Intelligent detection method, system, device and product based on portrait monitoring

Abstract

The invention discloses an intelligent detection method, system, device and product based on portrait monitoring, and belongs to the field of portrait detection, and the method comprises the steps: obtaining image data, and detecting the image data based on an image data quality detection model; preprocessing the image data reaching a preset image quality standard to obtain basic image data; extracting the basic image data by adopting a texture feature extraction model to obtain a texture direction feature code; segmenting the basic image data by using an image segmentation algorithm to obtain a plurality of pieces of first sub-image data, calculating the occurrence frequency of each texture direction feature code in the first sub-image data, and constructing a feature vector of the basic image data according to the occurrence frequency; and classifying the feature vectors based on a portrait detection model to obtain a classification result, judging whether the classification result contains a portrait or not, and if yes, labeling the portrait by using a predefined anchor frame. The method can reduce the influence of the illumination condition on the detection accuracy, improves the detection efficiency, and achieves the efficient detection.

Description

Technical Field

[0001] The present invention belongs to the technical field of human portrait detection, and in particular relates to an intelligent detection method, system, device and product based on human portrait monitoring. Background Art

[0002] With the development and progress of science and technology, portrait detection has been widely used in image processing, artificial intelligence, machine learning and other fields. Portrait detection technology uses a camera or a camera to collect images or video streams containing portraits, and automatically detects and tracks portraits in the image, and then performs a series of related technologies on the detected portraits. Commonly used methods are feature-based and deep learning methods, which can identify the identities of different portraits through memory and comparison. However, the current portrait detection technology is affected by lighting conditions. For example, the degree of feature appearance of portraits detected during the day is different from that of portraits detected at night, and the degree of feature appearance of portraits detected indoors is different from that of portraits detected outdoors. These lighting conditions cause large differences in visual images, thereby affecting the accuracy and efficiency of portrait detection.

[0003] Therefore, how to provide an effective solution to achieve an intelligent detection method based on portrait monitoring that reduces the impact of lighting conditions has become a difficult problem that needs to be solved urgently in the existing technology. Summary of the invention

[0004] The purpose of the present invention is to provide an intelligent detection method, system, device and product based on portrait monitoring, so as to solve the problem in the prior art that lighting conditions affect the accuracy and efficiency of detection.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] In a first aspect, the present invention provides an intelligent detection method based on portrait monitoring, comprising:

[0007] Acquire image data, and detect the image data based on an image data quality detection model to obtain image data that meets a preset image data quality standard;

[0008] Performing preprocessing operations on image data that meet preset image data quality standards to obtain basic image data;

[0009] The texture feature extraction model is used to extract the texture features of the basic image data to obtain the texture direction feature encoding;

[0010] The basic image data is segmented using an image segmentation algorithm to obtain a plurality of first sub-image data, the occurrence frequency of each texture direction feature code in the first sub-image data is calculated, and a feature vector of the basic image data is constructed according to the occurrence frequency;

[0011] The feature vector is classified based on the portrait detection model to obtain the classification result, and it is determined whether the classification result contains the portrait. If so, the portrait is annotated using a predefined anchor box in the basic image data.

[0012] In a possible design, a texture feature extraction model is used to extract texture features from basic image data to obtain texture direction feature encoding, including:

[0013] Divide the basic image data into a plurality of 3×3 rectangular windows according to the pixel points to obtain a plurality of second sub-image data;

[0014] For each second sub-image data, extracting texture features of the second sub-image data based on a texture feature extraction model to obtain texture feature encoding of the second sub-image data;

[0015] Calculating gradient information of pixels in the second sub-image data;

[0016] The texture feature coding is integrated with the gradient information to obtain the texture direction feature coding of the second sub-image data.

[0017] In a possible design, extracting texture features from the second sub-image data based on a texture feature extraction model to obtain texture feature encoding of the second sub-image data includes:

[0018] Acquire the grayscale value of a central pixel of the second sub-image data and the grayscale values ​​of pixels in eight areas adjacent to the central pixel;

[0019] Compare the grayscale value of the central pixel of the second sub-image data with the grayscale values ​​of the pixels in eight neighborhoods to obtain a binary sequence;

[0020] The binary sequence is converted into a decimal number to obtain a texture feature code of the central pixel, and the texture feature code of the central pixel is used as the texture feature code of the second sub-image data.

[0021] In a possible design, the basic image data is segmented using an image segmentation algorithm to obtain a plurality of first sub-image data, the occurrence frequency of each texture direction feature code in the sub-image data is calculated, and a feature vector of the image data is obtained according to the occurrence frequency, including:

[0022] Segmenting the basic image data into a plurality of first sub-image data based on an image segmentation algorithm;

[0023] Calculating the occurrence frequency of each texture direction feature code in each first sub-image data;

[0024] A frequency statistical histogram is made according to the frequency of occurrence, all the histograms are connected, and the histogram features are used as the feature vectors of the basic image data.

[0025] In a possible design, the method further includes: training a portrait detection model, including:

[0026] Obtain multiple training samples;

[0027] A plurality of training samples are input into the portrait monitoring model for training to obtain a trained portrait detection model, and the trained portrait detection model is used to identify the feature vector.

[0028] In a possible design, a preprocessing operation is performed on the image data that meets the image data quality standard, including:

[0029] The image data that meets the image data quality standard is subjected to grayscale processing, denoising processing and cropping processing in sequence.

[0030] In a possible design, the texture feature extraction model is constructed based on the LBP algorithm;

[0031] The portrait detection model is built based on the AdaBoost model.

[0032] In a second aspect, the present invention provides an intelligent detection system based on portrait monitoring, comprising:

[0033] An acquisition detection module is used to acquire image data, detect the image data based on an image data quality detection model, obtain image data that meets a preset image data quality standard, and upload the image data that meets the image data quality standard to an image preprocessing module;

[0034] An image preprocessing module is used to perform preprocessing operations on image data that meets a preset image data quality standard to obtain basic image data, and upload the basic image data to a feature extraction module;

[0035] A feature extraction module is used to extract texture features from basic image data using a texture feature extraction model to obtain texture direction feature codes, and upload the texture direction feature codes to a feature integration module;

[0036] A feature integration module is used to segment the basic image data using an image segmentation algorithm to obtain a plurality of first sub-image data, calculate the occurrence frequency of each texture direction feature code in the first sub-image data, and construct a feature vector of the basic image data according to the occurrence frequency;

[0037] The classification and labeling module is used to classify the feature vector based on the portrait detection model, obtain the classification result, and determine whether the classification result contains the portrait. If it does, the portrait is labeled using a predefined anchor frame in the basic image data to achieve portrait monitoring.

[0038] In the third aspect, the present invention provides an intelligent detection device based on portrait monitoring, comprising a memory, a processor and a transceiver that are communicatively connected in sequence, wherein the memory is used to store computer programs, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the intelligent detection method based on portrait monitoring as described in any one of the above.

[0039] In a fourth aspect, the present invention provides a computer-readable storage medium having instructions stored thereon. When the instructions are run on a computer, the intelligent detection method based on portrait monitoring as described in any one of the above items is executed.

[0040] In a fifth aspect, the present invention provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute any one of the intelligent detection methods based on portrait monitoring as described above.

[0041] The beneficial effects of the present invention are as follows:

[0042] The present invention discloses an intelligent detection method, system, device and product based on portrait monitoring, comprising: acquiring image data, detecting the image data based on an image data quality detection model, and obtaining image data that meets a preset image data quality standard; performing preprocessing operations on the image data that meets the preset image data quality standard to obtain basic image data; using a texture feature extraction model to extract texture features from the basic image data to obtain texture direction feature codes; using an image segmentation algorithm to segment the basic image data to obtain multiple first sub-image data, calculating the occurrence frequency of each texture direction feature code in the first sub-image data, and constructing a feature vector of the basic image data according to the occurrence frequency; classifying the feature vector based on a portrait detection model to obtain a classification result, judging whether the classification result contains a portrait, and if so, using a predefined anchor frame to mark the portrait in the basic image data to achieve monitoring of the portrait. The present invention uses a texture feature extraction model to extract texture features from image data, obtains texture direction feature coding, reduces the impact of different lighting on portrait detection, and can better process image data when lighting changes unevenly. At the same time, it improves the accuracy of portrait detection, improves detection efficiency, and has strong robustness in the field of portrait detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 A flowchart of an intelligent detection method based on portrait monitoring provided by an embodiment of the present invention;

[0044] Figure 2 A module diagram of an intelligent detection system based on portrait monitoring provided by an embodiment of the present invention;

[0045] Figure 3 A schematic structural diagram of an intelligent detection device based on portrait monitoring provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be briefly introduced below in combination with the drawings and the description of the embodiments or the prior art. Obviously, the following description of the structure of the drawings is only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. It should be noted that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation of the present invention.

[0047] It should be understood that although the terms first and second, etc. may be used herein to describe various objects, these objects should not be limited by these terms. These terms are only used to distinguish one object from another object. For example, a first object can be referred to as a second object, and similarly, a second object can be referred to as a first object without departing from the scope of the exemplary embodiments of the present invention.

[0048] Example:

[0049] like Figure 1 As shown, the first aspect of this embodiment provides an intelligent detection method based on portrait monitoring, which can be executed by, but not limited to, a computer device or a virtual machine with certain computing resources, such as a personal computer or a smart phone, or a virtual machine; Figure 1 As shown, the intelligent detection method based on portrait monitoring may include but is not limited to the following steps:

[0050] S1. Acquire image data, receive preset image data quality standards, detect the image data based on the image data quality detection model, and obtain image data that meets the preset image data quality standards;

[0051] Among them, image data includes but is not limited to being collected by a camera with a resolution greater than or equal to a preset resolution, the preset image data quality standards include but are not limited to requirements for the resolution and focus level of the image data, and the image data quality detection model is constructed based on the naive Bayes classifier algorithm and the support vector machine algorithm.

[0052] S2. Performing a preprocessing operation on the image data that reaches the preset image data quality standard to obtain basic image data so as to extract features from the basic image data;

[0053] Specifically, preprocessing includes grayscale processing, denoising processing and cropping processing. Grayscale processing is the process of converting a color image into a grayscale image to simplify image data, prepare for subsequent processing, and improve computing speed; denoising processing includes but is not limited to removing noise generated by the acquisition equipment to improve the clarity and quality of the basic image data; cropping is to adjust the image data to a uniform size so that the image data can be further analyzed and processed.

[0054] S3. Using a texture feature extraction model to extract texture features from the basic image data, and obtain texture direction feature encoding, so as to integrate the features of the basic image data;

[0055] In a possible design, the texture feature extraction model is constructed based on the LBP algorithm;

[0056] Specifically, in step S3, a texture feature extraction model is used to extract texture features from the basic image data to obtain texture direction feature encoding so as to perform feature integration on the basic image data, including:

[0057] S301. Divide the basic image data into a plurality of 3×3 rectangular windows according to the pixels to obtain a plurality of second sub-image data;

[0058] In this embodiment, the pixel point is used as the geometric center of the rectangular window, the basic image data is divided into a plurality of 3×3 rectangular windows, and the basic image data in the rectangular windows is used as the second sub-image data.

[0059] S302. For each second sub-image data, extract texture features of the second sub-image data based on the texture feature extraction model to obtain texture feature encoding of the second sub-image data;

[0060] S303. Calculate the gradient information of the pixels in the second sub-image data;

[0061] Specifically, the gradient information includes the gradient magnitude and the gradient direction, where the gradient magnitude is:

[0062]

[0063] The gradient direction is:

[0064]

[0065] is the derivative of the pixel in the horizontal direction, that is, the gradient in the horizontal direction; is the derivative of the pixel in the vertical direction, that is, the gradient in the vertical direction.

[0066] S304. Fuse the texture feature code with the gradient information to obtain the texture direction feature code of the second sub-image data.

[0067] If the numerical difference range of the texture feature coding and the gradient information is large, the data is normalized so that the values ​​of the two are in the same range to facilitate their fusion. The normalized texture feature coding and the gradient information are spliced. If the dimension of the spliced ​​feature coding is too high, PCA (Principal Component Analysis) and other methods can be used to reduce the dimension in order to reduce the amount of calculation and avoid overfitting, so as to obtain the texture direction feature coding of the second sub-image data.

[0068] Furthermore, in step S302, texture feature extraction is performed on the second sub-image data based on the texture feature extraction model to obtain texture feature encoding of the second sub-image data, including:

[0069] S3021. Obtain the grayscale value of the central pixel of the second sub-image data and the grayscale values ​​of the pixels in eight areas adjacent to the central pixel;

[0070] S3022. Compare the grayscale value of the pixel at the center of the second sub-image data with the grayscale values ​​of the pixels in the eight neighborhoods to obtain a binary sequence;

[0071] The grayscale value of the central pixel of the second sub-image data is compared with the grayscale values ​​of the eight neighboring pixels. If the grayscale value of the neighboring pixel is greater than the grayscale value of the central pixel, the position of the pixel is marked as 1, otherwise it is marked as 0. Eight values ​​of 0 or 1 are generated in the second sub-image data. Starting from the upper left corner value, eight values ​​are collected clockwise to form an eight-bit binary number, which is a binary sequence.

[0072] S3023. Convert the binary sequence into a decimal number, and use the decimal number as the texture feature encoding of the pixel point.

[0073] S4. Use an image segmentation algorithm to segment the basic image data to obtain a plurality of first sub-image data, calculate the frequency of occurrence of each texture direction feature code in the first sub-image data, and construct a feature vector of the basic image data according to the frequency of occurrence;

[0074] The image segmentation algorithm includes but is not limited to using a Thresholding algorithm and a K-means algorithm.

[0075] Specifically, in step S4, the basic image data is segmented using an image segmentation algorithm to obtain a plurality of first sub-image data, the occurrence frequency of each texture direction feature code in the sub-image data is calculated, and a feature vector of the image data is obtained according to the occurrence frequency, including:

[0076] S401. Segmenting the basic image data into a plurality of first sub-image data based on an image segmentation algorithm;

[0077] S402. Calculate the occurrence frequency of each texture direction feature code in the first sub-image data;

[0078] The frequency of occurrence of different texture direction feature codes indicates the importance of different areas in the first sub-image data, wherein a high frequency indicates a dramatic change in grayscale value, which is used to represent detail parts; a low frequency indicates a slow change in grayscale value, which is used to represent contour parts and main structural information.

[0079] The occurrence frequencies of different texture direction feature codes in each first sub-image data are counted to obtain a frequency number equal to the number of first sub-image data.

[0080] S403. Create a frequency statistical histogram based on the frequency of occurrence, and connect all the histograms to form a feature vector of the basic image data.

[0081] A frequency statistical histogram is made according to the frequency of occurrence to obtain multiple frequency statistical histograms, and all the frequency statistical histograms are connected and used as feature vectors of basic image data.

[0082] S5. Classify the feature vector based on the portrait detection model to obtain the classification result, and determine whether the classification result contains the portrait. If so, use the predefined anchor frame in the basic image data to mark the portrait in order to monitor the portrait.

[0083] In a possible design, the method further includes: training a portrait detection model, including:

[0084] Obtain multiple training samples;

[0085] A plurality of training samples are input into the portrait monitoring model for training to obtain a trained portrait detection model, and the trained portrait detection model is used to identify the feature vector.

[0086] In a possible design, the portrait detection model is constructed based on an AdaBoost (Adaptive Boosting, iteration) model. The AdaBoost model constructs a strong classifier by combining multiple weak classifiers, thereby improving the accuracy of the model.

[0087] In summary, the intelligent detection method based on portrait monitoring of the present embodiment obtains image data, receives a preset image data quality standard, detects the image data based on an image data quality detection model, and obtains image data that meets the preset image data quality standard; performs preprocessing operations on the image data that meets the preset image data quality standard to obtain basic image data, so as to extract features from the basic image data; uses a texture feature extraction model to extract texture features from the basic image data to obtain texture direction feature codes, so as to perform feature integration on the basic image data; uses an image segmentation algorithm to segment the basic image data to obtain multiple first sub-image data, calculates the frequency of occurrence of each texture direction feature code in the first sub-image data, and constructs a feature vector of the basic image data according to the frequency of occurrence; classifies the feature vector based on the portrait detection model to obtain a classification result, and determines whether the classification result contains a portrait. If so, a predefined anchor frame is used in the basic image data to mark the portrait, so as to achieve monitoring of the portrait. This embodiment preprocesses the image data into a grayscale image, uses a texture feature extraction model to extract texture features from the image data, obtains texture direction feature encoding, reduces the impact of different lighting on portrait detection, and can better process image data even when lighting changes unevenly. At the same time, it improves the accuracy of portrait detection, improves detection efficiency, and has strong robustness in the field of portrait detection.

[0088] like Figure 2 As shown, the second aspect of this embodiment provides an intelligent detection system based on portrait monitoring, which is used to implement the intelligent detection method based on portrait monitoring in Example 1. The intelligent detection system includes:

[0089] An acquisition detection module is used to acquire image data, detect the image data based on an image data quality detection model, obtain image data that meets a preset image data quality standard, and upload the image data that meets the image data quality standard to an image preprocessing module;

[0090] An image preprocessing module is used to perform preprocessing operations on image data that meets a preset image data quality standard to obtain basic image data, and upload the basic image data to a feature extraction module;

[0091] A feature extraction module is used to extract texture features from basic image data using a texture feature extraction model to obtain texture direction feature codes, and upload the texture direction feature codes to a feature integration module;

[0092] A feature integration module is used to segment the basic image data using an image segmentation algorithm to obtain a plurality of first sub-image data, calculate the occurrence frequency of each texture direction feature code in the first sub-image data, and construct a feature vector of the basic image data according to the occurrence frequency;

[0093] The classification and labeling module is used to classify the feature vector based on the portrait detection model, obtain the classification result, and determine whether the classification result contains the portrait. If it does, the portrait is labeled using a predefined anchor frame in the basic image data to achieve portrait monitoring.

[0094] In summary, the present embodiment provides an intelligent detection system based on portrait monitoring, which uses an acquisition detection module to acquire image data, receives a preset image data quality standard, uses an image data quality detection model to detect the image data, and uploads the image data that meets the preset image data quality standard to an image preprocessing module; the image and processing module performs preprocessing operations on the image data that meets the preset image data quality standard to obtain basic image data, and uploads the basic image data to a feature extraction module; the feature extraction module uses a texture feature extraction model to extract texture features from the basic image data to obtain texture direction feature codes; the feature integration module uses an image segmentation algorithm to segment the basic image data to obtain multiple first sub-image data, calculates the occurrence frequency of each texture direction feature code in the first sub-image data, constructs a feature vector of the basic image data according to the occurrence frequency, and uploads the feature vector to a classification and annotation module; the classification and annotation module classifies the feature vector based on the portrait detection model, determines whether the image data contains a portrait, and if so, uses a predefined anchor frame to annotate the portrait in the image data to achieve portrait monitoring. This embodiment uses a texture feature extraction model to extract texture features from preprocessed image data, obtains texture direction feature encoding, avoids the influence of different lighting conditions on portrait detection, and can better process image data even when lighting changes unevenly, thereby improving the accuracy and efficiency of portrait detection, making this embodiment more robust in the field of portrait detection.

[0095] The third aspect of this embodiment provides an intelligent detection device based on portrait monitoring, including a memory, a processor and a transceiver that are communicatively connected in sequence, wherein the memory is used to store computer programs, the transceiver is used to send and receive messages, and the processor is used to read the computer program to execute the intelligent detection method based on portrait monitoring as described in the first aspect of the embodiment.

[0096] For specific examples, the memory may include but is not limited to random access memory (RAM), read-only memory (ROM), flash memory, first-in-first-out memory (FIFO) and / or first-in-last-out memory (FILO), etc.; the processor may be but is not limited to a microprocessor of the STM32F105 series, an ARM (Advanced RISC-Machines), an X86 or other architecture processor, or a processor with an integrated NPU (neural-network processing units); the transceiver may be but is not limited to a WiFi (Wireless Fidelity) wireless transceiver, a Bluetooth wireless transceiver, a General Packet Radio Service (GPRS) wireless transceiver, a ZigBee protocol (a low-power local area network protocol based on the IEEE802.15.4 standard, ZigBee) wireless transceiver, a 3G transceiver, a 4G transceiver and / or a 5G transceiver, etc.

[0097] The working process, working details and technical effects of the device provided in the third aspect of this embodiment can be referred to the basic intelligent detection method described in the first aspect of the embodiment, and will not be repeated here.

[0098] In a fourth aspect of this embodiment, a computer-readable storage medium is provided, on which instructions are stored, and when the instructions are executed on a computer, the intelligent detection method based on portrait monitoring as described in the first aspect of the embodiment is executed. The computer-readable storage medium refers to a carrier for storing data, which may include but is not limited to computer-readable storage media such as floppy disks, optical disks, hard disks, flash memory, USB flash drives, and / or memory sticks, and the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.

[0099] The working process, working details and technical effects of the aforementioned computer-readable storage medium provided in the fourth aspect of this embodiment can be referred to the intelligent detection method described in the first aspect, and will not be repeated here.

[0100] A fifth aspect of this embodiment provides a computer program product, including a computer program or instructions, which, when executed by a computer, are used to implement the intelligent detection method based on portrait monitoring as described in the first aspect of the embodiment.

[0101] The working process, working details and technical effects of the aforementioned computer program product provided in the fourth aspect of this embodiment can be referred to the intelligent detection method described in the first aspect, and will not be repeated here.

[0102] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the protection scope of the present invention. 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 intelligent detection method based on portrait monitoring, characterized in that: include: Acquire image data, and detect the image data based on an image data quality detection model to obtain image data that meets a preset image data quality standard; Performing preprocessing operations on image data that meet preset image data quality standards to obtain basic image data; The texture feature extraction model is used to extract the texture features of the basic image data to obtain the texture direction feature encoding; The basic image data is segmented using an image segmentation algorithm to obtain a plurality of first sub-image data, the occurrence frequency of each texture direction feature code in the first sub-image data is calculated, and a feature vector of the basic image data is constructed according to the occurrence frequency; The feature vector is classified based on the portrait detection model to obtain the classification result, and it is determined whether the classification result contains the portrait. If so, the portrait is annotated using a predefined anchor box in the basic image data.

2. The intelligent detection method based on portrait monitoring according to claim 1 is characterized in that: The texture feature extraction model is used to extract texture features from the basic image data to obtain texture direction feature encoding, including: Divide the basic image data into a plurality of 3×3 rectangular windows according to the pixel points to obtain a plurality of second sub-image data; For each second sub-image data, extracting texture features of the second sub-image data based on a texture feature extraction model to obtain texture feature encoding of the second sub-image data; Calculating gradient information of pixels in the second sub-image data; The texture feature coding is integrated with the gradient information to obtain the texture direction feature coding of the second sub-image data.

3. The intelligent detection method based on portrait monitoring according to claim 2 is characterized in that: Extracting texture features from the second sub-image data based on the texture feature extraction model to obtain texture feature encoding of the second sub-image data includes: Acquire the grayscale value of a central pixel of the second sub-image data and the grayscale values ​​of pixels in eight areas adjacent to the central pixel; Compare the grayscale value of the central pixel of the second sub-image data with the grayscale values ​​of the pixels in eight neighborhoods to obtain a binary sequence; The binary sequence is converted into a decimal number to obtain a texture feature code of the central pixel, and the texture feature code of the central pixel is used as the texture feature code of the second sub-image data.

4. The intelligent detection method based on portrait monitoring according to claim 1 is characterized in that: The basic image data is segmented using an image segmentation algorithm to obtain a plurality of first sub-image data, the occurrence frequency of each texture direction feature code in the sub-image data is calculated, and a feature vector of the image data is obtained according to the occurrence frequency, including: Segmenting the basic image data into a plurality of first sub-image data based on an image segmentation algorithm; Calculating the occurrence frequency of each texture direction feature code in each first sub-image data; A frequency statistical histogram is made according to the frequency of occurrence, and all the histograms are connected to form the feature vector of the basic image data.

5. The intelligent detection method based on portrait monitoring according to claim 1 is characterized in that: The method further includes: training a portrait detection model, including: Obtain multiple training samples; A plurality of training samples are input into the portrait monitoring model for training to obtain a trained portrait detection model, and the trained portrait detection model is used to identify the feature vector.

6. The intelligent detection method based on portrait monitoring according to claim 1 is characterized in that: Perform preprocessing operations on image data that meet image data quality standards, including: The image data that meets the image data quality standard is subjected to grayscale processing, denoising processing and cropping processing in sequence.

7. The intelligent detection method based on portrait monitoring according to claim 1 is characterized in that: The portrait detection model is built based on the AdaBoost model.

8. An intelligent detection system based on portrait monitoring, characterized in that: include: An acquisition detection module is used to acquire image data, detect the image data based on an image data quality detection model, obtain image data that meets a preset image data quality standard, and upload the image data that meets the image data quality standard to an image preprocessing module; An image preprocessing module is used to perform preprocessing operations on image data that meets a preset image data quality standard to obtain basic image data, and upload the basic image data to a feature extraction module; A feature extraction module is used to extract texture features from basic image data using a texture feature extraction model to obtain texture direction feature codes, and upload the texture direction feature codes to a feature integration module; A feature integration module is used to segment the basic image data using an image segmentation algorithm to obtain a plurality of first sub-image data, calculate the occurrence frequency of each texture direction feature code in the first sub-image data, and construct a feature vector of the basic image data according to the occurrence frequency; The classification and labeling module is used to classify the feature vector based on the portrait detection model, obtain the classification result, and determine whether the classification result contains the portrait. If it does, the portrait is labeled using a predefined anchor frame in the basic image data to achieve portrait monitoring.

9. An intelligent detection device based on portrait monitoring, characterized in that: It includes a memory, a processor and a transceiver which are communicatively connected in sequence, wherein the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program to execute the intelligent detection method based on portrait monitoring as described in any one of claims 1 to 7.

10. A computer program product comprising a computer program or instructions, characterized in that When the computer program or the instruction is executed by a computer, the intelligent detection method based on portrait monitoring as claimed in any one of claims 1 to 7 is implemented.

Images