Fuzzy control-based video enhanced reasoning method, device and equipment and medium

Abstract

The invention relates to a video enhanced reasoning method and device based on fuzzy control, equipment and a medium. The method comprises the following steps: initializing a plurality of pre-training video reasoning models of different calculation scales; defining a fuzzy semantic set and a membership function of the input variable and the output variable; constructing a fuzzy rule table based on prior knowledge; collecting input data in real time, completing fuzzy reasoning, and defuzzifying through a gravity center method to output an accurate score; and dynamically selecting the optimal video reasoning model of the next frame according to the score value range and the anti-shake strategy. According to the method, the target continuity characteristic between video frames is fully utilized, dynamic balance between hardware resources and reasoning precision is achieved, the GPU utilization rate is obviously increased, the target detection rate is also obviously increased, the temperature is stabilized within a safety threshold value, and the method is suitable for real-time video scenes such as traffic monitoring and unmanned aerial vehicle inspection.

Description

Technical Field

[0001] The present invention relates to the field of video enhancement reasoning technology, and in particular to a video enhancement reasoning method, device, equipment and medium based on fuzzy control. Background Art

[0002] Current video inference technologies (such as object detection models YOLO and DETR) mainly rely on a single fixed model to process continuous video streams, which has significant flaws:

[0003] Imbalance in resource utilization and performance: Large-scale models, while highly accurate, require significant computing power, which can easily lead to GPU overload and temperature increase. Small-scale models, while less resource-intensive, struggle to cope with complex scenarios (such as densely packed objects), resulting in high rates of missed detections. In practical applications, hardware resource utilization and inference performance are difficult to optimize in a coordinated manner.

[0004] Ignoring inter-frame correlation: The object states (number, position) between adjacent video frames typically exhibit spatiotemporal continuity, resulting in a very low probability of sudden changes. However, existing methods perform independent inference on each frame, neither leveraging this property to reduce computational redundancy nor dynamically adjusting the model based on scene changes, resulting in wasted resources and poor inference efficiency.

[0005] For example, in traffic monitoring scenarios, where the number of vehicles fluctuates slowly, using a fixed large model will consistently occupy over 90% of GPU utilization, while using a fixed small model will increase the missed detection rate by 15% during peak hours. Although some studies have attempted multi-model switching, they lack a joint perception mechanism for hardware status (such as GPU temperature and utilization) and scene complexity (such as the number of targets). Furthermore, the switching strategy relies on hard thresholds, which can easily lead to frequent and invalid switching due to data jitter, thereby reducing system stability. Therefore, there is an urgent need for an intelligent video inference solution that can balance resource efficiency and inference accuracy in real time and fully utilize frame continuity. Summary of the Invention

[0006] The technical problem to be solved by the present invention is: the problem of mismatch between resource utilization and effect of video inference model.

[0007] In order to solve the above technical problems, the technical solution adopted by the present invention is: a video enhancement reasoning method based on fuzzy control, comprising the following steps:

[0008] S10, loading multiple pre-trained video inference models of different computational complexity;

[0009] S20, defining input parameters reflecting hardware resource status and scenario complexity, and output scoring parameters representing model selection tendency, and defining fuzzy sets for each parameter;

[0010] S30, configuring a membership function for each input parameter, and mapping the input parameter values ​​collected in real time to the corresponding fuzzy set;

[0011] S40, establishing a mapping rule between an input parameter fuzzy set and an output score fuzzy set based on an expert experience knowledge base of the video reasoning scenario;

[0012] S50, collecting input parameter values ​​of the video frame, mapping them to fuzzy sets through membership functions, and generating fuzzy subsets of output scores in combination with the mapping rules;

[0013] S60, performing a centroid calculation on the fuzzy subset of the output score to obtain an accurate model selection score;

[0014] S70: Determine the optimal video inference model for the next frame based on the model selection score of the current frame and the preset anti-switching jitter strategy.

[0015] Furthermore, in step S10, the plurality of pre-trained video inference models of different computational complexity include a small-scale model, a medium-scale model, and a large-scale model, and the medium-scale model is used as the initial frame inference model by default.

[0016] Furthermore, in step S20, the input parameters include:

[0017] GPU utilization, whose fuzzy set is defined as {low utilization, medium utilization, high utilization};

[0018] GPU temperature, whose fuzzy set is defined as {temperature low, temperature medium, temperature high};

[0019] The number of targets in the current frame, whose fuzzy set is defined as {small number, medium number, large number};

[0020] The output scoring parameter is the model recommendation score, and its fuzzy semantic label is {low score, medium score, high score}.

[0021] Furthermore, in step S30, the membership function is a piecewise linear function, which is used to map the numerical value of the input variable to a membership in the interval [0, 1].

[0022] Furthermore, in step S40: the mapping rules are stored in the form of a fuzzy rule table, and the fuzzy rule table is established according to the video reasoning experience knowledge base and measurement data.

[0023] Furthermore, in step S70, the mapping relationship between the model selection score and the inference model is:

[0024] When the score is in the interval [0,30], the large-scale model is selected;

[0025] When the score is in the range of (30,50], the medium-sized model is selected;

[0026] When the score is in the range of (50,100], a small-scale model is selected.

[0027] Furthermore, in step S70, the anti-switching jitter strategy is: the model switching operation is performed only when the model selection results of n consecutive frames are consistent and different from the current model.

[0028] The present invention also provides a video enhancement reasoning device based on fuzzy control, comprising:

[0029] Model loading module, used to load multiple pre-trained video inference models of different computational complexity;

[0030] The fuzzy variable definition module is used to define input parameters that reflect the hardware resource status and scenario complexity, as well as output scoring parameters that represent the model selection tendency, and define fuzzy sets for each parameter;

[0031] The membership configuration module is used to configure the membership function for each input parameter and map the input parameter values ​​collected in real time to the corresponding fuzzy set;

[0032] The rule base module is used to establish the mapping rules between the input parameter fuzzy set and the output score fuzzy set based on the expert experience knowledge base of the video reasoning scenario;

[0033] A fuzzy control module is used to collect input parameter values ​​of video frames, map them to fuzzy sets through membership functions, and generate fuzzy subsets of output scores based on the mapping rules;

[0034] The score conversion module is used to perform centroid calculation on the fuzzy subset of the output score to obtain an accurate model selection score;

[0035] The model decision module is used to determine the optimal inference model for the next frame based on the model selection score of the current frame and the preset anti-switching jitter strategy.

[0036] The present invention also provides a computer device, which includes a memory and a processor. The memory stores a computer program, and the processor implements the above-mentioned video enhancement reasoning method based on fuzzy control when executing the computer program.

[0037] The present invention also provides a storage medium storing a computer program. When the computer program is executed by a processor, the video enhancement reasoning method based on fuzzy control as described above can be implemented.

[0038] The beneficial effects of the present invention are as follows: the present invention significantly solves the core problems of imbalance between resource utilization and effect and lack of inter-frame correlation in video reasoning by initializing multi-scale models, defining fuzzy variables and mapping rules, constructing fuzzy control rules, real-time dynamic reasoning and integrating anti-shake strategies: on the one hand, based on fuzzy perception of hardware status (GPU utilization, temperature) and scene complexity (number of targets), the optimal scale model is dynamically switched, so that it is automatically downgraded to a small model to prevent overload when the load is high, and upgraded to a large model to improve accuracy when the load is low, thereby achieving coordinated optimization of resource utilization and reasoning effect; on the other hand, the characteristic of continuous change of targets between video frames is utilized, and frequent switching is suppressed through anti-shake strategies, so as to avoid excessively frequent switching of models, ensure the stability of the reasoning process, and improve the target detection rate compared with a single model in scenarios such as drone inspection and traffic monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The specific structure of the present invention is described in detail below with reference to the accompanying drawings.

[0040] Figure 1 A flowchart of a video enhancement reasoning method based on fuzzy control according to an embodiment of the present invention;

[0041] Figure 2 This is a diagram of the fuzzy control architecture of an embodiment of the present invention;

[0042] Figure 3 A membership function diagram of input and output quantities according to an embodiment of the present invention;

[0043] Figure 4 This is a schematic diagram of the center of gravity method according to an embodiment of the present invention;

[0044] Figure 5 This is a comparison diagram of the reasoning effects of an embodiment of the present invention;

[0045] Figure 6 A comparison chart of inference chip utilization in an embodiment of the present invention

[0046] Figure 7 This is a temperature comparison diagram of the inference device according to an embodiment of the present invention;

[0047] Figure 8 This is a block diagram of a video enhancement inference device based on fuzzy control according to an embodiment of the present invention;

[0048] Figure 9 A schematic block diagram of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0049] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0050] It will be understood that when used in this specification and the appended claims, the terms “comprises” and “comprising” indicate the presence of described features, integers, steps, operations, elements and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof.

[0051] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the present invention. As used in the specification and appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms unless the context clearly indicates otherwise.

[0052] It should be further understood that the term "and / or" used in the present description and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0053] The first embodiment of the present invention is: a video enhancement reasoning method based on fuzzy control, comprising the following steps:

[0054] S10. Load multiple pre-trained video inference models of different computational complexity.

[0055] In a specific embodiment, in step S10, the plurality of pre-trained video inference models of different computational complexity include a small-scale model, a medium-scale model, and a large-scale model, and the medium-scale model is used as the initial frame inference model by default.

[0056] In this embodiment, several trained video inference models with different parameter sizes are selected. Taking three as an example, they are recorded as a small-scale model (model_n), a medium-scale model (model_m), and a large-scale model (model_l). The medium-scale model model_m is selected as the initial frame inference model and loaded into the memory.

[0057] S20. Define input parameters that reflect hardware resource status and scenario complexity, as well as output scoring parameters that represent model selection tendencies, and define fuzzy sets for each parameter.

[0058] In a specific embodiment, in step S20, the input parameters include:

[0059] GPU utilization, whose fuzzy set is defined as {low utilization, medium utilization, high utilization};

[0060] GPU temperature, whose fuzzy set is defined as {temperature low, temperature medium, temperature high};

[0061] The number of targets in the current frame, whose fuzzy set is defined as {small number, medium number, large number};

[0062] The output scoring parameter is the model recommendation score, and its fuzzy semantic label is {low score, medium score, high score}.

[0063] In this embodiment, the input parameters refer to:

[0064] GPU utilization (GPU_use), GPU temperature (GPU_temp), and the number of objects inferred in the current video frame (Obj_num).

[0065] Their ranges are GPU_use∈[0,100],GPU_temp∈[t1,t2],Obj_num∈[0,n1];

[0066] Where t1 and t2 are the minimum and maximum values ​​of GPU_temp, respectively. Generally speaking, they are the lowest and highest operating temperatures of the GPU. n1 is the maximum number of possible targets in the current video stream.

[0067] The fuzzy distribution is:

[0068] GPU_use: {Low Utilization, Medium Utilization, High Utilization} = {UL, UM, UH};

[0069] GPU_temp: {temperature low, temperature medium, temperature high} = {TL, TM, TH};

[0070] Obj_num:{small quantity, medium quantity, large quantity} = {NL, NM, NH};

[0071] Output refers to:

[0072] The score of the current frame is Score, which ranges from Score∈[0,100];

[0073] The fuzzy distribution is:

[0074] Score: {low score, medium score, high score} = {SL, SM, SH}.

[0075] Figure 2This is the core architecture diagram of fuzzy control. In this embodiment, the input of the fuzzy controller is the GPU temperature, GPU utilization, and the number of inference targets in the current frame; the output is the score of the current frame inference. The knowledge base refers to the expert knowledge or experience during video inference. For details, see the fuzzy rule table in Table 1. The number of combinations conforms to the following rules:

[0076]

[0077] Among them, m i is the fuzzy set number of the i-th input quantity.

[0078] S30: configuring a membership function for each input parameter, and mapping the input parameter values ​​collected in real time to the corresponding fuzzy set.

[0079] In a specific embodiment, in step S30, the membership function is a piecewise linear function, which is used to map the numerical value of the input variable to a membership in the interval [0, 1].

[0080] In this embodiment, a membership function is defined based on input parameters and output quantities. The input parameters are fuzzy mapped, and the specific values ​​are mapped into fuzzy subsets. The membership function refers to a function that maps the specific input quantity to a fuzzy set. The membership functions of the input quantities GPU_use, GPU_temp, and Obj_num are denoted as μ1, μ2, and μ3, respectively. Their expressions are as follows:

[0081]

[0082] Among them, x is the input quantity, max(x) is the maximum value of the input quantity, that is, the right limit of the input quantity, min(x) means taking the minimum value of x, that is, the left limit of the input quantity, a, b, c, d are all constants, satisfying a <b<c<d。

[0083] Figure 3 This is a membership function for three input quantities, responsible for mapping the input deterministic quantities into fuzzy sets. The horizontal axis represents the specific numerical value of the input or output quantity, while the vertical axis represents the relative size of the corresponding fuzzy subset. For example, when GPU_use is 35% utilized, its model set is {UL,UM,UH} = {0.3,0.3,0}, indicating that its degree of belonging to UL is 0.3, UM is 0.3, and UH is 0.

[0084] S40. Based on the expert experience knowledge base of the video reasoning scenario, a mapping rule between the input parameter fuzzy set and the output score fuzzy set is established.

[0085] In a specific embodiment, in step S40: the mapping rules are stored in the form of a fuzzy rule table, and the fuzzy rule table is established according to a video reasoning experience knowledge base and measurement data.

[0086] In this embodiment, establishing video reasoning fuzzy control rules based on the video reasoning experience knowledge base and measurement data refers to human experience-based control based on the summary of the corresponding task rules. For example, it is generally believed that if the GPU utilization rate is relatively low, it means that there is a large redundancy in computing power, and a larger-scale model should be switched to fully utilize the computing power, otherwise a small-scale model should be given priority; if the GPU temperature is relatively low, it means that there is still redundancy in performance, and a larger-scale model should be switched to fully squeeze the performance, otherwise a small-scale model should be considered; if the number of targets to be inspected in the current frame is large, it is considered that the detection pressure is high, and a large-scale model should be switched to improve the detection effect and detect as many targets as possible, otherwise a small-scale model should be given priority.

[0087] In fact, it is specifically described by a rule table, as shown in Rule Table 1:

[0088] Table 1: Fuzzy rules table

[0089]

[0090]

[0091] S50 , collecting input parameter values ​​of the video frame, mapping them to a fuzzy set through a membership function, and generating a fuzzy subset of the output score in combination with the mapping rule.

[0092] In this embodiment, the membership function is defined based on the input and output quantities, and the input quantity is fuzzy mapped, and the specific value is mapped into the fuzzy subset. The fuzzy mapping refers to the operation of mapping the input quantity x to the fuzzy set based on the membership function.

[0093] S60. Perform a centroid calculation on the fuzzy subset of the output score to obtain an accurate model selection score.

[0094] In this embodiment, fuzzy reasoning is performed and the centroid method is used to obtain the output score of the fuzzy controller based on the membership function of the output quantity. The membership function of the output quantity refers to a function that clarifies the output fuzzy quantity (fuzzy subset) into a specific numerical value. Its expression is as follows:

[0095]

[0096] Among them, x is the specific value of the output, max(x) is the maximum value of the output, that is, the right limit of the output, min(x) means taking the minimum value of x, that is, the left limit of the output, a, b, c, d are all constants, satisfying a <b<c<d。

[0097] According to the membership function of the output quantity, the centroid method is used to obtain the output score of the fuzzy controller, which refers to the fuzzy subset {s1, s2, s3} of Score, where s.

[0098]

[0099] Among them, X * Indicates the exact value of the output Score after clarity, μ S (x) is the membership function of the output fuzzy set, x is the value of the output quantity, and the value range of the X output quantity Score is [0-100].

[0100] In order to more clearly explain the reasoning process of the fuzzy controller, an example is given to illustrate the entire calculation process of its fuzzy reasoning:

[0101] Assuming that the specific values ​​of the current measured input are GPU_use=30, GPU_temp=36, and Obj_num=15, according to their respective membership functions, their fuzzy subsets can be calculated as follows:

[0102] {UL, UM, UH} = {0.4, 0, 0}

[0103] {TL, TM, TH} = {0.1, 0.4, 0}

[0104] {NL, NM, NH} = {0, 1, 0}

[0105] According to the fuzzy subset values ​​and the fuzzy rule table, the second rule (the second row of the rule table) and the fifth rule (the fifth row of the rule table) are activated, namely UL, TL, NM, SH and UL, TM, NM, SM. The activation strengths are calculated as SH = min{0.4, 0.1, 1} = 0.1 and SM = min{0.4, 0.4, 1} = 0.4. According to the membership function of the output quantity and the centroid calculation formula, the score Score ≈ 54 can be obtained. The centroid calculation diagram is shown below. Figure 4 shown.

[0106] S70: Determine the optimal video inference model for the next frame based on the model selection score of the current frame and the preset anti-switching jitter strategy.

[0107] In a specific embodiment, in step S70, the mapping relationship between the model selection score and the inference model is:

[0108] When the score is in the interval [0,30], the large-scale model is selected;

[0109] When the score is in the range of (30,50], the medium-sized model is selected;

[0110] When the score is in the range of (50,100], a small-scale model is selected.

[0111] In this embodiment, the best inference model for the next frame is determined based on the output score of the current frame and whether the switching criteria are met. The best model is calculated when the score is between 0 and 30, which is a large-scale model; 31-50 is a medium-scale model; and 51-100 is a small-scale model. That is:

[0112] if 0 <Score≤30,then model_n

[0113] if 30 <Score≤50,then model_m

[0114] if 51 <Score≤100,then model_l

[0115] In a specific embodiment, in step S70, the anti-switching jitter strategy is: the model switching operation is performed only when the model selection results of n consecutive frames are consistent and different from the current model.

[0116] In this embodiment, the best inference model for the next frame is determined based on the output score of the current frame and whether the switching criteria are met. The switching criteria refers to the anti-shake strategy. Specifically, it avoids model calculation score jitter that causes incorrect model switching and excessive and frequent invalid switching. Model switching is only performed when the best models calculated for n consecutive frames are the same and different from the previous model. Figure 5 This is a comparison of the number of targets inferred in the same frame of the same video by the method of the present invention and the general video inference method. In this example, in order to be fair to the YOLOv8n model used in the general solution, the method of the present invention uses three models of different scales, namely YOLOv8n, YOLOv8s, and YOLOv8 l models. They all belong to the YOLOv8 series, but with different sizes. The dataset used is the benchmark dataset Visdrone2019, the video stream is the video stream in Visdrone2019, the implementation platform is a personal computer equipped with an AMD9950x CPU and an NVIDIA 4070Tis GPU, and the inference method is GPU inference. It can be seen that the method used in the present invention is far ahead of the general solution in terms of the number of targets detected in the same frame, which fully demonstrates the superiority of the method of the present invention.

[0117] Figure 6 This is a dynamic curve of GPU utilization for each frame of inference during video inference. The experimental conditions are as follows: Figure 5 Consistent, from Figure 6 It can be seen that the GPU utilization rate of the method proposed in the present invention is much higher than that of the general method, which fully proves that the method proposed in the present invention can effectively improve the efficiency of hardware utilization.

[0118] Figure 7 This is the GPU temperature dynamic curve of each frame inference during the video inference process. The experimental conditions are Figure 5 Consistent. Figure 7 It can be seen that although the method of the present invention will slightly increase the GPU temperature, due to its real-time dynamic switching characteristics, the temperature is quickly and effectively controlled (never exceeding 42 degrees). This strategy is especially effective on some platforms with low computing power.

[0119] In summary, the enhanced video reasoning method based on fuzzy control proposed in the present invention can implement dynamic switching model reasoning according to the hardware resource conditions and reasoning conditions, which can effectively improve the video reasoning effect and effectively improve the hardware resource utilization.

[0120] like Figure 8 As shown, an embodiment of the present invention further provides a video enhancement reasoning device based on fuzzy control, comprising:

[0121] A model loading module 10 is used to load multiple pre-trained video inference models of different computational complexity;

[0122] The fuzzy variable definition modulo 20 is used to define input parameters that reflect the hardware resource status and scenario complexity, as well as output scoring parameters that represent the model selection tendency, and define fuzzy sets for each parameter;

[0123] The membership configuration module 30 is used to configure a membership function for each input parameter and map the input parameter values ​​collected in real time to the corresponding fuzzy set;

[0124] A rule base module 40 is used to establish a mapping rule between an input parameter fuzzy set and an output score fuzzy set based on an expert experience knowledge base of a video reasoning scenario;

[0125] A fuzzy control module 50 is used to collect input parameter values ​​of video frames, map them to fuzzy sets through membership functions, and generate fuzzy subsets of output scores in combination with the mapping rules;

[0126] The score conversion module 60 is used to perform a centroid calculation on the fuzzy subset of the output score to obtain an accurate model selection score;

[0127] The model decision module 70 is used to determine the optimal inference model for the next frame based on the model selection score of the current frame and the preset anti-switching jitter strategy.

[0128] It should be noted that those skilled in the art can clearly understand that the specific implementation process of the above-mentioned video enhancement reasoning device based on fuzzy control can refer to the corresponding description in the aforementioned method embodiment. For the convenience and brevity of the description, it will not be repeated here.

[0129] The above-mentioned video enhancement reasoning device based on fuzzy control can be realized in the form of a computer program. The computer program can be used in Figure 9 Runs on the computer equipment shown.

[0130] See also Figure 9 , Figure 9 This is a schematic block diagram of a computer device provided in an embodiment of the present application. The computer device 500 can be a terminal or a server. The terminal can be a smart phone, tablet computer, laptop computer, desktop computer, personal digital assistant, wearable device, or other electronic device with communication capabilities. The server can be a standalone server or a server cluster consisting of multiple servers.

[0131] See Figure 9 The computer device 500 includes a processor 502 , a memory, and a network interface 505 connected via a system bus 501 , wherein the memory may include a non-volatile storage medium 503 and an internal memory 504 .

[0132] The non-volatile storage medium 503 can store an operating system 5031 and a computer program 5032. The computer program 5032 includes program instructions, which, when executed, can enable the processor 502 to execute a video enhancement reasoning method based on fuzzy control.

[0133] The processor 502 is used to provide computing and control capabilities to support the operation of the entire computer device 500.

[0134] The internal memory 504 provides an environment for the operation of the computer program 5032 in the non-volatile storage medium 503. When the computer program 5032 is executed by the processor 502, the processor 502 can execute a video enhancement reasoning method based on fuzzy control.

[0135] The network interface 505 is used to communicate with other devices through the network. Figure 9 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present application, and does not constitute a limitation on the computer device 500 to which the solution of the present application is applied. The specific computer device 500 may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0136] The processor 502 is configured to run a computer program 5032 stored in the memory to implement the above-mentioned video enhancement reasoning method based on fuzzy control.

[0137] It should be understood that in the embodiment of the present application, the processor 502 may be a central processing unit (CPU), and the processor 502 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0138] Those skilled in the art will appreciate that all or part of the steps in the method of the above-described embodiment can be implemented by instructing the relevant hardware through a computer program. The computer program includes program instructions, which can be stored in a storage medium that is computer-readable. The program instructions are executed by at least one processor in the computer system to implement the steps in the method of the above-described embodiment.

[0139] Therefore, the present invention further provides a storage medium. The storage medium may be a computer-readable storage medium. The storage medium stores a computer program, wherein the computer program includes program instructions. When executed by a processor, the program instructions cause the processor to perform the above-described video enhancement inference method based on fuzzy control.

[0140] The storage medium may be any computer-readable storage medium that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a magnetic disk, or an optical disk.

[0141] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the above description has generally described the composition and steps of each example according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0142] In the several embodiments provided herein, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the various units is merely a logical functional division, and actual implementation may employ other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be omitted or not implemented.

[0143] The steps in the methods of the embodiments of the present invention may be adjusted in order, combined, or deleted as needed. The units in the devices of the embodiments of the present invention may be combined, divided, or deleted as needed. Furthermore, the functional units in the various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit.

[0144] If this integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the existing technology, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes a number of instructions for causing a computer device (which can be a personal computer, terminal, or network device, etc.) to execute all or part of the steps of the method described in various embodiments of the present invention.

[0145] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and such modifications or substitutions are intended to be within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be subject to the scope of protection of the claims.

Claims

1. A video enhancement reasoning method based on fuzzy control, characterized in that: The following steps are involved: S10, loading multiple pre-trained video inference models of different computational complexity; S20, defining input parameters reflecting hardware resource status and scenario complexity, and output scoring parameters representing model selection tendency, and defining fuzzy sets for each parameter; S30, configuring a membership function for each input parameter, and mapping the input parameter values ​​collected in real time to the corresponding fuzzy set; S40, establishing a mapping rule between an input parameter fuzzy set and an output score fuzzy set based on an expert experience knowledge base of the video reasoning scenario; S50, collecting input parameter values ​​of the video frame, mapping them to fuzzy sets through membership functions, and generating fuzzy subsets of output scores in combination with the mapping rules; S60, performing a centroid calculation on the fuzzy subset of the output score to obtain an accurate model selection score; S70: Determine the optimal inference model for the next frame based on the model selection score of the current frame and the preset anti-switching jitter strategy.

2. The video enhancement reasoning method based on fuzzy control according to claim 1 is characterized in that: In step S10, the plurality of pre-trained video inference models of different computational complexity include a small-scale model, a medium-scale model, and a large-scale model, and the medium-scale model is used as the initial frame inference model by default.

3. The video enhancement reasoning method based on fuzzy control according to claim 1 is characterized in that: In step S20, the input parameters include: GPU utilization, whose fuzzy set is defined as {low utilization, medium utilization, high utilization}; GPU temperature, whose fuzzy set is defined as {temperature low, temperature medium, temperature high}; The number of targets in the current frame, whose fuzzy set is defined as {small number, medium number, large number}; The output scoring parameter is the model recommendation score, and its fuzzy semantic label is {low score, medium score, high score}.

4. The video enhancement reasoning method based on fuzzy control according to claim 1 is characterized in that: In step S30, the membership function is a piecewise linear function, which is used to map the numerical value of the input variable to a membership in the interval [0, 1].

5. The video enhancement reasoning method based on fuzzy control according to claim 1 is characterized in that: In step S40, the mapping rules are stored in the form of a fuzzy rule table, and the fuzzy rule table is established based on the video reasoning experience knowledge base and measurement data.

6. The video enhancement reasoning method based on fuzzy control according to claim 1 is characterized in that: In step S70, the mapping relationship between the model selection score and the inference model is: When the score is in the interval [0,30], the large-scale model is selected; When the score is in the range of (30,50], the medium-sized model is selected; When the score is in the range of (50,100], a small-scale model is selected.

7. The video enhancement reasoning method based on fuzzy control according to claim 1 is characterized in that: In step S70, the anti-switching jitter strategy is: the model switching operation is performed only when the model selection results of n consecutive frames are consistent and different from the current model.

8. A video enhancement reasoning device based on fuzzy control, characterized in that: include: Model loading module, used to load multiple pre-trained video inference models of different computational complexity; The fuzzy variable definition module is used to define input parameters that reflect the hardware resource status and scenario complexity, as well as output scoring parameters that represent the model selection tendency, and define fuzzy sets for each parameter; The membership configuration module is used to configure the membership function for each input parameter and map the input parameter values ​​collected in real time to the corresponding fuzzy set; The rule base module is used to establish the mapping rules between the input parameter fuzzy set and the output score fuzzy set based on the expert experience knowledge base of the video reasoning scenario; A fuzzy control module is used to collect input parameter values ​​of video frames, map them to fuzzy sets through membership functions, and generate fuzzy subsets of output scores based on the mapping rules; The score conversion module is used to perform centroid calculation on the fuzzy subset of the output score to obtain an accurate model selection score; The model decision module is used to determine the optimal inference model for the next frame based on the model selection score of the current frame and the preset anti-switching jitter strategy.

9. A computer device, characterized in that: The computer device includes a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the video enhancement reasoning method based on fuzzy control according to any one of claims 1 to 7 is implemented.

10. A storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the video enhancement reasoning method based on fuzzy control according to any one of claims 1 to 7 can be implemented.

Images