters such as color temperature and color rendering index. After receiving instantaneous illumination information, the cloud platform in this embodiment uses signal processing techniques such as Fourier transform to perform frequency domain analysis on the illumination intensity data, thereby identifying the frequency and amplitude of light flicker and quantifying them as light flicker characteristic parameters. Simultaneously, by analyzing the spectral data in the instantaneous illumination information, the spectral characteristic parameters such as color temperature and color rendering index of the current lighting environment are calculated. For example, the cloud platform can utilize a preset spectral database to compare the collected spectral data with the spectral characteristics of known light sources, thereby more accurately quantifying spectral characteristics.

[0073] S3, the cloud platform calculates the imaging parameters of industrial cameras based on quantized light flicker characteristics and quantized spectral characteristics. The imaging parameters include exposure time, gain and white balance.

[0074] In this embodiment, imaging parameters are key parameters that industrial cameras need to set when capturing images. These include exposure time, which determines the length of time the photosensitive element receives light; gain, which adjusts the amplification of the image signal to enhance brightness; and white balance, which corrects image color deviations to make them closer to true colors. The cloud platform in this embodiment has a built-in parameter calculation module. This module calculates the optimal exposure time, gain, and white balance parameters based on preset rules and algorithms, using quantized light flicker characteristics and spectral characteristics as input. For example, if high-frequency flicker is detected, the system may prioritize a shorter exposure time to avoid stripe effects; if a cool spectrum is detected, the white balance parameters will be adjusted accordingly to correct the color. In another implementation, the cloud platform can also employ a model-based optimization method. This optimization method uses a model simulating the imaging process of an industrial camera, using quantized light flicker characteristics and spectral characteristics as environmental input, and through iterative optimization algorithms, finds the optimal combination of imaging parameters that achieves the best image quality indicators (such as contrast and signal-to-noise ratio). For example, this model is a simulator built based on physical optics principles, capable of predicting image performance under specific lighting conditions with different imaging parameters.

[0075] S4. Adjust the camera settings of the industrial camera according to the imaging parameter instructions, and capture the identification code image of the image data bag with the adjusted industrial camera.

[0076] In this embodiment, the industrial camera has a pre-defined parameter adjustment strategy library. When the instruction sent by the cloud platform contains a specific strategy identifier, the industrial camera selects and executes the corresponding parameter adjustment process from the local strategy library based on that identifier. For example, if the instruction indicates the use of "high-frequency flash suppression mode," the industrial camera may automatically enable a series of pre-defined exposure time fine-tuning and image post-processing algorithms.

[0077] The technical solution of the above embodiments introduces a cloud platform for intelligent quantification of instantaneous illumination information and adaptive calculation and adjustment of imaging parameters. Before capturing the image of the image data bag identification code, the industrial camera acquires the instantaneous illumination information of the current lighting environment and uploads this information to the cloud platform for quantification of light flicker characteristics and spectral characteristics. Based on these quantification results, the cloud platform intelligently calculates the optimal imaging parameters for the industrial camera, including exposure time, gain, and white balance. Subsequently, the cloud platform sends these imaging parameter instructions to the industrial camera, which adjusts its own settings according to the instructions and captures the image data bag identification code image with the optimized parameters. This series of steps forms a closed-loop adaptive adjustment system, enabling the industrial camera to respond in real time to changes in the lighting environment, thereby consistently capturing high-quality, artifact-free image data bag identification code images under various complex lighting conditions. Therefore, when the lighting system is replaced with new LED lamps, causing changes in light flicker and spectral composition, this method can promptly detect these changes and automatically adjust the camera's exposure time to avoid stripe effects and adjust the white balance to correct color deviations, ensuring the clarity and accuracy of the identification code image. Another understandable point is that the centralized processing capabilities of the cloud platform enable unified management and optimization of multiple industrial cameras, further improving the robustness and efficiency of the entire system.

[0078] In one possible design, Figure 2 This is a flowchart illustrating step S2 according to an exemplary embodiment. (Refer to the attached document.) Figure 2 Step S2 includes:

[0079] S201. The cloud platform receives instantaneous illumination information, divides the instantaneous illumination information into regions, and obtains the partitioned instantaneous illumination information of multiple illumination regions.

[0080] In this embodiment, the instantaneous illumination information refers to real-time data on light intensity, color, flicker, and other characteristics acquired by the industrial camera from the current lighting environment before capturing the image of the image data bag identification code. Dividing the instantaneous illumination information into regions involves segmenting the entire lighting scene into several smaller local areas with relatively uniform illumination characteristics. For example, image segmentation algorithms (such as thresholding, edge detection, or cluster analysis) can be used to identify different illumination regions, or a preset grid division method can be used to more finely analyze non-uniform lighting environments.

[0081] S202. Quantify the flicker characteristics and spectral characteristics of the instantaneous illumination information of each partition to obtain the quantization results of multiple illumination areas;

[0082] In this embodiment, the fluctuation frequency and color composition distribution of light are independently analyzed for each locally illuminated area. The flicker characteristics are quantified using frequency domain analysis methods such as Fourier transform of the light intensity over time, for example, identifying the dominant flicker frequency and modulation depth. Spectral characteristics are quantified by analyzing the energy distribution of light at different wavelengths, for example, by calculating color temperature, color rendering index, or the relative intensity of specific wavelength bands. The detailed illumination characteristics of each local area are then obtained from the quantification results.

[0083] S203. Based on the degree of influence of each local illumination area on the image data bag identification code imaging, the quantization results of multiple local illumination areas are weighted and fused to obtain the dominant light flicker characteristics and dominant spectral characteristics.

[0084] In this embodiment, when comprehensively considering the illumination characteristics of all local areas, areas with a greater impact on the imaging of the identification code are given higher weights. For example, areas closer to the identification code label, areas with higher illumination intensity, or areas whose historical data shows a greater impact on the recognition success rate will be assigned higher weights in their quantification results. The assessment of the degree of influence can be based on spatial distance, illumination intensity, contrast analysis, or machine learning models. This ensures that the final quantified light flicker characteristics and spectral characteristics can most accurately reflect the actual illumination conditions of the area where the identification code is located.

[0085] S204. The dominant light flicker characteristics and dominant spectral characteristics are used as the light flicker characteristics and spectral characteristics for quantifying the current lighting environment;

[0086] In this embodiment, through the above-mentioned process of region division, independent quantization and weighted fusion, a light flicker characteristic and spectral characteristic that can represent the most significant impact on the imaging of identification codes in the entire complex lighting environment are finally obtained.

[0087] In the technical solution of the above embodiments, the instantaneous illumination information is divided into regions, and the flicker characteristics and spectral characteristics of each region are quantified independently, thereby addressing the non-uniform lighting problem commonly found in actual industrial scenarios. By introducing a weighted fusion mechanism, intelligent evaluation and weight allocation can be performed based on the degree of influence of different local illumination regions on the identification code imaging, making the final quantized dominant light flicker characteristics and dominant spectral characteristics more targeted and representative, thus avoiding quantization deviations caused by interference from the illumination characteristics of non-critical areas. Ultimately, this improves the accuracy and robustness of quantifying the flicker characteristics and spectral characteristics of light under complex, non-uniform lighting environments.

[0088] In one possible design, Figure 3 This is a flowchart illustrating step S3 according to an exemplary embodiment. (Refer to the attached diagram.) Figure 3 Step S3 includes:

[0089] S301, Identifying the label characteristics of the image data bag identification code;

[0090] In this embodiment, the physical and optical properties of the identification code label are determined through image analysis, sensor detection, or other preset information acquisition methods. Label characteristics include, but are not limited to, the label's material (e.g., paper, plastic, metal), surface treatment (e.g., matte, glossy, frosted), color, contrast, size, printing precision, and the presence of reflective or transparent overlays. For example, label characteristics can be determined by pre-scanning and analyzing the identification code image to extract features such as texture, color histogram, and reflectivity.

[0091] S302. Select an imaging parameter adjustment strategy based on the tag characteristics;

[0092] In this embodiment, based on the identified tag characteristics, one or more imaging parameter adjustment schemes most suitable for the current tag are selected from a preset strategy library. For example, for highly reflective tags, a strategy of reducing gain and adjusting white balance to suppress highlight clipping and optimize color reproduction is selected; for low-contrast tags, a strategy of extending exposure time or increasing gain is selected to enhance image brightness and contrast. It is understood that the imaging parameter adjustment strategies can be pre-trained through experiments or machine learning and stored in the cloud platform's database.

[0093] S303, combining quantized light flicker characteristics, quantized spectral characteristics, and imaging parameter adjustment strategies, calculates the exposure time, gain, and white balance parameters of an industrial camera.

[0094] In this embodiment, the dynamic characteristics of ambient light (flicker and spectrum) are comprehensively considered along with the inherent characteristics of the tag and their corresponding adjustment strategies. A preset optimization algorithm or decision model is used to derive the optimal exposure time, gain, and white balance parameters. For example, a multi-objective optimization algorithm is employed to balance the influence between different parameters while satisfying objectives such as image sharpness, color accuracy, and recognition rate, thereby generating a set of imaging parameters best suited to the current environment and tag characteristics.

[0095] In the technical solution of the above embodiments, by introducing the identification of the characteristics of the image data bag identification code label and selecting the corresponding imaging parameter adjustment strategy accordingly, the recognition accuracy and robustness of the image data bag identification code are improved. Specifically, after the industrial camera acquires instantaneous illumination information, the cloud platform not only quantifies the light flicker characteristics and spectral characteristics, but also further identifies the inherent characteristics of the identification code label. Based on a deep understanding of the label characteristics, the system selects or generates the most suitable imaging parameter adjustment strategy for the label. For example, for labels with strong reflective characteristics, the system selects a strategy to suppress reflection; for labels with darker colors, it selects a strategy to enhance brightness. This dual consideration of ambient light and label characteristics allows the calculated exposure time, gain, and white balance parameters to more accurately adapt to actual imaging needs, thereby ensuring high-quality identification code images can be obtained in various complex scenarios.

[0096] In one example, suppose an automated production line needs to identify two different types of image data bag identification codes: one is an identification code printed on a matte paper label, and the other is an identification code printed on a reflective metal label under a transparent plastic film.

[0097] First, when the industrial camera captures an image of a document bag with a matte paper label, the cloud platform identifies the label's characteristics as "matte paper, low reflectivity." Based on this characteristic, the system selects an imaging parameter adjustment strategy that "enhances contrast." Simultaneously, by combining the quantified light flicker and spectral characteristics of the current environment, the system may calculate a relatively long exposure time, appropriate gain, and standard white balance to ensure clear image details and accurate color reproduction.

[0098] Subsequently, when the industrial camera captures an image of the data bag containing the reflective metallic tag, the cloud platform identifies its tag characteristics as "highly reflective and metallic." Based on this characteristic, the system selects an imaging parameter adjustment strategy that "suppresses highlights and reduces glare." Considering the same ambient lighting characteristics, the system may calculate a shorter exposure time, lower gain, and adjust the white balance to compensate for any color cast that might be caused by metallic reflection, thereby effectively avoiding overexposure and loss of detail in highlight areas and ensuring that the identification code remains readable even under strong reflective conditions.

[0099] In this way, the proposed solution can dynamically adjust imaging parameters according to the characteristics of different tags, thereby achieving the best recognition results in diverse practical application scenarios.

[0100] In one possible design, Figure 4 This is a flowchart illustrating step S4 according to an exemplary embodiment. (Refer to the attached document.) Figure 4 Step S4 includes:

[0101] S401. Based on real-time motion data of the image data bag, predict the exposure path of the image data bag within the exposure window of the industrial camera's recognition area.

[0102] In this embodiment, the real-time motion data of the image data bag is acquired by various sensors, such as encoders mounted on the conveyor belt, vision tracking systems, or laser rangefinders, to monitor the speed, position, and acceleration of the image data bag in real time. This data is used to accurately predict the motion trajectory of the image data bag within the industrial camera's recognition area, from entry to exit, over the entire exposure window period, i.e., the exposure path. The exposure window period refers to the total time required for the industrial camera to capture an image once, including the exposure time and the necessary image processing and transmission time.

[0103] S402. Based on the dynamically updated local light field spatial distribution map, predict the light intensity, flicker frequency, and instantaneous changes in spectral composition that the image data bag may experience along the exposure path.

[0104] In this embodiment, the dynamically updated local light field spatial distribution map refers to a data map formed by refining, real-time monitoring, and modeling the lighting environment within the industrial camera's recognition area. This distribution map is acquired in real time by multiple distributed light sensor arrays, or obtained through pre-calibration and periodic updates, to reflect the light intensity, flicker characteristics, and spectral composition at different locations within the recognition area. By combining the predicted exposure path with this local light field spatial distribution map, the instantaneous changes in light intensity, flicker frequency, and spectral composition that the identification tag may experience at different times and locations during the movement of the image data bag can be accurately predicted.

[0105] S403. Based on the prediction results, calculate the parameter sequence of the industrial camera to dynamically compensate for instantaneous changes in illumination;

[0106] In this embodiment, the industrial camera parameter sequence is not a single imaging parameter, but a series of pre-calculated, continuous imaging parameters for illumination changes at different times or locations along the exposure path. These parameters include exposure time, gain, and white balance, designed to ensure that the identification code image is captured with optimal imaging quality throughout the entire exposure window of the image data bag's movement.

[0107] S404. Preload the industrial camera parameter sequence into the industrial camera's local cache;

[0108] In this embodiment, the local cache setting enables the industrial camera to quickly access and switch the required imaging parameters without relying on real-time instructions from the cloud platform.

[0109] S405. The industrial camera automatically switches to the corresponding industrial camera parameters in the industrial camera parameter sequence to capture the image data bag identification code when the image data bag moves through the exposure path according to the preset trigger conditions.

[0110] In this embodiment, when the image data bag moves through the preset exposure path, the industrial camera automatically switches from the local cache to the industrial camera parameters corresponding to the current position or time point in the parameter sequence according to preset triggering conditions, such as the signal of the image data bag entering a specific area, timestamp matching, or position sensor triggering, thereby achieving accurate capture of the image data bag identification code image.

[0111] In the technical solution of the above embodiments, by introducing a prediction mechanism for the motion trajectory of the image data bag and the instantaneous changes in the local light field, and dynamically adjusting the industrial camera parameters accordingly, the capture quality and recognition success rate of the image data bag identification code image are improved in dynamic lighting environments and high-speed motion scenarios. First, the exposure path of the image data bag is predicted by real-time motion data, enabling the system to predict the precise position of the object within the recognition area. Second, combined with a dynamically updated local light field spatial distribution map, the system can predict the instantaneous changes in illumination that the image data bag will experience along the exposure path. Based on these prediction results, the calculated industrial camera parameter sequence can specifically compensate for the instantaneous changes in illumination, ensuring that the camera parameters always match the current lighting conditions throughout the entire motion of the image data bag. Finally, by preloading the parameter sequence into the local cache of the industrial camera and automatically switching it according to preset trigger conditions, real-time communication delays are avoided, ensuring the timeliness of parameter switching and the continuity and accuracy of image capture, ultimately ensuring that clear and stable identification code images can be obtained even in complex and changing environments.

[0112] In one example, suppose on a high-speed logistics sorting line, an image data bag passes through a 0.5-meter-long industrial camera recognition area at a speed of 2 meters per second. Multiple LED lights are installed above this recognition area; these lights may exhibit slight flicker, and the local illumination intensity and spectral composition may fluctuate momentarily due to obstructions from the conveyor belt or changes in ambient light.

[0113] First, encoders and photoelectric sensors mounted on the conveyor belt acquire real-time speed and position data of the image data bag. Based on this real-time motion data, the cloud platform predicts the precise exposure path the image data bag will take within a 0.25-second exposure window (e.g., an exposure time of 50ms, plus processing and transmission time).

[0114] Simultaneously, a pre-deployed array of miniature light sensors within the identified area continuously collects and dynamically updates the local light field spatial distribution map. This map records in detail the light intensity, flicker frequency, and spectral composition of each small grid within the area. The cloud platform matches the predicted exposure path with this dynamic light field distribution map, predicting the instantaneous changes in light intensity, flicker frequency, and spectral composition that the image data bag may experience every 10 milliseconds along the exposure path.

[0115] Based on these predictions, the cloud platform calculates a sequence of 25 industrial camera parameters (0.25 seconds / 10 milliseconds). Each parameter set corresponds to a specific time point or location on the exposure path, used to dynamically compensate for instantaneous changes in illumination at that point. For example, when the illumination intensity suddenly increases at a certain location, the corresponding parameter set will reduce the exposure time or gain; when the spectral composition is bluish at a certain location, the corresponding parameter set will adjust the white balance to compensate.

[0116] Subsequently, this sequence of industrial camera parameters is preloaded into the industrial camera's local cache. When the image data bag enters the recognition area and triggers the photoelectric sensor, the industrial camera automatically and seamlessly switches to the corresponding industrial camera parameters in the parameter sequence to capture the image data bag identification code as the image data bag moves through the exposure path, based on preset trigger conditions (e.g., position information based on encoder feedback or an internal timer). For example, when the image data bag moves to 1 / 5 of the way along the path, the camera automatically switches to the 5th parameter set in the sequence for exposure, ensuring that each frame of the image achieves optimal exposure and color performance throughout the entire movement, thereby significantly improving the identification code recognition success rate.

[0117] In one possible design, step S204 is followed by:

[0118] S204A1, the cloud platform receives instantaneous photo information, identifies the target area of ​​the instantaneous photo information, and determines the main identification area where the image data bag identification code label is located;

[0119] In this embodiment, target region identification refers to accurately locating and determining the core region, i.e., the main identification region, of the image data bag identification code label within the image or data represented by instantaneous illumination information using techniques such as image processing, pattern recognition, or machine learning. For example, methods such as feature matching, deep learning object detection models, or preset region template matching can be used to identify the location of the identification code label. This allows subsequent illumination characteristics to be quantified and focused on the region most critical to the identification code imaging, eliminating interference from illumination information in other non-critical regions.

[0120] S204A2. Quantify the stroboscopic and spectral characteristics of the instantaneous illumination information of the main identification area to obtain the quantization result of the main identification area.

[0121] In this embodiment, instead of performing a general quantification of the entire lighting environment, a refined analysis of the light characteristics of the local area where the identification code label is located is performed. For example, Fourier transform is performed on the lighting data within the main identification area to analyze the flicker frequency and amplitude, and spectral analysis is conducted simultaneously to obtain the spectral composition distribution of the area. This yields more accurate local lighting characteristics directly related to the identification code label.

[0122] S204A3. Correct the quantization results of the main identifier area to obtain the corrected dominant ray flicker characteristics and dominant spectral characteristics;

[0123] In this embodiment, the correction process includes, but is not limited to: compensating based on the pre-stored optical properties of the identification code label material (such as reflectivity, absorptivity, color characteristics, etc.); fine-tuning the illumination quantization results based on auxiliary sensor data such as ambient temperature and humidity; or optimizing the quantization results using historical data and machine learning models to eliminate the influence of measurement errors or environmental noise. Ultimately, this ensures that the final dominant ray flicker characteristics and dominant spectral characteristics can most realistically and accurately reflect the actual performance of the identification code label under the current illumination conditions.

[0124] In the technical solution of the above embodiments, the accuracy of quantizing the flicker characteristics and spectral characteristics of light in the current lighting environment is improved by introducing the steps of target area identification and quantization result correction. First, through target area identification, the focus of illumination characteristic quantization is precisely locked on the main identification area where the image data bag identification code label is located, avoiding the dilution or interference of non-critical area illumination information on the overall quantization result. Second, independent flicker and spectral characteristic quantization is performed on the main identification area, ensuring that the acquired illumination data is highly correlated with the imaging requirements of the identification code label. Finally, by correcting the quantization results, the physical optical characteristics of the identification code label itself and the potential impact of environmental factors on illumination perception are further considered, thereby making the final quantized dominant light flicker characteristics and dominant spectral characteristics closer to the actual imaging conditions of the identification code label.

[0125] In one possible design, step S204 is followed by:

[0126] S204B1: The cloud platform receives instantaneous illumination information, performs high-frequency sampling on the instantaneous illumination information, and obtains a continuous instantaneous illumination data stream.

[0127] In this embodiment, high-frequency sampling refers to continuously acquiring illumination data at a rate much higher than the conventional sampling frequency to ensure that rapid, instantaneous changes in the illumination environment can be captured. Through high-frequency sampling, a continuous stream of instantaneous illumination data can be obtained, containing detailed information on changes over time, such as illumination intensity, flicker frequency, and spectral composition. For example, the sampling frequency can reach hundreds or even thousands of times per second to cope with millisecond-level fluctuations in ambient illumination.

[0128] S204B2. Perform real-time analysis on instantaneous illumination data streams to identify the start and end points of changes in illumination characteristics;

[0129] In this embodiment, real-time analysis is achieved through various signal processing techniques, such as moving average, differential analysis, wavelet transform, or machine learning algorithms, to detect abnormal fluctuations or trend changes in the data stream. When a significant change in light intensity, flicker, or spectral composition exceeding a preset threshold is detected, it is marked as the starting point of the change in lighting characteristics; when the lighting characteristics return to stability or enter a new stable state, it is marked as the ending point. This allows for precise definition of the specific time period during which a significant change in the lighting environment occurs.

[0130] S204B3. Dynamically adjust the length and position of the quantization window based on the start and end points;

[0131] In this embodiment, the quantization window refers to the time period used for quantizing flicker and spectral characteristics. When a change in illumination characteristics is detected, the length of the quantization window can be shortened to focus more on the current stable illumination state, or the position of the quantization window can be moved to the stable region after the change to exclude unstable data during the change process. For example, if a sudden dimming of the illumination is detected, the quantization window is immediately adjusted to the stable stage after the dimming, avoiding the inclusion of data during the dimming process in the calculation, thereby improving the accuracy of the quantization results.

[0132] S204B4. Within the dynamically adjusted quantization window, the flicker characteristics and spectral characteristics of the instantaneous illumination information are quantized to obtain the flicker characteristics and spectral characteristics of the light in the current lighting environment after adjustment.

[0133] In this embodiment, the quantization process dynamically selects the data segment that best represents the current stable lighting environment based on real-time detected changes in illumination. In this way, the flicker characteristics and spectral characteristics of the current lighting environment are obtained after adjustment, and these characteristics can more accurately reflect the actual, dynamically changing lighting conditions.

[0134] In the technical solution of the above embodiments, by introducing high-frequency sampling and real-time analysis of instantaneous illumination information, rapid and transient changes in the lighting environment can be effectively identified, thereby improving the accuracy of quantifying the flicker characteristics and spectral characteristics of light in dynamic or unstable lighting environments. When a significant change in illumination characteristics is detected, the system dynamically adjusts the length and position of the quantization window according to the start and end points of the change. This dynamic adjustment mechanism ensures that the data used for quantifying flicker characteristics and spectral characteristics always comes from relatively stable or representative illumination time periods, thus avoiding the inclusion of unstable data during illumination changes in the calculation. Consequently, in dynamic lighting environments, the quantization results can more accurately reflect the current actual lighting environment characteristics.

[0135] In one example, suppose on an automated transport line, as an image data bag passes through the recognition area of ​​an industrial camera, the overhead lighting fixtures may occasionally experience slight flickering or momentary drops in brightness due to power fluctuations or aging. Traditional quantization methods might quantize the illumination characteristics within a fixed 100-millisecond window. If a momentary drop in brightness occurs within this window, the quantization result will be dragged down by this brief anomaly, leading to an overly long calculated exposure time and thus overexposing the identification code image.

[0136] Using the scheme described in this application, the cloud platform samples instantaneous illumination information at a frequency of 1000 times per second. When a momentary drop in brightness is detected in the lighting fixtures, for example, between 50 and 70 milliseconds, the system immediately identifies this change range. Subsequently, the quantization window is dynamically adjusted. For example, the original 100-millisecond window is adjusted to include only the data before the brightness drop (0-49 milliseconds) or after the brightness stabilizes (71-100 milliseconds), or the window is shortened to include only the most stable illumination data segment. In this way, the quantized light flicker characteristics and spectral characteristics more accurately reflect the actual situation of the image data bag identification code under stable illumination conditions. This allows the industrial camera to calculate more accurate exposure time, gain, and white balance parameters, ultimately capturing clear and accurate identification code images and avoiding image quality degradation caused by instantaneous changes in illumination.

[0137] In one possible design, Figure 5 This is a flowchart illustrating step S2 according to another exemplary embodiment. (Refer to the attached document.) Figure 5 Step S2 includes:

[0138] S211. The cloud platform receives instantaneous illumination information and obtains the lighting equipment identifier contained in the instantaneous illumination information;

[0139] In this embodiment, after receiving instantaneous illumination information, the cloud platform can parse the lighting device identifier from the information. This identifier can be a unique serial number pre-configured in the lighting device, a MAC address, or device model information obtained through image recognition or other methods. The instantaneous illumination information is then associated with a specific lighting device through this identifier.

[0140] S212. Based on the lighting equipment identification, query the operating time, ambient temperature, and ambient humidity data of the lighting equipment, and based on the operating time, ambient temperature, and ambient humidity data, evaluate the luminous efficiency and spectral output drift of the lighting equipment.

[0141] In this embodiment, based on the acquired lighting device identifier, the cloud platform accesses a preset device database or IoT platform to query the operating time, ambient temperature, and ambient humidity data of the lighting device corresponding to that identifier. This data is uploaded in real time by the lighting device's own sensors or collected through environmental sensors. Based on this historical operating data and environmental parameters, a model is established or a preset performance degradation curve is consulted to evaluate the current luminous efficiency and spectral output drift of the lighting device. For example, as operating time increases, the luminous efficiency of LED beads may decrease, and the spectral composition may also shift slightly; excessively high or low ambient temperatures may also affect its performance.

[0142] S213. Quantify the instantaneous illumination information by its stroboscopic and spectral characteristics;

[0143] In this embodiment, the received instantaneous illumination data undergoes signal processing, such as Fourier transform analysis of its frequency components to identify flicker characteristics, or analysis of its wavelength distribution using spectrometer data to determine spectral characteristics. The aim is to initially obtain the physical characteristics of the current illumination environment.

[0144] S214. Based on the luminous efficiency and spectral output drift, the quantization results are compensated to obtain the light flicker characteristics and spectral characteristics of the current lighting environment;

[0145] In this embodiment, for example, if the evaluation results show that the luminous efficiency has decreased by 5%, a corresponding correction is made to the quantized illuminance value; if the spectral output drifts to a certain band, the weight or distribution of the quantized spectral characteristics is adjusted. This compensation mechanism aims to correct the deviation between the instantaneous illumination information and the actual ideal illumination characteristics caused by changes in the performance of the lighting equipment itself, thereby obtaining light flicker characteristics and spectral characteristics that are closer to reality.

[0146] In the technical solutions of the above embodiments, by introducing lighting equipment identification and its operating time, ambient temperature, and ambient humidity data, the luminous efficiency and spectral output drift of the lighting equipment can be evaluated. Because lighting equipment experiences performance degradation and spectral drift during long-term operation or under specific environmental conditions, the characteristics of the emitted light deviate from the initial design values. Relying solely on instantaneous illumination information is insufficient to capture these long-term or environment-related changes. By evaluating these drifts and applying them to compensate for the quantification results of instantaneous illumination information, this application can effectively correct errors introduced by changes in the performance of the lighting equipment itself. This allows the quantified light flicker characteristics and spectral characteristics to more accurately reflect the true state of the current lighting environment, providing a foundation for the precise calculation of subsequent industrial camera imaging parameters.

[0147] In one possible design, step S2 further includes:

[0148] S221. The cloud platform receives instantaneous illumination information, performs frequency domain analysis on the instantaneous illumination information, and identifies the light flicker components and spectral components of multiple frequency bands as sample components.

[0149] In this embodiment, after receiving the instantaneous illumination information acquired by the industrial camera, the cloud platform first performs frequency domain analysis on the information. Frequency domain analysis refers to converting the time-domain signal (i.e., the instantaneous illumination information) into a frequency-domain signal to reveal the various frequency components it contains. Using mathematical tools such as Fourier transform, the complex instantaneous illumination signal is decomposed into light flicker components and spectral components of different frequencies. These components are identified and used as sample components. This deconstructs the complex signal in the original illumination information, facilitating subsequent differentiation and processing of signals from different sources.

[0150] S222. Import the sample components of multiple frequency bands into the preset non-lighting equipment interference feature library, and identify and obtain the non-associated flicker components and non-associated spectral components related to non-lighting equipment.

[0151] In this embodiment, the non-lighting equipment interference feature library stores typical flicker patterns, spectral characteristics, or noise fingerprints associated with non-lighting equipment. For example, this feature library contains characteristic data of electromagnetic noise frequencies generated by specific types of motors, specific spectral shifts caused by conveyor belt reflections, or other known environmental interference sources. After importing sample components from multiple frequency bands into this feature library, non-associated flicker components and non-associated spectral components related to non-lighting equipment are identified through methods such as pattern matching and signal correlation analysis. This accurately distinguishes interference signals caused by non-lighting equipment, preventing them from affecting the quantification of the actual lighting environment characteristics.

[0152] S223. From the sample components of multiple frequency bands, remove the non-correlated stroboscopic components and non-correlated spectral components to obtain the pure sample light components of multiple frequency bands;

[0153] In this embodiment, digital signal processing techniques, such as filter design, signal reconstruction, or subtraction operations, are used to remove the identified non-correlated interference components from the original sample components. This yields clean sample light components across multiple frequency bands, which contain only illumination information generated by the lighting equipment that truly reflects the characteristics of the lighting environment.

[0154] S224. Quantify the pure sample light components of multiple frequency bands to obtain the light flicker characteristics and spectral characteristics of the current lighting environment;

[0155] In this embodiment, based on these interference-free pure light components, the precise light flicker characteristics (e.g., flicker frequency, flicker depth) and spectral characteristics (e.g., color temperature, spectral distribution) of the current lighting environment are calculated. This ultimately provides an accurate and reliable description of the lighting environment characteristics, enabling precise adjustment of industrial camera imaging parameters.

[0156] In the technical solution of the above embodiments, by introducing a mechanism for identifying and removing interference from non-lighting devices, interference from non-lighting devices in instantaneous illumination information is filtered out, significantly improving the accuracy of quantification of the flicker characteristics and spectral characteristics of the current lighting environment. First, frequency domain analysis is performed on the instantaneous illumination information, decomposing the complex time-domain signal into easily identifiable and processable frequency band components. Second, by comparing these sample components with a preset non-lighting device interference feature library, unrelated flicker components and unrelated spectral components generated by non-lighting devices are accurately identified. Then, these components are removed from the original sample components to obtain pure sample light components. Finally, quantization is performed based on these pure sample light components, ensuring that the obtained light flicker characteristics and spectral characteristics truly reflect the lighting environment and avoiding quantization distortion caused by interference.

[0157] In summary, the image data bag based on a cloud platform provided in this embodiment of the invention acquires instantaneous illumination information of the current lighting environment and transmits this information to the cloud platform. Based on the received instantaneous illumination information, the cloud platform can accurately quantify the light flicker characteristics and spectral characteristics of the current lighting environment. Furthermore, the cloud platform calculates the imaging parameters of the industrial camera and adjusts the camera settings of the industrial camera according to the calculated imaging parameters, ultimately capturing the identification code image of the image data bag with the adjusted industrial camera. The method provided in this application can dynamically sense and adapt to complex changes in the lighting environment, and by adjusting the exposure time, gain, and white balance of the industrial camera in real time, it ensures the capture of high-quality, artifact-free, and high-contrast identification code images under different lighting conditions, significantly improving the overall recognition accuracy of the image data bag. Simultaneously, the improved recognition accuracy shortens the extraction cycle of a single image data bag, thereby improving the overall rapid extraction efficiency.

[0158] Example 2

[0159] Embodiment 2 of the present invention provides an intelligent recognition system for image data bags based on a cloud platform. Figure 6 This is a block diagram illustrating a cloud-based image portfolio intelligent recognition system according to an exemplary embodiment. Figure 6 As shown, the system includes:

[0160] The illumination information acquisition module 01 is used to acquire instantaneous illumination information of the current lighting environment through an industrial camera before capturing the image of the image data bag identification code.

[0161] The illumination characteristic quantization module 02 is used to acquire instantaneous illumination information and, based on the instantaneous illumination information, quantify the light flicker characteristics and spectral characteristics of the current lighting environment;

[0162] The imaging parameter calculation module 03 is used to calculate the imaging parameters of the industrial camera based on the quantized light flicker characteristics and quantized spectral characteristics. The imaging parameters include exposure time, gain and white balance.

[0163] The imaging parameter adjustment module 04 is used to adjust the camera settings of the industrial camera according to the imaging parameter instructions, and to capture the identification code image of the image data bag with the adjusted industrial camera.

[0164] In summary, the cloud-based intelligent image data bag recognition system provided in this invention acquires instantaneous illumination information of the current lighting environment and transmits this information to the cloud platform. Based on the received instantaneous illumination information, the cloud platform can accurately quantify the light flicker characteristics and spectral characteristics of the current lighting environment. Furthermore, the cloud platform calculates the imaging parameters of the industrial camera and adjusts the camera settings according to the calculated imaging parameters, ultimately capturing the identification code image of the image data bag with the adjusted industrial camera. The method provided in this application can dynamically sense and adapt to complex changes in the lighting environment, and by adjusting the exposure time, gain, and white balance of the industrial camera in real time, it ensures the capture of high-quality, artifact-free, and high-contrast identification code images under different lighting conditions, significantly improving the overall recognition accuracy of the image data bag. Simultaneously, the improved recognition accuracy shortens the extraction cycle of a single image data bag, thereby improving the overall rapid extraction efficiency, ultimately forming a highly adaptable, highly accurate, and efficient intelligent recognition solution.

[0165] Example 3

[0166] Embodiment 3 of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the method provided in Embodiment 1.

[0167] The readable storage medium may be more specifically adopted, including but not limited to: portable disk, hard disk, random access memory, read-only memory, erasable programmable read-only memory, optical storage device, magnetic storage device, or any suitable combination thereof.

[0168] In a possible implementation, the present invention can also be implemented as a program product comprising program code that, when the program product is run on a terminal device, causes the terminal device to perform steps implementing the method provided in Embodiment 1.

[0169] The program code for executing the present invention can be written in any combination of one or more programming languages. The program code can be executed entirely on the user device, partially on the user device, as a standalone software package, partially on the user device and partially on a remote device, or entirely on a remote device.

[0170] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0171] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A cloud-based intelligent recognition method for image data bags, characterized in that, Includes the following steps: Before capturing the image of the image data bag identification code, obtain the instantaneous illumination information of the current lighting environment; The cloud platform acquires the instantaneous illumination information and, based on the instantaneous illumination information, quantifies the light flicker characteristics and spectral characteristics of the current lighting environment; The cloud platform calculates the imaging parameters of the industrial camera based on quantized light flicker characteristics and quantized spectral characteristics, wherein the imaging parameters include exposure time, gain, and white balance. The camera settings of the industrial camera are adjusted according to the imaging parameter instructions, and the adjusted industrial camera is used to capture the identification code image of the image data bag. The step of adjusting the camera settings of the industrial camera according to the imaging parameter instructions, and capturing the identification code image of the image data bag with the adjusted industrial camera, includes: Based on real-time motion data of the image data bag, the exposure path of the image data bag within the exposure window period of the industrial camera's recognition area is predicted; Based on the dynamically updated local light field spatial distribution map, the instantaneous changes in light intensity, flicker frequency, and spectral composition that the image data bag may experience along the exposure path are predicted. Based on the predicted results, calculate the industrial camera parameter sequence to dynamically compensate for the instantaneous changes in illumination; The industrial camera parameter sequence is preloaded into the local cache of the industrial camera; According to preset triggering conditions, when the image data bag moves through the exposure path, the industrial camera automatically switches to the corresponding industrial camera parameters in the industrial camera parameter sequence to capture the image data bag identification code image.

2. The method according to claim 1, characterized in that, The cloud platform acquires the instantaneous illumination information and, based on the instantaneous illumination information, quantifies the light flicker characteristics and spectral characteristics of the current lighting environment, including: The cloud platform receives the instantaneous illumination information, divides the instantaneous illumination information into regions, and obtains the partitioned instantaneous illumination information of multiple illumination regions; The instantaneous illumination information of each partition is quantized for flicker characteristics and spectral characteristics to obtain quantization results for multiple illumination regions. Based on the degree of influence of each local illumination region on the image data bag identification code imaging, the quantization results of multiple local illumination regions are weighted and fused to obtain the dominant light flicker characteristics and dominant spectral characteristics. The dominant light flicker characteristics and the dominant spectral characteristics are used as the light flicker characteristics and spectral characteristics of the current lighting environment.

3. The method according to claim 1, characterized in that, The cloud platform calculates the imaging parameters of the industrial camera based on quantized light flicker characteristics and quantized spectral characteristics, including: Identify the tag characteristics of the image data bag identification code; Based on the characteristics of the tag, select an imaging parameter adjustment strategy; By combining the quantized light flicker characteristics, the quantized spectral characteristics, and the imaging parameter adjustment strategy, the exposure time, gain, and white balance parameters of the industrial camera are calculated.

4. The method according to claim 2, characterized in that, The step of using the dominant ray flicker characteristics and the dominant spectral characteristics as the quantification of the ray flicker characteristics and spectral characteristics of the current lighting environment, followed by: The cloud platform receives the instantaneous light information, identifies the target area of ​​the instantaneous light information, and determines the main identification area where the image data bag identification code label is located; The instantaneous illumination information of the main identification area is quantized using stroboscopic and spectral characteristics to obtain the quantization result of the main identification area. The quantization results of the main identifier region are corrected to obtain the corrected dominant spectral flicker characteristics and dominant spectral characteristics.

5. The method according to claim 2, characterized in that, The method of using the dominant ray flicker characteristics and the dominant spectral characteristics as the ray flicker characteristics and spectral characteristics of the current lighting environment further includes: The cloud platform receives the instantaneous illumination information, performs high-frequency sampling on the instantaneous illumination information, and obtains a continuous instantaneous illumination data stream; The instantaneous illumination data stream is analyzed in real time to identify the start and end points of changes in illumination characteristics; The length and position of the quantization window are dynamically adjusted based on the start and end points. Within the dynamically adjusted quantization window, the instantaneous illumination information is quantized for flicker and spectral characteristics to obtain the adjusted flicker and spectral characteristics of the current lighting environment.

6. The method according to claim 1, characterized in that, The cloud platform acquires the instantaneous illumination information and, based on the instantaneous illumination information, quantifies the light flicker characteristics and spectral characteristics of the current lighting environment, including: The cloud platform receives the instantaneous illumination information and obtains the lighting device identifier contained in the instantaneous illumination information; Based on the lighting equipment identifier, query the operating time, ambient temperature, and ambient humidity data of the lighting equipment, and evaluate the luminous efficiency and spectral output drift of the lighting equipment based on the operating time, ambient temperature, and ambient humidity data; The instantaneous illumination information is quantified based on its stroboscopic and spectral characteristics. The quantization results are compensated based on the luminous efficiency and the spectral output drift to obtain the light flicker characteristics and spectral characteristics of the current lighting environment.

7. The method according to claim 1, characterized in that, The cloud platform acquires the instantaneous illumination information and, based on the instantaneous illumination information, quantifies the light flicker characteristics and spectral characteristics of the current lighting environment, including: The cloud platform receives the instantaneous illumination information, performs frequency domain analysis on the instantaneous illumination information, and identifies the light flicker components and spectral components of multiple frequency bands as sample components; Sample components from multiple frequency bands are imported into a pre-defined non-lighting equipment interference feature library to identify and obtain non-associated flicker components and non-associated spectral components related to non-lighting equipment. From the sample components of multiple frequency bands, the unrelated stroboscopic components and the unrelated spectral components are stripped to obtain the pure sample light components of multiple frequency bands; The light components of pure samples in multiple frequency bands are quantified to obtain the flicker characteristics and spectral characteristics of the current lighting environment.

8. An intelligent recognition system for image document bags, characterized in that, The system includes: The illumination information acquisition module is used to acquire instantaneous illumination information of the current lighting environment through an industrial camera before capturing the image of the image data bag identification code. The illumination characteristic quantization module is used to acquire the instantaneous illumination information and, based on the instantaneous illumination information, quantify the light flicker characteristics and spectral characteristics of the current lighting environment; An imaging parameter calculation module is used to calculate the imaging parameters of the industrial camera based on quantized light flicker characteristics and quantized spectral characteristics, wherein the imaging parameters include exposure time, gain, and white balance. An imaging parameter adjustment module is used to adjust the camera settings of the industrial camera according to the imaging parameter instructions, and to capture the identification code image of the image data bag with the adjusted industrial camera. The step of adjusting the camera settings of the industrial camera according to the imaging parameter instructions, and capturing the identification code image of the image data bag with the adjusted industrial camera, includes: Based on real-time motion data of the image data bag, the exposure path of the image data bag within the exposure window period of the industrial camera's recognition area is predicted; Based on the dynamically updated local light field spatial distribution map, the instantaneous changes in light intensity, flicker frequency, and spectral composition that the image data bag may experience along the exposure path are predicted. Based on the predicted results, calculate the industrial camera parameter sequence to dynamically compensate for the instantaneous changes in illumination; The industrial camera parameter sequence is preloaded into the local cache of the industrial camera; According to preset triggering conditions, when the image data bag moves through the exposure path, the industrial camera automatically switches to the corresponding industrial camera parameters in the industrial camera parameter sequence to capture the image data bag identification code image.

9. A computer-readable storage medium, characterized in that, It stores a computer program thereon, which, when executed by a processor, implements the steps of the method according to any one of claims 1-7.

Images