Intelligent processing system and method for batch uploading of pictures

Abstract

The invention discloses an intelligent processing system and method for batch uploading of pictures. The intelligent processing system comprises a client module and a server module, the client module is used for providing a graphical user interface, receiving user input and executing local processing of pictures; and the server module is used for receiving and storing the pictures and related data uploaded from the client. According to the method, the defects in the prior art are effectively overcome by introducing an automatic processing flow of intelligent identification and strategy driving. The system performs local intelligent analysis before picture uploading, and can automatically filter low-quality, repeated or non-compliant pictures, so that uploading and storage of invalid data are greatly reduced, and the post-operation cost of a user and the redundant load of a server are reduced. On the basis of the recognized picture content, the system can automatically execute differentiated processing strategies, such as picture optimization, classification, specific label or watermark addition and the like, so that personalized and intelligent processing is realized, and the picture management efficiency and precision are improved.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, specifically to an intelligent system and method for batch image uploading. Background Technology

[0002] In current internet applications, batch image upload functionality has become a standard feature in systems such as social networks, e-commerce, and content management platforms. However, most existing batch upload technologies have simple and rigid processes, typically only providing basic functions such as file selection, initial client-side validation (e.g., format, size), and sequential upload. When users upload a large number of images, the system cannot intelligently understand the image content, resulting in inefficient upload processes. For example, the system cannot automatically identify and filter out blurry, duplicate, or sensitive images, requiring users to manually filter and process them after upload, increasing the user's workload and time cost. Furthermore, due to the lack of pre-analysis of image content, the system cannot perform differentiated processing based on the image's inherent attributes (e.g., type, theme) (such as automatically adding watermarks to product images or intelligently beautifying portrait images), nor can it optimize upload strategies, leading to insufficient network bandwidth utilization. This is especially true when processing large numbers of high-resolution images, resulting in lengthy upload times, concentrated server processing pressure, and a poor overall user experience. Summary of the Invention

[0003] Therefore, the present invention provides an intelligent system and method for batch image uploading to solve the problems in the prior art.

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] A smart image batch upload system, comprising a client module and a server module;

[0006] The client module provides a graphical user interface, receives user input, and performs local image processing.

[0007] The server module is used to receive and store images and related data uploaded from the client;

[0008] The client module includes:

[0009] The image acquisition and caching unit is used to receive and temporarily store the original image set selected by the user.

[0010] The intelligent recognition engine integrates a lightweight neural network model for image quality assessment, content classification, and duplicate detection analysis.

[0011] The strategy decision-maker is used to assign appropriate processing strategies to each image based on the output of the intelligent recognition engine.

[0012] An image processor is used to perform image filtering, tagging, and image enhancement operations according to the instructions of a policy decision-maker.

[0013] The upload management unit is responsible for managing the upload queue, scheduling upload tasks, and implementing network-adaptive upload control.

[0014] A report generator that produces a visual upload processing report after the upload task is completed.

[0015] Preferably, the server module further includes a user behavior analysis unit, used to collect and analyze user feedback on the uploaded processing report, especially the final disposal choice for "pending confirmation" images; the user behavior analysis unit feeds back the analysis results to the strategy decision-maker of the client module, used to optimize and adjust the matching rules of subsequent image processing strategies, so as to realize the self-learning and continuous optimization of the processing strategy.

[0016] Preferably, the lightweight neural network model in the intelligent recognition engine adopts an image classification model based on the MobileNetV3 architecture, and has been optimized for mobile devices through model quantization and pruning. The model is pre-trained on the server side using massive image data, and is incrementally updated and fine-tuned periodically through feedback data collected by the user behavior analysis unit. It is then synchronized to each client through a model distribution mechanism to continuously improve recognition accuracy and adapt to new image types.

[0017] Preferably, the upload management unit includes a network status awareness subunit and a queue scheduling subunit. The network status awareness subunit continuously detects the current network round-trip latency and packet loss rate, and constructs a comprehensive network health score based on the available bandwidth estimate. The queue scheduling subunit dynamically adjusts the number of concurrent upload threads based on this network health score, increasing the number of concurrent threads to accelerate uploads when the score is high and decreasing the number of concurrent threads to maintain connection stability when the score is low. At the same time, this subunit also sorts the upload queue according to the file size and priority label of the images using a weighted fair queue algorithm to ensure that small files and high-priority images are transmitted in a timely manner.

[0018] To achieve the above objectives, the present invention also provides the following technical solution:

[0019] A smart method for batch image uploading includes the following steps:

[0020] S1, User Interaction and Image Collection Acquisition: Receives batch image selection operations from users through a graphical user interface and obtains the original image collection to be uploaded;

[0021] S2, Local Preprocessing and Intelligent Recognition Analysis: Local preprocessing and intelligent recognition analysis are performed on each image in the original image set; the local preprocessing includes at least parsing the image's metadata information; the intelligent recognition analysis is performed by a lightweight neural network model loaded onto the client, and the recognition content includes at least image quality assessment, main content classification, and duplicate image detection;

[0022] S3, Automatic matching and allocation of processing strategies: Based on the intelligent recognition and analysis results of step S2, an appropriate processing strategy is automatically matched and allocated for each image; the processing strategy includes at least one or more of the following: marking blurry or low-quality images as "to be confirmed", identifying detected duplicate image groups and recommending the retention of the optimal version, automatically generating keyword tags based on the main content classification results, and specifying subsequent image enhancement operation parameters for specific types of images according to preset rules;

[0023] S4, Automated processing execution driven by strategy: According to the processing strategy allocated in step S3, perform corresponding automated processing operations on the images in the original image set to generate a filtered and processed queue to be uploaded; the automated processing operations include: moving images marked "pending confirmation" to a separate list to await the user's final decision, automatically selecting the highest quality image from the duplicate image group and keeping it in the queue while removing other images, and attaching the keyword tags generated in step S3 to the images;

[0024] S5, Intelligent Upload Scheduling with Network Adaptation: Initiate batch upload tasks to upload images from the queue generated in step S4 to the server; during the upload process, monitor network bandwidth and server response status in real time, and dynamically adjust the number of concurrent upload threads and the priority of upload order based on the monitoring results and the attributes of the images themselves.

[0025] S6, Upload Result Processing Report Generation and Feedback: After the upload task is completed, the system generates and presents an upload processing report; the report lists in detail the successfully uploaded images, the filtered images and the reasons for filtering, as well as a summary of the automated processing operations performed, and provides the user with the final operation interface for the "pending confirmation" image list.

[0026] Preferably, in step S2, the image quality assessment specifically includes: calculating the image's sharpness index, brightness distribution histogram, and color contrast, and comparing these indices with a preset quality threshold to output a comprehensive quality score; the main content classification identifies the image as one or more categories such as landscape, portrait, document, food, and goods; the duplicate image detection is achieved by calculating the perceptual hash value of the image and comparing the Hamming distance between the hash values.

[0027] Preferably, in step S3, the "automatic generation of keyword tags based on the main content classification results" specifically involves: establishing a mapping relationship between content classification and keyword database; when an image is classified as "mechanical equipment", an "industrial" tag is automatically generated; when it is classified as "fruits and vegetables", an "agricultural and sideline products" tag is automatically generated; and simultaneously, geographic location tags are generated by combining the geographic location information in the image metadata.

[0028] Preferably, in step S3, the "specifying subsequent image enhancement operation parameters for specific types of images according to preset rules" includes: specifying higher sharpening parameters and black-and-white binarization parameters for images classified as "documents" to improve text readability; specifying skin smoothing and beautification parameters for images classified as "portraits"; and specifying saturation enhancement and contrast optimization parameters for images classified as "landscapes".

[0029] Preferably, in step S5, the "dynamically adjusting the number of concurrent upload threads and the priority of upload order" specifically includes:

[0030] S51: Calculate the ideal throughput of a single upload thread based on the real-time monitored network available bandwidth, and dynamically increase or decrease the number of concurrent upload threads accordingly to maximize the use of available bandwidth without causing network congestion.

[0031] S52: Set upload priority based on image file size and processing strategy complexity. Images with small file sizes that do not require complex image enhancement processing on the client side are uploaded first so that users can see the first batch of upload results faster; while images with large file sizes or that require time-consuming processing are uploaded later.

[0032] This invention offers the following advantages: By introducing an automated processing flow driven by intelligent recognition and strategies, it effectively overcomes the shortcomings of existing technologies. First, the system performs local intelligent analysis before image upload, automatically filtering low-quality, duplicate, or non-compliant images, significantly reducing the upload and storage of invalid data and lowering post-processing costs for users and redundant server load. Second, based on the recognized image content, the system automatically executes differentiated processing strategies, such as image optimization, classification, adding specific tags or watermarks, achieving personalized and intelligent processing and improving the efficiency and accuracy of image management. Finally, by dynamically adjusting the upload queue and concurrency based on image attributes, size, and network conditions, this invention optimizes network bandwidth utilization, shortens the overall upload time, and provides users with clear processing reports through intelligent feedback on upload results, significantly improving the automation level, processing efficiency, and user experience of batch image uploads. Attached Figure Description

[0033] To more intuitively illustrate the prior art and this application, exemplary drawings are provided below. It should be understood that the specific shapes and structures shown in the drawings should not generally be regarded as limiting conditions for implementing this application; for example, based on the technical concept disclosed in this application and the exemplary drawings, those skilled in the art are able to easily make conventional adjustments or further optimizations to the addition / reduction / classification, specific shapes, positional relationships, connection methods, size ratios, etc. of certain units (components).

[0034] Figure 1 A module diagram of an intelligent image batch upload processing system provided in this application embodiment;

[0035] Figure 2 This is a flowchart of an intelligent method for batch uploading images, provided in an embodiment of this application. Detailed Implementation

[0036] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. It should be understood that these embodiments are merely for further explanation of the present invention and should not be construed as limiting the scope of protection of the present invention. Technical engineers in the field can make some non-essential improvements and adjustments to the present invention based on the above-described content. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0037] Please see Figure 1 A smart image batch upload processing system, comprising a client module and a server module;

[0038] The client module provides a graphical user interface, receives user input, and performs local image processing.

[0039] The server module is used to receive and store images and related data uploaded from the client;

[0040] The client module includes:

[0041] The image acquisition and caching unit is used to receive and temporarily store the original image set selected by the user.

[0042] The intelligent recognition engine integrates a lightweight neural network model for image quality assessment, content classification, and duplicate detection analysis.

[0043] The strategy decision-maker is used to assign appropriate processing strategies to each image based on the output of the intelligent recognition engine.

[0044] An image processor is used to perform image filtering, tagging, and image enhancement operations according to the instructions of a policy decision-maker.

[0045] The upload management unit is responsible for managing the upload queue, scheduling upload tasks, and implementing network-adaptive upload control.

[0046] A report generator that produces a visual upload processing report after the upload task is completed.

[0047] In this system implementation, the client module acts as a fully functional miniature processing center, executing pipeline operations sequentially through its various internal units. The image acquisition and caching unit first takes over the original image selected by the user. Subsequently, the intelligent recognition engine uses an integrated lightweight model to perform deep analysis of the image. The analysis results are submitted to the policy decision-maker, which assigns action instructions to each image based on a built-in rule base. The image processor performs specific tasks such as filtering, labeling, and enhancement. The processed image queue is then taken over by the upload management unit, which transmits the data stream to the server in the most optimal way. The server module is responsible for receiving and storing images and can expand user behavior analysis functions to optimize the system. The entire system operates on a highly decoupled and collaborative principle: the client completes most of the intelligent work, effectively reducing server load; the server provides support and storage. This design not only improves processing efficiency (because images can be processed without round-trip transmission) but also strengthens the privacy protection of the user's original data.

[0048] The server module also includes a user behavior analysis unit, which collects and analyzes user feedback on the uploaded processing report, especially the final disposal choice for "pending confirmation" images. The user behavior analysis unit feeds the analysis results back to the strategy decision-maker of the client module, which is used to optimize and adjust the matching rules of subsequent image processing strategies, so as to realize the self-learning and continuous optimization of processing strategies.

[0049] The user behavior analysis unit analyzes users' final actions regarding system recommendations (such as keeping or deleting "pending confirmation" images), using this feedback as training data to iteratively optimize the rules of the client-side policy decision-maker. This allows the system to increasingly align with users' personal preferences, achieving personalized intelligence.

[0050] The lightweight neural network model in the intelligent recognition engine adopts an image classification model based on the MobileNetV3 architecture, and has been optimized for mobile devices through model quantization and pruning. The model is pre-trained on the server side using massive image data, and is incrementally updated and fine-tuned periodically through feedback data collected by the user behavior analysis unit. It is then synchronized to each client through the model distribution mechanism to continuously improve the recognition accuracy and adapt to new image types.

[0051] It utilizes MobileNetV3, a lightweight model architecture optimized for mobile devices, and further compresses the model through techniques such as quantization and pruning to ensure efficient operation on mobile phones and other devices. Furthermore, it elucidates a mechanism that combines server-side pre-training with incremental client-side updates, forming a closed-loop learning system that continuously evolves and becomes smarter with use.

[0052] The upload management unit includes a network status awareness subunit and a queue scheduling subunit. The network status awareness subunit continuously detects the current network round-trip latency and packet loss rate, and constructs a comprehensive network health score based on the estimated available bandwidth. The queue scheduling subunit dynamically adjusts the number of concurrent upload threads based on this network health score, increasing the number of concurrent threads to accelerate uploads when the score is high and decreasing the number of concurrent threads to maintain connection stability when the score is low. At the same time, this subunit also sorts the upload queue according to the file size and priority label of the images using a weighted fair queue algorithm to ensure that small files and high-priority images are transmitted in a timely manner.

[0053] The network status awareness subunit assesses "network health" by comprehensively considering bandwidth, latency, and packet loss rate, rather than relying on a single metric. The queue scheduling subunit then uses this health assessment and image attributes, employing advanced scheduling algorithms such as weighted fair queues, to intelligently determine the upload order and concurrency control, thereby ensuring that upload tasks are both fast and stable in complex and ever-changing network environments.

[0054] like Figure 2 As shown, a smart method for batch uploading images includes the following steps:

[0055] S1, User Interaction and Image Collection Acquisition: Receives batch image selection operations from users through a graphical user interface and obtains the original image collection to be uploaded;

[0056] S2, Local Preprocessing and Intelligent Recognition Analysis: Local preprocessing and intelligent recognition analysis are performed on each image in the original image set; the local preprocessing includes at least parsing the image's metadata information; the intelligent recognition analysis is performed by a lightweight neural network model loaded onto the client, and the recognition content includes at least image quality assessment, main content classification, and duplicate image detection;

[0057] S3, Automatic matching and allocation of processing strategies: Based on the intelligent recognition and analysis results of step S2, an appropriate processing strategy is automatically matched and allocated for each image; the processing strategy includes at least one or more of the following: marking blurry or low-quality images as "to be confirmed", identifying detected duplicate image groups and recommending the retention of the optimal version, automatically generating keyword tags based on the main content classification results, and specifying subsequent image enhancement operation parameters for specific types of images according to preset rules;

[0058] S4, Automated processing execution driven by strategy: According to the processing strategy allocated in step S3, perform corresponding automated processing operations on the images in the original image set to generate a filtered and processed queue to be uploaded; the automated processing operations include: moving images marked "pending confirmation" to a separate list to await the user's final decision, automatically selecting the highest quality image from the duplicate image group and keeping it in the queue while removing other images, and attaching the keyword tags generated in step S3 to the images;

[0059] S5, Intelligent Upload Scheduling with Network Adaptation: Initiate batch upload tasks to upload images from the queue generated in step S4 to the server; during the upload process, monitor network bandwidth and server response status in real time, and dynamically adjust the number of concurrent upload threads and the priority of upload order based on the monitoring results and the attributes of the images themselves.

[0060] S6, Upload Result Processing Report Generation and Feedback: After the upload task is completed, the system generates and presents an upload processing report; the report lists in detail the successfully uploaded images, the filtered images and the reasons for filtering, as well as a summary of the automated processing operations performed, and provides the user with the final operation interface for the "pending confirmation" image list.

[0061] In this method, the user selects images in batches via a graphical interface (S1). The system then initiates the core intelligent recognition and analysis on the user's local device (S2). This step, through a lightweight AI model integrated into the client, pre-examines the images without transmitting any data, assessing their quality, classifying content, and identifying duplicates, providing data support for subsequent decisions. Based on the recognition results, the system enters the strategy decision-making stage (S3), "tailor-making" a processing plan for each image, such as automatically marking low-quality images, recommending the retention of the best duplicates, or assigning image optimization parameters, achieving personalized and automated processing. Next, the system strictly executes the predetermined strategy (S4), completing image filtering, classification, and preprocessing, forming a refined queue to be uploaded. In the upload stage (S5), the system does not simply transmit data randomly, but introduces a network adaptive mechanism, dynamically adjusting the concurrency and priority based on real-time network conditions and image attributes, much like an intelligent traffic control system, ensuring an efficient and stable upload process. Finally, the system generates a detailed report (S6) that provides a closed-loop summary of the entire process. This report not only informs the user of the results but also provides the user with an entry point to make final interventions in the system's decisions, ensuring the smoothness of human-machine collaboration.

[0062] In step S2, the image quality assessment specifically includes: calculating the image's sharpness index, brightness distribution histogram, and color contrast, and comparing these indices with preset quality thresholds to output a comprehensive quality score; the main content classification identifies the image into one or more categories such as landscape, portrait, document, food, and goods; the duplicate image detection is achieved by calculating the image's perceptual hash value and comparing the Hamming distance between hash values. Quality assessment requires a comprehensive consideration of multiple dimensions such as sharpness, brightness, and contrast; content classification must be specific to a particular scene; and duplicate detection uses the efficient and computationally efficient perceptual hash and Hamming distance techniques. For example, the system calculates the image's blurriness; if it exceeds a threshold, it is considered poor quality; by comparing hash values, images with a Hamming distance less than 5 can be identified as duplicates.

[0063] In step S3, the "automatic generation of keyword tags based on the main content classification results" specifically involves: establishing a mapping relationship between content classification and a keyword database. When an image is classified as "mechanical equipment," the "industrial" tag is automatically generated; when classified as "fruits and vegetables," the "agricultural and sideline products" tag is automatically generated. Simultaneously, geographic location tags are generated by combining the geographic location information in the image metadata. Automated tagging is achieved by establishing a classification-tag mapping table. Furthermore, it integrates geographic location information from the metadata into the tagging system, allowing a portrait photo taken in Paris to simultaneously obtain both "portrait" and "Paris, France" tags, greatly enriching the semantic information of the image.

[0064] In step S3, "specifying subsequent image enhancement operation parameters for specific types of images according to preset rules" includes: specifying higher sharpening parameters and black-and-white binarization parameters for images classified as "documents" to improve text readability; specifying skin smoothing and beautification parameters for images classified as "portraits"; and specifying saturation enhancement and contrast optimization parameters for images classified as "landscapes".

[0065] Image enhancement strategies embody refined processing, pre-setting differentiated enhancement parameters for the characteristics of different image types: sharpening and binarizing documents to improve readability, beautifying portraits to optimize visual appeal, and enhancing colors in landscapes to make them more vivid. This ensures that automated processing is not a "one-size-fits-all" approach, but rather intelligent optimization with scene awareness.

[0066] In step S5, the "dynamically adjusting the number of concurrent upload threads and the priority of upload order" specifically includes:

[0067] S51: Calculate the ideal throughput of a single upload thread based on the real-time monitored network available bandwidth, and dynamically increase or decrease the number of concurrent upload threads accordingly to maximize the use of available bandwidth without causing network congestion.

[0068] S52: Set upload priority based on image file size and processing strategy complexity. Images with small file sizes that do not require complex image enhancement processing on the client side are uploaded first so that users can see the first batch of upload results faster; while images with large file sizes or that require time-consuming processing are uploaded later.

[0069] The dynamic upload strategy analogizes the upload process to intelligent traffic management: on the one hand, it adjusts the number of "lanes" based on real-time bandwidth (such as increasing threads under high-speed Wi-Fi and reducing threads under weak networks); on the other hand, it sets the priority of "vehicles" based on image size and processing complexity (such as uploading a small preview image first, and then uploading the original large image), thereby maximizing the utilization of network resources and optimizing the user experience.

[0070] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An intelligent processing system for batch uploading pictures, characterized in that, The client module and the server module are included; The client module is used for providing a graphical user interface, receiving user input, and performing local processing of pictures; The server module is used for receiving and storing pictures and related data uploaded from the client; The client module includes: A picture collection and cache unit is used for receiving and temporarily storing a set of original pictures selected by a user; An intelligent recognition engine integrated with a lightweight neural network model is used for quality assessment, content classification, and duplicate detection analysis of pictures; A policy decision maker is used for assigning appropriate processing strategies to each picture according to the output of the intelligent recognition engine; A picture processor is used for performing filtering, marking, and image enhancement operations of pictures according to the instructions of the policy decision maker; An upload management unit is responsible for managing an upload queue, scheduling upload tasks, and implementing network adaptive upload control; A report generator is used for generating a visual upload processing report after the completion of an upload task. 2.The picture batch uploading intelligent processing system of claim 1, wherein, The server module further includes a user behavior analysis unit used for collecting and analyzing user feedback operations on the upload processing report; the user behavior analysis unit feeds back the analysis results to the policy decision maker of the client module, so as to optimize and adjust the matching rules of subsequent picture processing strategies, and realize self-learning and continuous optimization of the processing strategies. 3.The picture batch uploading intelligent processing system of claim 1, wherein, The lightweight neural network model in the intelligent recognition engine adopts an image classification model based on the MobileNetV3 architecture, and is optimized in model quantization and pruning for mobile terminal devices; the model is pre-trained by using massive image data on the server side, and is periodically updated and fine-tuned by feedback data collected by the user behavior analysis unit, and is then synchronized to each client through a model distribution mechanism, so as to continuously improve the recognition accuracy and adapt to new picture types.

4. The system of claim 1, wherein, The upload management unit includes a network state perception subunit and a queue scheduling subunit; the network state perception subunit continuously detects the round-trip time and packet loss rate of the current network, and constructs a comprehensive network health score in combination with the available bandwidth estimation value; the queue scheduling subunit dynamically adjusts the number of concurrent upload threads according to the network health score, increases the number of concurrent threads to speed up the upload when the score is high, and reduces the number of concurrent threads to maintain the stability of the connection when the score is low; at the same time, the subunit also sorts the upload queue by using a weighted fair queue algorithm according to the file size and priority label of the pictures, to ensure that small files and high-priority pictures are transmitted in time.

5. An intelligent processing method for batch uploading pictures, characterized in that, The steps include: S1, user interaction and picture collection: receiving user batch picture selection operations through a graphical user interface, and obtaining a set of original pictures to be uploaded; S2, local preprocessing and intelligent recognition analysis: performing local preprocessing and intelligent recognition analysis on each picture in the set of original pictures; the local preprocessing at least includes parsing metadata information of the pictures; The intelligent recognition analysis is performed by using the lightweight neural network model loaded to the client, and the identified content at least includes picture quality assessment, subject content classification, and duplicate picture detection; S3, processing strategy automatic matching and distribution: based on the intelligent recognition analysis result of step S2, automatically matching and distributing appropriate processing strategies for each picture; the processing strategies at least include one or more of the following: marking the pictures with low quality or low resolution as "to be confirmed", identifying and recommending to keep the best version of the detected duplicate picture group, automatically generating keyword tags according to the subject content classification result, and specifying subsequent image enhancement operation parameters for specific type pictures according to preset rules; S4, automatic processing execution under the strategy driving: according to the processing strategies distributed in step S3, performing corresponding automatic processing operations on the pictures in the original picture set to generate a filtered and processed queue to be uploaded; The automatic processing operations include: moving the pictures marked as "to be confirmed" to an independent list for user's final decision, automatically selecting the highest quality picture from the duplicate picture group to keep in the queue and removing other pictures, and adding keyword tags generated in step S3 to the pictures; S5, network adaptive intelligent upload scheduling: starting a batch upload task to upload the pictures in the queue to be uploaded generated in step S4 to the server; during the uploading process, the network bandwidth and server response state are monitored in real time, and the number of concurrent threads and the priority of the upload order are dynamically adjusted based on the monitoring results and the attributes of the pictures themselves; S6, upload result processing report generation and feedback: after the uploading task is completed, the system generates and presents an upload processing report; the report lists the successfully uploaded pictures, the filtered pictures and the filtering reasons, and the summary of the executed automatic processing operations in detail, and provides the user with a final operation interface for the "to be confirmed" picture list.

6. The intelligent processing method for bulk uploading pictures according to claim 5, characterized in that, In step S2, the picture quality evaluation specifically includes: calculating the definition index, brightness distribution histogram and color contrast of the picture, and comparing these indexes with the preset quality threshold to output a comprehensive quality score; the subject content classification is to identify the picture as one or more categories of landscape, portrait, document, food, and commodity; the duplicate picture detection is achieved by calculating the perceptual hash value of the picture and comparing the Hamming distance between the hash values.

7. The intelligent processing method for bulk uploading pictures of claim 5, wherein, In step S3, the automatically generated keyword tags according to the subject content classification result specifically include: establishing a mapping relationship between content classification and keyword library, automatically generating "industrial" label when the picture is classified as "mechanical equipment", and automatically generating "agricultural and sideline products" label when it is classified as "fruits and vegetables"; at the same time, the geographical location label is supplemented according to the geographical location information in the picture metadata.

8. The intelligent processing method for bulk uploading pictures according to claim 7, characterized in that, In step S3, the subsequent image enhancement operation parameters specified for specific type pictures according to preset rules include: specifying higher sharpening parameters and black and white binary parameters for pictures classified as "documents" to improve the readability of the text; specifying skin smoothing and beautification parameters for pictures classified as "portraits"; and specifying saturation enhancement and contrast optimization parameters for pictures classified as "landscapes".

9. The intelligent processing method for bulk uploading pictures of claim 5, wherein, In step S5, the "dynamic adjustment of the number of concurrent threads and the priority of the upload order" specifically includes: S51: According to the real-time monitored network available bandwidth, the ideal throughput of a single uploading thread is calculated, and the number of concurrent uploading threads is dynamically increased or decreased accordingly to maximize the utilization of available bandwidth without causing network congestion; S52: According to the file size of the picture and the complexity of the processing strategy, the uploading priority is set, wherein the picture with small file size and without complex image enhancement processing on the client is uploaded in priority, so that the user can see the first batch of uploading results faster; and the picture with large file size or requiring time-consuming processing is uploaded later.

Images