Multi-target analysis method, device, equipment, storage medium and product

Abstract

The present application discloses a multi-target analysis method, apparatus, device, storage medium and product, which relates to the field of artificial intelligence technology, including: obtaining multiple groups of pictures, wherein the multiple groups of pictures contain multiple targets, determining the fractal dimension D of each target in the multiple groups of pictures to classify the multiple targets and obtain classification results, and based on the classification results, calling different preset AI models from a preset AI model set to analyze targets of different categories to obtain multi-target analysis results. The present application classifies targets in an image according to the fractal dimension, and uses different preset AI models to analyze different targets according to different target categories, thereby avoiding superimposing multiple different models to analyze different targets in the image and improving the efficiency of multi-target analysis.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to multi-target analysis methods, devices, equipment, storage media and products. Background Art

[0002] In related technologies, when analyzing multiple targets in an image, it is usually necessary to superimpose multiple different models to complete the task, which leads to low target analysis efficiency. Therefore, how to improve the efficiency of analyzing multiple targets in an image is an issue that needs to be solved urgently.

[0003] The above content is only used to assist in understanding the technical solution of this application and does not constitute an admission that the above content is related technology. Summary of the Invention

[0004] The main purpose of this application is to provide a multi-target analysis method, device, equipment, storage medium and product, aiming to solve the technical problem of how to improve the efficiency of analyzing multiple targets in an image.

[0005] To achieve the above objectives, this application proposes a multi-objective analysis method, which includes:

[0006] Acquire multiple groups of images, wherein the multiple groups of images contain multiple targets;

[0007] Determining a fractal dimension D of each target in the plurality of groups of images to classify the plurality of targets and obtain a classification result;

[0008] Based on the classification results, different preset AI models are called from the preset AI model set to analyze targets of different categories to obtain multi-target analysis results.

[0009] In one embodiment, the step of determining the fractal dimension D of each object in the plurality of groups of images further includes any one of the following:

[0010] Each target is processed using the double blanket coverage model to obtain the fractal dimension D of each target;

[0011] Based on the differential box counting algorithm, fractal calculation is performed on the multiple targets to obtain the fractal dimension D of each target.

[0012] In one embodiment, the step of processing each target using the double-blanket coverage model to obtain the fractal dimension D of each target further includes:

[0013] Based on the double-blanket coverage model, two upper and lower blankets are constructed, wherein the blankets are used to cover the image corresponding to each target;

[0014] Adjust the scales of the upper and lower blankets to obtain the surface areas of the upper and lower blankets corresponding to different sizes;

[0015] Based on the surface area, the fractal dimension is derived to obtain the fractal dimension D of each target.

[0016] In one embodiment, the step of performing fractal calculation on the multiple targets based on the differential box counting algorithm to obtain the fractal dimension D of each target further includes:

[0017] Expand the image corresponding to each target into a three-dimensional surface, wherein the x-axis and y-axis corresponding to the three-dimensional surface are spatial coordinates, and the z-axis is the image grayscale value;

[0018] Divide the x-plane and y-plane corresponding to the three-dimensional surface into The grid is based on The three-dimensional surface is covered with boxes to obtain the total number of boxes;

[0019] Adjust the size of L to get different total number of boxes based on different values of L;

[0020] Based on the least square method and the total number of different boxes, the fractal dimension D is obtained.

[0021] In one embodiment, after the step of obtaining the multi-target analysis results, the method further includes:

[0022] Performing mathematical calculation of information fusion on the multi-objective analysis results to obtain mathematical calculation results;

[0023] Obtain work task requirements, and based on the work task requirements, perform logical analysis and logical judgment on the mathematical calculation results to obtain accurate information required by the work task requirements.

[0024] In one embodiment, before the step of analyzing targets of different categories by calling different preset AI models from a preset AI model set based on the classification result to obtain a multi-target analysis result, the step further includes:

[0025] Acquire a sample target to be analyzed, where the analysis result corresponding to the sample target is a first analysis result;

[0026] Analyze the sample target using the current AI model to obtain a second analysis result;

[0027] Determining an analysis result loss value based on the first analysis result, the second analysis result, and a preset loss function;

[0028] Determine whether the loss value of the analysis result is greater than a preset analysis loss threshold;

[0029] If it is greater, adjust the parameters of the current AI model, and based on the adjusted current AI model, return to the step of using the current AI model to analyze the sample target and obtain the second analysis result, until the analysis result loss value is less than or equal to the preset analysis loss threshold, and obtain the preset AI model.

[0030] In addition, to achieve the above objectives, the present application also proposes a multi-target analysis device, which includes:

[0031] An acquisition module, configured to acquire multiple sets of images, wherein the multiple sets of images contain multiple targets;

[0032] a classification module, the classification module being configured to determine a fractal dimension D of each target in the plurality of groups of images, so as to classify the plurality of targets and obtain a classification result;

[0033] An analysis module is used to call different preset AI models from a preset AI model set to analyze targets of different categories based on the classification results to obtain multi-target analysis results.

[0034] In one embodiment, the classification module further includes:

[0035] a first computing unit, configured to process each target using a double blanket coverage model to obtain a fractal dimension D of each target;

[0036] The second calculation unit is configured to perform fractal calculation on the multiple targets based on a differential box counting algorithm to obtain a fractal dimension D of each target.

[0037] In one embodiment, the classification module further includes:

[0038] a construction unit, configured to construct two upper and lower blankets based on a double-blanket covering model, wherein the blankets are used to cover the image corresponding to each target;

[0039] The first adjustment unit is used to adjust the sizes of the upper and lower blankets to obtain the surface areas of the upper and lower blankets corresponding to different sizes;

[0040] The derivation unit is configured to derive the fractal dimension based on the surface area to obtain the fractal dimension D of each target.

[0041] In one embodiment, the classification module further includes:

[0042] An expansion unit, configured to expand the image corresponding to each target into a three-dimensional surface, wherein the x-axis and y-axis corresponding to the three-dimensional surface are spatial coordinates, and the z-axis is the image grayscale value;

[0043] A division unit is used to divide the x plane and y plane corresponding to the three-dimensional surface into The grid is based on The three-dimensional surface is covered with boxes to obtain the total number of boxes;

[0044] The second adjustment unit is used to adjust the size of L, and obtain different total numbers of boxes based on different values of L;

[0045] An obtaining unit is used to obtain a fractal dimension D based on a least square method and the total number of different boxes.

[0046] In one embodiment, the multi-target analysis device further includes a calculation module, and the calculation module includes:

[0047] A calculation unit, configured to perform mathematical calculations of information fusion on the multi-objective analysis results to obtain mathematical calculation results;

[0048] The logic processing unit is used to obtain work task requirements and, based on the work task requirements, perform logic analysis and logic judgment on the mathematical calculation results to obtain accurate information required by the work task requirements.

[0049] In one embodiment, the multi-target analysis device further includes a model training module, and the model training module includes:

[0050] an acquiring unit, configured to acquire a sample target to be analyzed, wherein the analysis result corresponding to the sample target is a first analysis result;

[0051] an analyzing unit, configured to analyze the sample target using the current AI model to obtain a second analysis result;

[0052] a determining unit, configured to determine an analysis result loss value based on the first analysis result, the second analysis result, and a preset loss function;

[0053] A judgment unit, configured to judge whether the loss value of the analysis result is greater than a preset analysis loss threshold;

[0054] The training unit is used to adjust the parameters of the current AI model if it is greater than, and based on the adjusted current AI model, return to the step of using the current AI model to analyze the sample target and obtain a second analysis result, until the analysis result loss value is less than or equal to the preset analysis loss threshold, and the preset AI model is obtained.

[0055] In addition, to achieve the above-mentioned purpose, the present application also proposes a multi-target analysis device, which includes: a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the computer program is configured to implement the steps of the multi-target analysis method as described above.

[0056] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by the processor, the steps of the multi-target analysis method described above are implemented.

[0057] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, it implements the steps of the multi-target analysis method described above.

[0058] One or more technical solutions proposed in this application have at least the following technical effects:

[0059] The present application proposes a multi-target analysis method, device, equipment, storage medium and product, which relate to the field of artificial intelligence technology. Compared with the related technology, when analyzing multiple targets in an image, it is usually necessary to superimpose multiple different models to complete the task, which leads to low target analysis efficiency. In the present application, first, multiple groups of pictures are obtained, wherein the multiple groups of pictures contain multiple targets, and then the fractal dimension D of each target in the multiple groups of pictures is determined to classify the multiple targets and obtain classification results. Finally, based on the classification results, different preset AI models are called from the preset AI model set to analyze targets of different categories to obtain multi-target analysis results.

[0060] It can be understood that this application classifies targets in the image according to the fractal dimension, and uses different preset AI models to analyze different targets according to different target categories, avoiding the superposition of multiple different models to analyze different targets in the image, and improving the efficiency of multi-target analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0062] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0063] Figure 1 A schematic diagram of the process flow provided for the first embodiment of the multi-target analysis method of this application;

[0064] Figure 2 The original image of the river surface inspection by the drone provided in Example 1 of the multi-target analysis method of this application;

[0065] Figure 3 The river surface object recognition map provided in Example 1 of the multi-target analysis method of this application;

[0066] Figure 4 This is a floating garbage identification map in a river provided in Example 1 of the multi-target analysis method of this application;

[0067] Figure 5 A schematic diagram of a simplified multi-target analysis method according to the second embodiment of the present invention;

[0068] Figure 6 This is a schematic diagram of the module structure of the multi-target analysis device according to an embodiment of the present application;

[0069] Figure 7 Schematic diagram of the device structure of the hardware operating environment involved in the multi-target analysis method in the embodiment of the present application.

[0070] The purpose, features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0071] It should be understood that the specific embodiments described herein are merely used to explain the technical solutions of the present application and are not intended to limit the present application.

[0072] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0073] The main solutions of the embodiments of this application are:

[0074] In this embodiment, for ease of description, the following description is made with the multi-target analysis device as the execution subject.

[0075] Due to existing technologies: when analyzing multiple targets in an image, it is usually necessary to superimpose multiple different models to complete the task, which leads to low target analysis efficiency.

[0076] The present application provides a solution, which includes: first, obtaining multiple groups of pictures, wherein the multiple groups of pictures contain multiple targets; then, determining the fractal dimension D of each target in the multiple groups of pictures to classify the multiple targets and obtain classification results; finally, based on the classification results, calling different preset AI models from a preset AI model set to analyze targets of different categories and obtain multi-target analysis results.

[0077] It can be understood that this application classifies targets in the image according to the fractal dimension, and uses different preset AI models to analyze different targets according to different target categories, avoiding the superposition of multiple different models to analyze different targets in the image, and improving the efficiency of multi-target analysis.

[0078] It should be noted that the execution subject of this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, mobile phone, etc., or an electronic device or multi-target analysis device capable of performing the above functions. The following uses a multi-target analysis device as an example to illustrate this embodiment and the following embodiments.

[0079] Based on this, the present invention provides a multi-target analysis method. Figure 1 , Figure 1 This is a flow chart of the first embodiment of the multi-target analysis method of this application.

[0080] In this embodiment, the multi-target analysis method includes steps S10 to S30:

[0081] Step S10, obtaining multiple groups of pictures, wherein the multiple groups of pictures contain multiple targets;

[0082] It should be noted that the execution subject of this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, mobile phone, etc., or an electronic device or multi-target analysis device capable of performing the above functions. The following uses a multi-target analysis device as an example to illustrate this embodiment and the following embodiments.

[0083] In this embodiment, specific application scenarios may be:

[0084] On a river channel, it is necessary to collect pictures from drone cameras, such as Figure 2 As shown, it is determined whether there is floating garbage in the river. Then, the multi-target analysis device receives the image captured by the drone camera, and then the multi-target analysis device analyzes the targets in the captured image, and finally determines whether each target in the river is floating garbage.

[0085] It is understandable that for the images collected by the drone camera, the multi-target analysis device will select objects according to the task (preliminary recognition range, such as Figure 3 As shown), thereby reducing the computational complexity of the multi-target analysis equipment.

[0086] Step S20, determining the fractal dimension D of each target in the plurality of groups of images to classify the plurality of targets and obtain classification results;

[0087] It should be noted that a fractal is a geometric form with non-integer dimensions. Its core characteristic is a complex structure that repeats itself at different scales through self-similarity. The most fundamental characteristic of fractal theory is that it uses the perspective and mathematical methods of fractional dimensions to describe and study objective reality. In other words, it uses the mathematical tools of fractal dimension to describe and study objective reality. It transcends the traditional limitations of one-dimensional lines, two-dimensional surfaces, three-dimensional solids, and even four-dimensional space-time, moving closer to describing the true properties and states of complex systems and more in line with the diversity and complexity of objective reality.

[0088] It should be noted that the fractal dimension is a non-integer dimension value that reflects the self-similarity and space occupancy efficiency of the fractal structure under scale transformation.

[0089] It should be noted that, in the present application, after the fractal dimension D of each target is determined, the classification result of each target is determined according to a preset fractal dimension interval.

[0090] For example, the fractal dimension is divided into five intervals: [1, 1.2), [1.2, 1.4), [1.4, 1.6), [1.6, 18), [1.8, 2.0], each interval corresponds to a classification, where the simpler the line texture of the object image, the smaller the corresponding fractal dimension.

[0091] Determine the interval corresponding to each fractal dimension D to determine the classification corresponding to each target (classify according to the line texture complexity of the image corresponding to each target). Furthermore, according to different classifications, select different AI models to process different targets, where different AI models are used to process target images with different line texture complexity.

[0092] In addition, it should be noted that in this application, the characteristics of the target are characterized by the fractal dimension D, such as Figure 4 As shown, multiple targets are classified based on the characteristics of the targets. For targets of the same category, the multi-target analysis device uses the same preset AI model for analysis, avoiding the superposition of multiple models, improving the efficiency of multi-target analysis, and ultimately obtaining multi-target analysis results.

[0093] Specifically, the step of determining the fractal dimension D of each target in the multiple groups of images further includes any one of the following:

[0094] Each target is processed using the double blanket coverage model to obtain the fractal dimension D of each target;

[0095] Based on the differential box counting algorithm, fractal calculation is performed on the multiple targets to obtain the fractal dimension D of each target.

[0096] It's important to note that the Double Blanket Covering Model (DBCM) is an important method in fractal theory, primarily used to calculate the fractal dimension of an image to describe its texture roughness. Based on fractal geometry, the DBCM treats the image's grayscale values as a curved surface in three-dimensional space (x and y are spatial coordinates, and z is the grayscale value). The fractal dimension is estimated by measuring the volume changes of a "blanket" covering this surface.

[0097] It's important to note that the differential box counting method, based on the box-counting dimension theory, estimates the fractal dimension by counting the minimum number of boxes that cover the grayscale surface of an image. The core idea is to treat a two-dimensional grayscale image as a surface in three-dimensional space (x, y are spatial coordinates, z is the grayscale value), cover the surface with grids of different scales, and analyze the relationship between the number of boxes and scale.

[0098] Specifically, the step of processing each target using the double-blanket coverage model to obtain the fractal dimension D of each target further includes steps A01 to A03:

[0099] Step A01: constructing two upper and lower blankets based on a double-blanket coverage model, wherein the blankets are used to cover the image corresponding to each target;

[0100] Step A02: Adjust the sizes of the upper and lower blankets to obtain the surface areas of the upper and lower blankets corresponding to different sizes;

[0101] Step A03: deriving the fractal dimension based on the surface area to obtain the fractal dimension D of each target.

[0102] Specifically, the steps for processing each target using the double blanket coverage model to obtain the fractal dimension D of each target are as follows:

[0103] First, a "blanket" with a thickness of 2ε is constructed on the upper and lower ε of the image grayscale surface. The surface area of the blanket is the volume of the blanket divided by 2ε. Then, let Represents the gray value function, the upper and lower surfaces are represented by 、 Indicates that .

[0104] Furthermore, the upper and lower surfaces grow according to the following principles:

[0105]

[0106]

[0107] in, for and The distance between two points.

[0108] Furthermore, the volume of the "blanket" is:

[0109]

[0110] Furthermore, the surface area of the "blanket" is:

[0111]

[0112] After obtaining the surface area and volume of the "blanket", since the fractal surface area conforms to the relationship:

[0113]

[0114] get:

[0115] +

[0116] in, is the slope of the regression line equation, is a constant corresponding to the regression line equation, It has no practical significance in this application.

[0117] Furthermore, changing the scale The size of can be calculated as a series of , and then use the least square method to solve { , } Perform linear regression on the points and find the slope of the regression line , through the relationship between the slope of the straight line and the fractal dimension: The fractal dimension D can be calculated.

[0118] It can be understood that in this embodiment, two upper and lower “blankets” are constructed to cover the image surface, and the fractal dimension D is derived by calculating the surface area changes at different scales, providing a data basis for subsequent steps.

[0119] Specifically, the step of performing fractal calculation on the multiple targets based on the differential box counting algorithm to obtain the fractal dimension D of each target further includes steps B01 to B03:

[0120] Step B01, expanding the image corresponding to each target into a three-dimensional surface, wherein the x-axis and y-axis corresponding to the three-dimensional surface are spatial coordinates, and the z-axis is the image grayscale value;

[0121] Step B02: Divide the x-plane and y-plane corresponding to the three-dimensional curved surface into grids, and based on the grids, use boxes to cover the three-dimensional curved surface to obtain the total number of boxes;

[0122] Step B03, adjusting the size of L, and obtaining different total number of boxes based on different values of L;

[0123] Step B04: Obtain the fractal dimension D based on the least square method and the total number of different boxes.

[0124] Specifically, based on the differential box counting algorithm, the fractal calculation is performed on the multiple targets to obtain the fractal dimension D of each target as follows:

[0125] First, The image is divided into The word block, ,make ,in, is the side length of the segmented image, is the side length of the image before segmentation (original image), It is a key scale factor used to quantify the observation scale when analyzing the fractal properties of an image. ,in, For images The grayscale value at the position. The X, Y plane is divided into many pieces of size A grid with a column on each grid box, is the height of a single box, and the total gray level is G. The value is .

[0126] Located in The minimum and maximum values of the image grayscale in the grid fall on the mth and In a box, then:

[0127]

[0128] in, is the box corresponding to the maximum grayscale value, is the box corresponding to the minimum grayscale value, Represents the number of box layers, then the number of boxes required to cover the entire image can be calculated for:

[0129]

[0130] Furthermore, the fractal dimension D can be calculated as:

[0131]

[0132] Select a group , calculate a series of , and then use the least square method to find the point pair In linear regression, the calculated slope of the line is the fractional dimension D.

[0133] It can be understood that in this embodiment, the grayscale image is regarded as a three-dimensional surface (x, y are spatial coordinates, and the z axis is the grayscale value). By counting the number of boxes required to cover the surface at different scales, the fractal dimension D is calculated to provide a data basis for subsequent steps.

[0134] Step S30: Based on the classification result, different preset AI models are called from the preset AI model set to analyze targets of different categories to obtain multi-target analysis results.

[0135] It should be noted that the multi-target analysis device uses preset AI models to analyze different target categories, which mainly involves two parts: target detection and semantic segmentation. Specifically, the preset AI models include the RCNN series of target detection models, led by Cascade-RCNN, and the YOLO series of target detection models. Semantic segmentation models include U-Net, SegNet, DeepLabv3+, and Swin-UNet.

[0136] Specifically, after the step of obtaining the multi-target analysis results, steps S40 to S50 are further included:

[0137] Step S40, performing mathematical calculation of information fusion on the multi-objective analysis results to obtain mathematical calculation results;

[0138] Step S50 , obtaining work task requirements, and based on the work task requirements, performing logical analysis and logical judgment on the mathematical calculation results to obtain accurate information required by the work task requirements.

[0139] It can be understood that, taking the example of determining whether there is floating garbage in the river, first, the fractal dimension D of the image corresponding to the first target in the river is calculated, and then, based on the preset fractal dimension interval, it is determined to use the first AI model to identify the first target, and the recognition result of the first target is a plastic bottle. At this time, combined with the task of "determining whether there is floating garbage in the river", the recognition result of the first target is a plastic bottle, and combined with "plastic bottle in the river", it is determined that the first target is floating garbage.

[0140] The present application proposes a multi-target analysis method, device, equipment, storage medium and product, which relate to the field of artificial intelligence technology. Compared with the related technology, when analyzing multiple targets in an image, it is usually necessary to superimpose multiple different models to complete the task, which leads to low target analysis efficiency. In the present application, first, multiple groups of pictures are obtained, wherein the multiple groups of pictures contain multiple targets, and then the fractal dimension D of each target in the multiple groups of pictures is determined to classify the multiple targets and obtain classification results. Finally, based on the classification results, different preset AI models are called from the preset AI model set to analyze targets of different categories to obtain multi-target analysis results.

[0141] It can be understood that this application classifies targets in the image according to the fractal dimension, and uses different preset AI models to analyze different targets according to different target categories, avoiding the superposition of multiple different models to analyze different targets in the image, and improving the efficiency of multi-target analysis.

[0142] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above embodiment 1 can be referred to the above introduction and will not be described in detail later. Figure 5 Before step S30, the multi-target analysis method further includes steps S21 to S25:

[0143] Step S21, obtaining a sample target to be analyzed, where the analysis result corresponding to the sample target is the first analysis result;

[0144] It should be noted that the sample target refers to the object or data that needs to be analyzed. It can be any form of data, such as images, text, audio, video, etc. In this application, the sample target refers to the sample target image.

[0145] Step S22: Analyze the sample target using the current AI model to obtain a second analysis result;

[0146] It can be understood that the second analysis result is the result obtained after the current AI model analyzes the sample target. It is the result of model analysis and may differ from the actual result (the first analysis result).

[0147] Step S23, determining an analysis result loss value based on the first analysis result, the second analysis result, and a preset loss function;

[0148] It should be noted that a loss function is a mathematical function used to measure the difference or error between a predicted result and the actual result. In machine learning, loss functions are often used to evaluate model performance and guide the model training process. A preset loss function is a loss function defined before analysis and used to compare and evaluate analysis results. In this embodiment, the difference or error between two analysis results is represented by the analysis result loss value.

[0149] It is understandable that the results of the model analysis may differ from the actual results (the first analysis results). In this case, the model needs to be trained based on the results of the model analysis and the actual results until the model achieves the expected effect.

[0150] Step S24, determining whether the loss value of the analysis result is greater than a preset analysis loss threshold;

[0151] Step S25: If it is greater than, adjust the parameters of the current AI model, and based on the adjusted current AI model, return to the step of using the current AI model to analyze the sample target and obtain the second analysis result, until the analysis result loss value is less than or equal to the preset analysis loss threshold, and obtain the preset AI model.

[0152] It is understood that in this embodiment, through continuous adjustment and testing, the difference between the analysis results of the trained AI model and the actual analysis results is kept within a certain error range, thereby obtaining a preset AI model that meets the requirements. This process is crucial for improving the stability and reliability of the model, especially in application scenarios that require high consistency of model output.

[0153] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the multi-target analysis method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0154] This application also provides a multi-target analysis device, please refer to Figure 6 , the multi-target analysis device comprises:

[0155] An acquisition module 10 is configured to acquire multiple sets of images, wherein the multiple sets of images contain multiple targets;

[0156] a classification module 20, configured to determine a fractal dimension D of each target in the plurality of groups of images, so as to classify the targets and obtain a classification result;

[0157] The analysis module 30 is used to call different preset AI models from the preset AI model set based on the classification result to analyze targets of different categories and obtain multi-target analysis results.

[0158] In one embodiment, the classification module further includes:

[0159] a first computing unit, configured to process each target using a double blanket coverage model to obtain a fractal dimension D of each target;

[0160] The second calculation unit is configured to perform fractal calculation on the multiple targets based on a differential box counting algorithm to obtain a fractal dimension D of each target.

[0161] In one embodiment, the classification module further includes:

[0162] a construction unit, configured to construct two upper and lower blankets based on a double-blanket covering model, wherein the blankets are used to cover the image corresponding to each target;

[0163] The first adjustment unit is used to adjust the sizes of the upper and lower blankets to obtain the surface areas of the upper and lower blankets corresponding to different sizes;

[0164] The derivation unit is configured to derive the fractal dimension based on the surface area to obtain the fractal dimension D of each target.

[0165] In one embodiment, the classification module further includes:

[0166] An expansion unit, configured to expand the image corresponding to each target into a three-dimensional surface, wherein the x-axis and y-axis corresponding to the three-dimensional surface are spatial coordinates, and the z-axis is the image grayscale value;

[0167] A division unit is used to divide the x plane and y plane corresponding to the three-dimensional surface into The grid is based on The three-dimensional surface is covered with boxes to obtain the total number of boxes;

[0168] The second adjustment unit is used to adjust the size of L, and obtain different total numbers of boxes based on different values of L;

[0169] An obtaining unit is used to obtain a fractal dimension D based on a least square method and the total number of different boxes.

[0170] In one embodiment, the multi-target analysis device further includes a calculation module, and the calculation module includes:

[0171] A calculation unit, configured to perform mathematical calculations of information fusion on the multi-objective analysis results to obtain mathematical calculation results;

[0172] The logic processing unit is used to obtain work task requirements and, based on the work task requirements, perform logic analysis and logic judgment on the mathematical calculation results to obtain accurate information required by the work task requirements.

[0173] In one embodiment, the multi-target analysis device further includes a model training module, and the model training module includes:

[0174] an acquiring unit, configured to acquire a sample target to be analyzed, wherein the analysis result corresponding to the sample target is a first analysis result;

[0175] an analyzing unit, configured to analyze the sample target using the current AI model to obtain a second analysis result;

[0176] a determining unit, configured to determine an analysis result loss value based on the first analysis result, the second analysis result, and a preset loss function;

[0177] A judgment unit, configured to judge whether the loss value of the analysis result is greater than a preset analysis loss threshold;

[0178] The training unit is used to adjust the parameters of the current AI model if it is greater than, and based on the adjusted current AI model, return to the step of using the current AI model to analyze the sample target and obtain a second analysis result, until the analysis result loss value is less than or equal to the preset analysis loss threshold, and the preset AI model is obtained.

[0179] The multi-target analysis device provided in this application utilizes the multi-target analysis method described in the aforementioned embodiments to address the technical challenges of multi-target analysis. Compared to the prior art, the multi-target analysis device provided in this application achieves the same beneficial effects as the multi-target analysis method described in the aforementioned embodiments. Other technical features of the multi-target analysis device are the same as those disclosed in the aforementioned embodiments and are not further elaborated here.

[0180] The present application provides a multi-target analysis device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the multi-target analysis method in the above-mentioned embodiment one.

[0181] Reference below Figure 7, which shows a schematic diagram of the structure of a multi-target analysis device suitable for implementing the embodiments of the present application. The multi-target analysis device in the embodiments of the present application may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 7 The multi-target analysis device shown is merely an example and should not limit the functions and scope of use of the embodiments of the present application.

[0182] like Figure 7 As shown, the multi-target analysis device may include a processing device 1001 (e.g., a central processing unit, graphics processing unit, etc.), which can perform various appropriate actions and processes based on programs stored in a read-only memory (ROM) 1002 or programs loaded from a storage device 1003 into a random access memory (RAM) 1004. RAM 1004 also stores various programs and data required for the operation of the multi-target analysis device. Processing device 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007, such as a touchscreen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 1008, such as a liquid crystal display (LCD), speaker, vibrator, etc.; storage device 1003, such as a magnetic tape or hard disk; and communication devices 1009. Communication device 1009 can allow the multi-target analysis device to communicate with other devices wirelessly or wired to exchange data. Although the figure shows a multi-target analysis device with various systems, it should be understood that it is not required to implement or have all of the systems shown. More or fewer systems can be implemented or provided instead.

[0183] In particular, according to the embodiments disclosed in the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program comprising program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0184] The multi-target analysis device provided in this application utilizes the multi-target analysis method described in the above-mentioned embodiments to solve the technical problems. Compared with the prior art, the beneficial effects of the multi-target analysis device provided in this application are the same as those of the multi-target analysis method described in the above-mentioned embodiments. Other technical features of the multi-target analysis device are the same as those disclosed in the above-mentioned embodiments and are not further elaborated here.

[0185] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0186] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0187] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, a computer program) stored thereon, wherein the computer-readable program instructions are used to execute the multi-target analysis method in the above-mentioned embodiment.

[0188] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0189] The computer-readable storage medium may be included in the multi-target analysis device, or may exist independently without being incorporated into the multi-target analysis device.

[0190] The computer-readable storage medium carries one or more programs. When the one or more programs are executed by the multi-target analysis device, the multi-target analysis device:

[0191] Acquire multiple groups of images, wherein the multiple groups of images contain multiple targets;

[0192] Determining a fractal dimension D of each target in the plurality of groups of images to classify the plurality of targets and obtain a classification result;

[0193] Based on the classification results, different preset AI models are called from the preset AI model set to analyze targets of different categories to obtain multi-target analysis results.

[0194] Computer program code for performing the operations of the present application may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0195] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, program segment or a part of code, and the module, program segment or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented by a dedicated hardware-based system that performs the specified function or operation, or can be implemented by a combination of dedicated hardware and computer instructions.

[0196] The modules described in the embodiments of the present application may be implemented in software or hardware, wherein the name of a module does not necessarily limit the unit itself.

[0197] The computer-readable storage medium provided in this application stores computer-readable program instructions (i.e., a computer program) for executing the multi-target analysis method described above, thereby solving the technical problems of multi-target analysis. Compared to the prior art, the beneficial effects of the computer-readable storage medium provided in this application are similar to those of the multi-target analysis method provided in the aforementioned embodiments and are not further elaborated here.

[0198] The present application also provides a computer program product, comprising a computer program, which implements the steps of the multi-target analysis method as described above when executed by a processor.

[0199] The computer program product provided in this application can solve the technical problem of multi-target analysis. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the multi-target analysis method provided in the above embodiment, and will not be repeated here.

[0200] The above description is only part of the embodiments of the present application and does not limit the patent scope of the present application. All equivalent structural transformations made by using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect application in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A multi-objective analysis method, characterized in that: The multi-target analysis method comprises: Acquire multiple groups of images, wherein the multiple groups of images contain multiple targets; Determining a fractal dimension D of each target in the plurality of groups of images, and classifying the plurality of targets based on a preset fractal dimension interval and the fractal dimension D to obtain a classification result; Based on the classification results, different preset AI models are called from the preset AI model set to analyze targets of different categories to obtain multi-target analysis results.

2. The multi-target analysis method according to claim 1, wherein: The step of determining the fractal dimension D of each target in the plurality of groups of images further includes any one of the following: Each target is processed using the double blanket coverage model to obtain the fractal dimension D of each target; Based on the differential box counting algorithm, fractal calculation is performed on the multiple targets to obtain the fractal dimension D of each target.

3. The multi-target analysis method according to claim 2, wherein: The step of processing each target using the double blanket coverage model to obtain the fractal dimension D of each target further includes: Based on the double-blanket coverage model, two upper and lower blankets are constructed, wherein the blankets are used to cover the image corresponding to each target; Adjust the scales of the upper and lower blankets to obtain the surface areas of the upper and lower blankets corresponding to different sizes; Based on the surface area, the fractal dimension is derived to obtain the fractal dimension D of each target.

4. The multi-target analysis method according to claim 2, wherein: The step of performing fractal calculation on the multiple targets based on the differential box counting algorithm to obtain the fractal dimension D of each target further includes: Expand the image corresponding to each target into a three-dimensional surface, wherein the x-axis and y-axis corresponding to the three-dimensional surface are spatial coordinates, and the z-axis is the image grayscale value; Divide the x-plane and y-plane corresponding to the three-dimensional surface into The grid is based on The three-dimensional surface is covered with boxes to obtain the total number of boxes; Adjust the size of L to get different total number of boxes based on different values of L; Based on the least square method and the total number of different boxes, the fractal dimension D is obtained.

5. The multi-target analysis method according to claim 1, wherein: After the step of obtaining the multi-target analysis results, the method further includes: Performing mathematical calculation of information fusion on the multi-objective analysis results to obtain mathematical calculation results; Obtain work task requirements, and based on the work task requirements, perform logical analysis and logical judgment on the mathematical calculation results to obtain accurate information required by the work task requirements.

6. The multi-target analysis method according to claim 1, wherein: Before the step of analyzing targets of different categories by calling different preset AI models from a preset AI model set based on the classification result to obtain a multi-target analysis result, the method further includes: Acquire a sample target to be analyzed, where the analysis result corresponding to the sample target is a first analysis result; Analyze the sample target using the current AI model to obtain a second analysis result; Determining an analysis result loss value based on the first analysis result, the second analysis result, and a preset loss function; Determine whether the loss value of the analysis result is greater than a preset analysis loss threshold; If it is greater, adjust the parameters of the current AI model, and based on the adjusted current AI model, return to the step of using the current AI model to analyze the sample target and obtain the second analysis result, until the analysis result loss value is less than or equal to the preset analysis loss threshold, and obtain the preset AI model.

7. A multi-target analysis device, characterized in that: The multi-target analysis device comprises: An acquisition module, configured to acquire multiple sets of images, wherein the multiple sets of images contain multiple targets; a classification module, the classification module being configured to determine a fractal dimension D of each target in the plurality of groups of images, and classify the plurality of targets based on a preset fractal dimension interval and the fractal dimension D to obtain a classification result; An analysis module is used to call different preset AI models from a preset AI model set to analyze targets of different categories based on the classification results to obtain multi-target analysis results.

8. A multi-target analysis device, characterized in that: The device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the multi-target analysis method according to any one of claims 1 to 6.

9. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the multi-target analysis method according to any one of claims 1 to 6 are implemented.

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 steps of the multi-target analysis method according to any one of claims 1 to 6 are implemented.

Images