Image quality detection method and system

By improving the deep learning algorithm model and combining the feature pyramid network and multi-scale attention mechanism, the noise sensitivity and interpretability problems of image quality detection in the existing technology are solved, and higher accuracy and stronger adaptability of image quality detection are achieved.

CN119006990APending Publication Date: 2024-11-22STATE GRID INTELLIGENCE TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411138431.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-19
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

Existing image quality detection technologies are sensitive to noise, easily affected by external interference, have poor interpretability of detection results, lack versatility and adaptability, and have low detection accuracy.

Method used

An improved deep learning algorithm model is adopted, which performs image quality detection through backbone network, neck network and head network. It combines feature pyramid network and multi-scale attention mechanism, uses grayscale image histogram method to judge image brightness, and designs new loss function and data augmentation technology.

Benefits of technology

It improves the accuracy and interpretability of image quality detection, reduces sensitivity to noise, and enhances the intelligence and adaptability of detection results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119006990A_ABST
    Figure CN119006990A_ABST
Patent Text Reader

Abstract

The invention provides an image quality detection method and system, and belongs to the technical field of image processing. The image quality detection method comprises the following steps: obtaining a first image quality detection result after an image to be detected sequentially passes through a backbone network, a neck network and a head network of a deep learning network; feature extraction is carried out in the neck network through a feature pyramid network, and multi-scale information is extracted in the neck network through a multi-scale attention mechanism; according to the method, various quality problems in the image can be automatically learned, extracted and recognized, the sensitivity to noise is low, the interpretability of a quality detection result is high, and the precision of image quality detection is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to an image quality detection method and system. Background Art

[0002] The statements in this section merely provide background art related to the present invention and do not necessarily constitute prior art.

[0003] Patent No. CN113628203B discloses an image quality detection scheme, which determines whether there is a quality problem in the image to be detected by the ratio of the number of overexposed and overdark areas to the total number of areas; Patent No. CN113409278B discloses an image quality detection scheme, which determines the quality of the image to be processed based on the area information of the text area and the Euclidean distance between each text area and the light source, and determines the quality detection result of the image to be processed based on the quality of the image to be processed and a preset quality condition.

[0004] The existing technology represented by the above scheme has the following problems: it is sensitive to noise in the image and easily affected by external interference, which reduces the accuracy of detection; the detection results are poorly interpretable and have certain black box characteristics, which is not conducive to users understanding the reasons and basis of the detection results; it is dependent on specific environments, equipment or scenarios, and is not universal and applicable, which limits its application in different scenarios; the detection accuracy of existing image quality detection algorithms is generally not high and the adaptability is poor. Summary of the invention

[0005] In order to address the shortcomings of the prior art, the present invention provides an image quality detection method and system, which can automatically learn, extract and identify various quality problems in images through an improved deep learning algorithm model. It has low sensitivity to noise, weak dependence on specific environments, equipment or scenes, and strong interpretability of quality detection results, thereby improving the accuracy of image quality detection.

[0006] In order to achieve the above object, the present invention adopts the following technical solution:

[0007] In a first aspect, the present invention provides an image quality detection method.

[0008] An image quality detection method includes the following steps:

[0009] Acquire an image to be detected;

[0010] The image to be tested passes through the backbone network, neck network and head network of the deep learning network in turn to obtain the first image quality detection result; feature extraction is performed in the neck network through the feature pyramid network, and multi-scale information is extracted in the neck network through a multi-scale attention mechanism.

[0011] As a further limitation of the first aspect of the present invention, the backbone network includes a first convolutional layer, a first MBConv1 layer, a first MBConv6 layer, a second MBConv6 layer, a third MBConv6 layer, a fourth MBConv6 layer, a fifth MBConv6 layer, a sixth MBConv6 layer, a first MBConv6 layer, a second convolutional layer and a first fully connected layer connected in sequence;

[0012] The neck network includes an SPPF module, an FPN module, a first connection layer, a first c2f module, a SENet network, a second connection layer, a second c2f module, a third convolutional layer, a third connection layer, a third c2f module, a fourth convolutional layer, a fourth connection layer and a fourth c2f module, which are connected in sequence;

[0013] The head network includes three branches, the first branch includes the first Dropout layer, the fifth convolution layer and the first GIoU layer connected in sequence, the second branch includes the second Dropout layer, the sixth convolution layer and the second GIoU layer connected in sequence, and the third branch includes the third Dropout layer, the seventh convolution layer and the third GIoU layer connected in sequence;

[0014] The fourth MBConv6 layer is connected to the second connection layer, the second convolutional layer is connected to the first connection layer, the first fully connected layer is connected to the SPPF module, the SPPF module is connected to the fourth connection layer, and the first c2f module is connected to the third connection layer;

[0015] The first c2f module is connected to the first Dropout layer of the first branch, the third c2f module is connected to the second Dropout layer of the second branch, and the fourth c2f module is connected to the third Dropout layer of the third branch.

[0016] As a further limitation of the first aspect of the present invention, the first image quality detection result includes: image blur, image color cast, noise interference, stripe interference, video occlusion and screen freeze.

[0017] As a further limitation of the first aspect of the present invention, the following process is also included:

[0018] The image to be detected is calculated using a grayscale image histogram method to obtain a grayscale mean and a grayscale variance of the grayscale image corresponding to the image to be detected, and a second image quality detection result is obtained based on the grayscale mean and the grayscale variance.

[0019] As a further limitation of the first aspect of the present invention, the second image quality detection result includes: signal loss, image too bright, image too dark, black and white image and low contrast.

[0020] In a second aspect, the present invention provides an image quality detection system.

[0021] An image quality detection system, comprising:

[0022] The image acquisition unit is configured to: acquire an image to be detected;

[0023] The quality detection unit is configured as follows: after the image to be detected passes through the backbone network, neck network and head network of the deep learning network in sequence, a first image quality detection result is obtained; feature extraction is performed in the neck network through a feature pyramid network, and multi-scale information is extracted in the neck network through a multi-scale attention mechanism.

[0024] As a further limitation of the second aspect of the present invention, the backbone network includes a first convolutional layer, a first MBConv1 layer, a first MBConv6 layer, a second MBConv6 layer, a third MBConv6 layer, a fourth MBConv6 layer, a fifth MBConv6 layer, a sixth MBConv6 layer, a first MBConv6 layer, a second convolutional layer and a first fully connected layer connected in sequence;

[0025] The neck network includes an SPPF module, an FPN module, a first connection layer, a first c2f module, a SENet network, a second connection layer, a second c2f module, a third convolutional layer, a third connection layer, a third c2f module, a fourth convolutional layer, and a fourth connection layer, which are connected in sequence;

[0026] The head network includes three branches, the first branch includes the first Dropout layer, the fifth convolution layer and the first GIoU layer connected in sequence, the second branch includes the second Dropout layer, the sixth convolution layer and the second GIoU layer connected in sequence, and the third branch includes the third Dropout layer, the seventh convolution layer and the third GIoU layer connected in sequence;

[0027] The fourth MBConv6 layer is connected to the second connection layer, the second convolutional layer is connected to the first connection layer, the first fully connected layer is connected to the SPPF module, the SPPF module is connected to the fourth connection layer, and the first c2f module is connected to the third connection layer;

[0028] The first c2f module is connected to the first Dropout layer of the first branch, the third c2f module is connected to the second Dropout layer of the second branch, and the fourth c2f module is connected to the third Dropout layer of the third branch.

[0029] In a third aspect, the present invention provides a computer device, comprising: a processor and a computer-readable storage medium;

[0030] a processor adapted to execute a computer program;

[0031] A computer-readable storage medium, wherein a computer program is stored in the computer-readable storage medium, and when the computer program is executed by the processor, the image quality detection method as described in the first aspect of the present invention is implemented.

[0032] In a fourth aspect, the present invention provides a computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program, and the computer program is suitable for being loaded by a processor and executing the image quality detection method as described in the first aspect of the present invention.

[0033] In a fifth aspect, the present invention provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the image quality detection method as described in the first aspect of the present invention.

[0034] Compared with the prior art, the present invention has the following beneficial effects:

[0035] The present invention innovatively proposes an image quality detection method and designs a new deep learning model. Through the improved deep learning algorithm model, various quality problems in the image can be automatically learned, extracted and identified, including image blur, stripes, snowflakes, etc., thereby realizing intelligent detection and classification of image quality; it has low sensitivity to noise, strong interpretability of quality detection results, and improves the accuracy of image quality detection; it has higher accuracy and intelligence, can better adapt to complex and changeable image quality detection scenarios, and provide more reliable and effective solutions for image processing and application fields.

[0036] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0038] Figure 1 A schematic diagram of a flow chart of an image quality detection method provided in Embodiment 1 of the present invention;

[0039] Figure 2 A schematic diagram of the structure of a deep learning network provided in Example 1 of the present invention;

[0040] Figure 3 A schematic diagram of the process of grayscale image histogram detection provided in Example 1 of the present invention;

[0041] Figure 4 A schematic diagram of an image quality detection system provided in Embodiment 2 of the present invention;

[0042] Figure 5 A schematic diagram of a computer device provided in Example 3 of the present invention. DETAILED DESCRIPTION

[0043] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0044] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0045] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0046] Embodiment 1:

[0047] This implementation proposes an image quality detection method. Deep learning has a wide range of applications in image processing and recognition, and has outstanding performance in the detection of image blur, image color cast, noise interference, stripe interference, etc. Deep learning can be trained with a large number of abnormal images and clear images, and a new model can be obtained by using an improved backbone network, neck network and head network. This model can effectively detect the degree of abnormality of the image and can effectively achieve this goal.

[0048] Specifically, Figure 1 As shown, the following process is included:

[0049] S1: Acquire the image to be detected;

[0050] S2: After the image to be detected passes through the backbone network, neck network and head network of the deep learning network in sequence, a first image quality detection result is obtained.

[0051] The deep learning network of this implementation method includes:

[0052] 1) Backbone network: Use the more powerful backbone network EfficientNet to extract richer features, perform fine-tuning or transfer learning for target detection tasks, and improve the performance of the network on target detection tasks.

[0053] 2) Neck network, adds a multi-scale attention mechanism SENet, enhances feature fusion and extracts multi-scale information, and combines the attention mechanism and feature pyramid network (FPN) to better process features of different scales.

[0054] 3) Head network, adjust the design of Anchor Box and optimize it according to the characteristics of the target detection task.

[0055] 4) Loss function: Use the improved loss function GIoU loss to improve the accuracy of bounding box regression and combine it with multi-task learning to improve the generalization ability and accuracy of the model.

[0056] 5) Data enhancement and regularization: Use the more advanced data enhancement technology MixUp to increase the diversity of training data, and add regularization technologies such as Dropout layer and Batch Normalization to prevent overfitting.

[0057] In this implementation, more specifically, Figure 2 As shown, the backbone network includes a first convolutional layer, a first MBConv1 layer, a first MBConv6 layer, a second MBConv6 layer, a third MBConv6 layer, a fourth MBConv6 layer, a fifth MBConv6 layer, a sixth MBConv6 layer, a first MBConv6 layer, a second convolutional layer and a first fully connected layer, which are connected in sequence;

[0058] The neck network includes an SPPF module, an FPN module, a first connection layer, a first c2f module, a SENet network, a second connection layer, a second c2f module, a third convolutional layer, a third connection layer, a third c2f module, a fourth convolutional layer, a fourth connection layer and a fourth c2f module, which are connected in sequence;

[0059] The head network includes three branches, the first branch includes the first Dropout layer, the fifth convolution layer and the first GIoU layer connected in sequence, the second branch includes the second Dropout layer, the sixth convolution layer and the second GIoU layer connected in sequence, and the third branch includes the third Dropout layer, the seventh convolution layer and the third GIoU layer connected in sequence;

[0060] The fourth MBConv6 layer is connected to the second connection layer, the second convolutional layer is connected to the first connection layer, the first fully connected layer is connected to the SPPF module, the SPPF module is connected to the fourth connection layer, and the first c2f module is connected to the third connection layer;

[0061] The first c2f module is connected to the first Dropout layer of the first branch, the third c2f module is connected to the second Dropout layer of the second branch, and the fourth c2f module is connected to the third Dropout layer of the third branch.

[0062] As a further limitation of the first aspect of the present invention, the first image quality detection result includes: image blur, image color cast, noise interference, stripe interference, video occlusion and screen freeze.

[0063] In this implementation, MBConv1 refers to the MBCon layer without expanding the network width, and MBConv6 refers to the layer with expanding the network width to 6 times.

[0064] In this implementation, the c2f module includes the following functions:

[0065] (1) Feature transformation: The C2f module performs feature transformation on the input data through two convolutional layers (cv1 and cv2). These convolution operations help extract features of different levels and abstractions in the input data.

[0066] (2) Branch processing: The C2f module divides the input data into two branches for processing, one of which is directly passed to the output, and the other is processed by multiple Bottleneck modules. Such branch design helps to increase the nonlinear ability and representation ability of the network, thereby improving the network's ability to model complex data.

[0067] (3) Feature fusion: The C2f module achieves feature fusion by splicing features from different branches in the channel dimension. The spliced ​​features will contain information from different branches, enriching the expressive power of the features.

[0068] The SENet in this implementation is called Squeeze-and-Excitation Networks. SENet is inserted into the deep learning network of this implementation to explicitly model the interdependence between feature channels without introducing new spatial dimensions to fuse feature channels. The importance of each feature channel is automatically acquired through learning, and then useful features are promoted according to this importance and features that are not very useful for the current task are suppressed. The core idea of ​​SENet is to learn feature weights through the network according to the loss, so that effective feature maps have large weights and invalid or ineffective feature maps have small weights, so as to train the model to achieve better results.

[0069] The training steps of a deep learning network include:

[0070] S2.1: Data preparation.

[0071] a) Collect image datasets containing different quality issues, such as blur, stripes, snow, black and white images, etc.;

[0072] b) Write image functions such as blur, stripes, snow, black and white images, and automatically generate sample images;

[0073] c) Perform data preprocessing, including image size standardization, noise removal, etc.;

[0074] d) Label the dataset, mark the type and location of quality issues in the images, and provide labeled data for training the model.

[0075] S2.2: Model construction.

[0076] Build a deep learning network and design the network structure to identify quality issues such as blur, stripes, snow, black and white images, etc.

[0077] S2.3: Data division and training.

[0078] a) Divide the dataset into training set, validation set and test set;

[0079] b) Use the training set to train the model and optimize the model weights and parameters through back propagation algorithm and gradient descent;

[0080] c) Use the validation set to adjust hyperparameters to optimize model performance and prevent overfitting.

[0081] S2.4: Model evaluation and tuning.

[0082] a) Use the test set to evaluate the performance of the trained model, including metrics such as accuracy, recall, and F1 score;

[0083] b) Based on the evaluation results, the model is tuned to improve its performance in image quality detection tasks.

[0084] S2.5: Image quality detection.

[0085] a) Deploy the trained model to the actual environment and perform quality detection on the input images;

[0086] b) The model can automatically detect and identify quality problems such as blur, stripes, snow, black and white in the image, and output the corresponding recognition results (image blur, image color cast, noise interference, stripe interference, video occlusion, screen freeze).

[0087] S2.6: Experimental verification and optimization.

[0088] a) Conduct experimental verification in different scenarios, collect feedback data and continuously optimize the model;

[0089] b) Based on experimental results and user needs, further adjust model parameters, network structure and training data to continuously improve the accuracy and stability of image quality detection.

[0090] Through the above process, the present invention realizes a method for detecting and identifying image quality based on a deep learning network, which can effectively identify and locate quality problems in images, and provides an intelligent and efficient solution for image quality management and application fields.

[0091] like Figure 3 As shown in the figure, the brightness discrimination technology based on the grayscale image histogram determines the brightness of the image by calculating and analyzing the grayscale histogram of the image, that is, the distribution of the pixel grayscale values. In the histogram, the horizontal axis represents the grayscale level (usually 0 to 255). By analyzing the histogram, the approximate brightness of the image can be obtained. In addition to analyzing the pixel distribution, it also calculates the grayscale mean and variance of the image. Among them, the grayscale mean is the average value of all pixel grayscale values, which can represent the overall brightness of the image; the grayscale variance is the variance of the pixel grayscale value, which can represent the contrast or clarity of the image. A higher grayscale mean indicates a brighter image, and a higher grayscale variance indicates that the image has a higher contrast or clarity. The combination of these two parameters can more accurately determine the brightness and contrast of the image, which is helpful for image analysis and processing (signal loss, image too bright, image too dark, black and white image, low contrast, etc.).

[0092] Embodiment 2:

[0093] like Figure 4 As shown, this implementation provides an image quality detection system, including:

[0094] The image acquisition unit is configured to: acquire an image to be detected;

[0095] The quality detection unit is configured as follows: after the image to be detected passes through the backbone network, neck network and head network of the deep learning network in sequence, a first image quality detection result is obtained; feature extraction is performed in the neck network through a feature pyramid network, and multi-scale information is extracted in the neck network through a multi-scale attention mechanism.

[0096] The specific results and training process of the deep learning network are described in Example 1 and will not be repeated here.

[0097] It is understandable that the above-mentioned units can be separately or completely combined into one or several other units to constitute, or one (some) of the units can be further divided into multiple smaller units in function to constitute, which can achieve the same operation without affecting the realization of the technical effects of the embodiments of the present application. The above-mentioned units are divided based on logical functions. In practical applications, the function of one unit can also be implemented by multiple units, or the functions of multiple units can be implemented by one unit. In other embodiments of the present application, the image quality detection system may also include other units. In practical applications, these functions can also be implemented with the assistance of other units, and can be implemented by the collaboration of multiple units.

[0098] According to another embodiment of the present application, the system described in this embodiment can be constructed, and the image quality detection method of the embodiment of the present application can be implemented by running a computer program (including program code) capable of executing the steps involved in the corresponding method described in Example 1 on a general computing device such as a computer including processing elements and storage elements such as a central processing unit (CPU), random access memory (RAM), and read-only memory (ROM). The computer program can be recorded on, for example, a computer-readable recording medium, and loaded into the above-mentioned computing device through the computer-readable recording medium and run therein.

[0099] Embodiment 3:

[0100] like Figure 5 As shown, this implementation provides an electronic device, which includes a processor 1001, a communication interface 1002, and a computer-readable storage medium 1003. The processor 1001, the communication interface 1002, and the computer-readable storage medium 1003 may be connected via a bus or other means.

[0101] Among them, the communication interface 1002 is used to receive and send data, the computer-readable storage medium 1003 can be stored in the memory of the electronic device, the computer-readable storage medium 1003 is used to store a computer program, the computer program includes program instructions, and the processor 1001 is used to execute the program instructions stored in the computer-readable storage medium 1003.

[0102] Processor 1001 (or CPU (Central Processing Unit)) is the computing core and control core of the electronic device, which is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions to implement corresponding method processes or corresponding functions.

[0103] The processor 1001 is configured to execute the following process:

[0104] An image to be detected is obtained, and the image to be detected passes through the backbone network, neck network and head network of the deep learning network in sequence to obtain a first image quality detection result; feature extraction is performed in the neck network through a feature pyramid network, and multi-scale information is extracted in the neck network through a multi-scale attention mechanism.

[0105] The specific results and training process of the deep learning network are described in Example 1 and will not be repeated here.

[0106] Embodiment 4:

[0107] This implementation provides a computer-readable storage medium (Memory), which is a memory device in an electronic device for storing programs and data. It is understandable that the computer-readable storage medium here can include both built-in storage media in the electronic device and, of course, extended storage media supported by the electronic device. The computer-readable storage medium provides a storage space that stores the processing system of the electronic device.

[0108] In addition, the storage space also stores one or more instructions suitable for being loaded and executed by the processor, and these instructions may be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here may be a high-speed RAM memory, or a non-volatile memory (non-volatile memory), such as at least one disk storage; optionally, it may also be at least one computer-readable storage medium located away from the aforementioned processor.

[0109] In one embodiment, the computer-readable storage medium stores one or more instructions; the processor loads and executes the one or more instructions stored in the computer-readable storage medium to implement the following process:

[0110] An image to be detected is obtained, and the image to be detected passes through the backbone network, neck network and head network of the deep learning network in sequence to obtain a first image quality detection result; feature extraction is performed in the neck network through a feature pyramid network, and multi-scale information is extracted in the neck network through a multi-scale attention mechanism.

[0111] The specific results and training process of the deep learning network are described in Example 1 and will not be repeated here.

[0112] Embodiment 5:

[0113] The present implementation provides a computer program product or a computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the electronic device performs the following process:

[0114] An image to be detected is obtained, and the image to be detected passes through the backbone network, neck network and head network of the deep learning network in sequence to obtain a first image quality detection result; feature extraction is performed in the neck network through a feature pyramid network, and multi-scale information is extracted in the neck network through a multi-scale attention mechanism.

[0115] The specific results and training process of the deep learning network are described in Example 1 and will not be repeated here.

[0116] A person skilled in the art can appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in this application can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. A person skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0117] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loading and executing computer program instructions on a computer, the process or function according to the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network or other programmable device. The computer instruction can be stored in a computer-readable storage medium or transmitted by a computer-readable storage medium. The computer instruction can be transmitted from a website site, a computer, a server or a data center to another website site, a computer, a server or a data center by wired (for example, coaxial cable, optical fiber, digital line (DSL)) or wireless (for example, infrared, wireless, microwave, etc.) mode. The computer-readable storage medium can be any available medium that a computer can access or a data processing device such as a server, a data center, etc. that contains one or more available media integration. Available media can be magnetic media (for example, floppy disk, hard disk, tape), optical media (for example, DVD), or semiconductor media (for example, solid-state drive (Solid State Disk, SSD)) and the like.

[0118] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An image quality detection method, characterized in that: The process includes: Acquire an image to be detected; The image to be tested passes through the backbone network, neck network and head network of the deep learning network in turn to obtain the first image quality detection result; feature extraction is performed in the neck network through the feature pyramid network, and multi-scale information is extracted in the neck network through a multi-scale attention mechanism.

2. The image quality detection method according to claim 1, characterized in that: The backbone network includes a first convolutional layer, a first MBConv1 layer, a first MBConv6 layer, a second MBConv6 layer, a third MBConv6 layer, a fourth MBConv6 layer, a fifth MBConv6 layer, a sixth MBConv6 layer, a first MBConv6 layer, a second convolutional layer and a first fully connected layer, which are connected in sequence; The neck network includes an SPPF module, an FPN module, a first connection layer, a first c2f module, a SENet network, a second connection layer, a second c2f module, a third convolutional layer, a third connection layer, a third c2f module, a fourth convolutional layer, a fourth connection layer and a fourth c2f module, which are connected in sequence; The head network includes three branches, the first branch includes the first Dropout layer, the fifth convolution layer and the first GIoU layer connected in sequence, the second branch includes the second Dropout layer, the sixth convolution layer and the second GIoU layer connected in sequence, and the third branch includes the third Dropout layer, the seventh convolution layer and the third GIoU layer connected in sequence; The fourth MBConv6 layer is connected to the second connection layer, the second convolutional layer is connected to the first connection layer, the first fully connected layer is connected to the SPPF module, the SPPF module is connected to the fourth connection layer, and the first c2f module is connected to the third connection layer; The first c2f module is connected to the first Dropout layer of the first branch, the third c2f module is connected to the second Dropout layer of the second branch, and the fourth c2f module is connected to the third Dropout layer of the third branch.

3. The image quality detection method according to claim 1, characterized in that: The first image quality detection result includes: image blur, image color cast, noise interference, stripe interference, video occlusion and screen freeze.

4. The image quality detection method according to claim 1, characterized in that: It also includes the following processes: The image to be detected is calculated using a grayscale image histogram method to obtain a grayscale mean and a grayscale variance of the grayscale image corresponding to the image to be detected, and a second image quality detection result is obtained based on the grayscale mean and the grayscale variance.

5. The image quality detection method according to claim 2, characterized in that: The second image quality detection result includes: signal loss, image too bright, image too dark, black and white image, and low contrast.

6. An image quality detection system, characterized in that: include: The image acquisition unit is configured to: acquire an image to be detected; The quality detection unit is configured as follows: after the image to be detected passes through the backbone network, neck network and head network of the deep learning network in sequence, a first image quality detection result is obtained; feature extraction is performed in the neck network through a feature pyramid network, and multi-scale information is extracted in the neck network through a multi-scale attention mechanism.

7. The image quality detection system according to claim 6, characterized in that: The backbone network includes a first convolutional layer, a first MBConv1 layer, a first MBConv6 layer, a second MBConv6 layer, a third MBConv6 layer, a fourth MBConv6 layer, a fifth MBConv6 layer, a sixth MBConv6 layer, a first MBConv6 layer, a second convolutional layer and a first fully connected layer, which are connected in sequence; The neck network includes an SPPF module, an FPN module, a first connection layer, a first c2f module, a SENet network, a second connection layer, a second c2f module, a third convolutional layer, a third connection layer, a third c2f module, a fourth convolutional layer, and a fourth connection layer, which are connected in sequence; The head network includes three branches, the first branch includes the first Dropout layer, the fifth convolution layer and the first GIoU layer connected in sequence, the second branch includes the second Dropout layer, the sixth convolution layer and the second GIoU layer connected in sequence, and the third branch includes the third Dropout layer, the seventh convolution layer and the third GIoU layer connected in sequence; The fourth MBConv6 layer is connected to the second connection layer, the second convolutional layer is connected to the first connection layer, the first fully connected layer is connected to the SPPF module, the SPPF module is connected to the fourth connection layer, and the first c2f module is connected to the third connection layer; The first c2f module is connected to the first Dropout layer of the first branch, the third c2f module is connected to the second Dropout layer of the second branch, and the fourth c2f module is connected to the third Dropout layer of the third branch.

8. A computer device, characterized in that: include: a processor and a computer readable storage medium; a processor adapted to execute a computer program; A computer-readable storage medium, wherein a computer program is stored in the computer-readable storage medium, and when the computer program is executed by the processor, the image quality detection method according to any one of claims 1 to 5 is implemented.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is suitable for being loaded by a processor and executing the image quality detection method according to any one of claims 1 to 5.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the image quality detection method according to any one of claims 1 to 5 is implemented.