A physical quantity position information management method and system

Through image recognition technology and sensor signal monitoring, the problem of time-consuming and risk monitoring of manual entry in temperature tag management is solved, and rapid encoding entry and real-time risk alarm is achieved, which improves the efficiency and safety of temperature tag management.

CN117522966BActive Publication Date: 2024-11-19COSCO SHIPPING +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202311528873.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-11-15
Publication Date
2024-11-19
Estimated Expiration
2043-11-15

AI Technical Summary

Technical Problem

In the prior art, temperature tag management relies on manual entry and inspection, which leads to time-consuming and labor-intensive, lacks a quick entry framework coding scheme, and is unable to monitor and alarm the risk of temperature tag falling off in time, affecting cargo temperature monitoring and safety.

Method used

通过图像识别技术获取框架和温度标签编码,结合传感器信号强度值监控潜在脱落风险,发出提示信息通知工作人员进行确认和补正。

Benefits of technology

It improves the coding and entry efficiency, reduces manual errors, realizes real-time risk monitoring of fall off and timely alarms, ensures continuous monitoring of cargo temperature, and reduces the incidence of accidents.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117522966B_ABST
    Figure CN117522966B_ABST
Patent Text Reader

Abstract

The present invention discloses a physical quantity position information management method and system, belonging to the technical field of management systems, the physical quantity includes the measurement and management of other physical quantity position information such as temperature, humidity, displacement, acceleration, load and signal strength, the method includes: obtaining a frame image with a temperature label attached; preprocessing the frame image; processing the frame image through an image recognition algorithm to obtain a frame code; scanning the temperature label attached to the frame to obtain the temperature label code; receiving the sensor signal sent by the temperature label to obtain the received sensor signal strength value; when the sensor signal strength value is less than a preset value, it is determined that the temperature label has a potential risk of falling off; issuing a prompt message to notify the staff to go to the temperature label with a potential risk of falling off according to the frame code and the temperature label code for manual confirmation and correction. The present invention can improve the efficiency of coding entry and timely discover potential risks of falling off.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of management systems, and in particular relates to a physical quantity position information management method and system. Background Art

[0002] At present, temperature labels are widely used in the fields of industry, medical treatment, food, etc. to monitor and record the temperature information of objects to ensure product quality and work safety. However, the existing physical quantity location information management methods have some limitations, such as the lack of real-time monitoring of the temperature label status, which makes the temperature label management difficult and the inability to monitor the effectiveness of the label, which may affect product traceability and operator work efficiency.

[0003] Traditional physical quantity location information management methods usually rely on manual input and manual inspection, which is time-consuming and labor-intensive.

[0004] When manually recording the frame code and temperature label code, a QR code is often set on the temperature label, and the temperature label code is automatically recorded by scanning the QR code. However, this method is difficult to use to record the frame code. It is impossible to set the QR code on the frame. On the one hand, most frames are made of metal, and the cost of engraving the QR code is too high. On the other hand, if the QR code is pasted on the metal, it is easy to fall off, which will make the subsequent frame code unrecognizable. Therefore, there is currently a lack of solutions for quickly entering the frame code.

[0005] Furthermore, manual inspection to check whether the temperature label has fallen off cannot provide timely alarms and intervention measures, and cannot guarantee a quick response when the risk of the temperature label falling off occurs. Once the temperature label falls off, the temperature of the goods will be lost and effective intervention cannot be made when an accident occurs. Summary of the invention

[0006] In order to solve the problems that the existing technology of managing temperature labels through manual entry and manual inspection is time-consuming and laborious, there is a lack of a solution for quickly entering framework codes when manually recording framework codes and temperature label codes, and it is impossible to provide timely alarms and intervention measures when checking whether the temperature labels have fallen off through manual inspections, and it is impossible to ensure that a quick response can be made when there is a risk of temperature labels falling off. Once the temperature labels fall off, the temperature of the goods will be lost, and risks are prone to loss of control. The present invention provides a physical quantity location information management method and system.

[0007] First aspect

[0008] The present invention provides a physical quantity position information management method, the physical quantity includes temperature, and the physical quantity position information management method includes:

[0009] S101: Acquire a frame image with a temperature label attached;

[0010] S102: Preprocessing the frame image, the preprocessing including: grayscale, filtering and noise reduction, brightness balance and contrast balance;

[0011] S103: Processing the frame image by using an image recognition algorithm to obtain a frame code;

[0012] S104: Scan the temperature label attached to the frame to obtain the temperature label code;

[0013] S105: receiving a sensor signal sent by the temperature tag, and obtaining a received sensor signal strength value;

[0014] S106: When the sensor signal strength value is less than a preset value, it is determined that the temperature label has a potential risk of falling off;

[0015] S107: A prompt message is issued to inform the staff to go to the temperature label where there is a potential risk of falling off according to the frame code and the temperature label code for manual confirmation and correction.

[0016] Second aspect

[0017] The present invention provides a temperature label management system, which is used to execute the physical quantity position information management method in the first aspect.

[0018] Compared with the prior art, the present invention has at least the following beneficial technical effects:

[0019] (1) In the present invention, the frame code and the temperature label code are accurately obtained through image recognition technology, which improves the efficiency of code entry, saves time, reduces labor costs, and can reduce manual entry errors and improve the accuracy of temperature label management.

[0020] (2) In the present invention, whether there is a potential risk of falling off is determined based on the sensor signal strength value emitted by the temperature tag. When the risk of falling off occurs, the system can immediately issue a prompt message to notify the staff to go and deal with it, thereby avoiding the expansion and delay of the problem. Real-time monitoring of the risk of falling off can be carried out, and timely alarms and intervention measures can be provided. It can ensure that a quick response is made when the risk of temperature label falling off occurs, realize continuous monitoring of the temperature of the goods, and reduce the accident rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The preferred implementation modes will be described below in a clear and understandable manner with reference to the accompanying drawings to further illustrate the above-mentioned characteristics, technical features, advantages and implementation methods of the present invention.

[0022] Figure 1It is a flow chart of a physical quantity position information management method provided by the present invention. DETAILED DESCRIPTION

[0023] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the specific implementation methods of the present invention will be described below with reference to the accompanying drawings. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings and other implementation methods can be obtained based on these drawings without creative work.

[0024] Example 1

[0025] In one embodiment, the reference specification Figure 1 , shows a flow chart of a physical quantity position information management method provided by the present invention. The physical quantity position information of the present invention refers to a temperature tag; of course, it can also be a humidity tag, a displacement tag, an acceleration tag, and a signal strength tag.

[0026] The physical quantity position information management method and system provided by the present invention are not limited to the management of temperature tags for measuring temperature, but can also be applied to the measurement and management of other physical quantity position information such as humidity, displacement, acceleration and signal strength.

[0027] The present invention provides a physical quantity position information management method, the physical quantity includes temperature, and the physical quantity position information management method includes:

[0028] S101: Acquire a frame image with a temperature label attached.

[0029] Specifically, the frame image with the temperature label attached thereto can be obtained through the camera.

[0030] S102: Preprocess the frame image.

[0031] Among them, preprocessing includes: grayscale, filtering and noise reduction, brightness balance and contrast balance.

[0032] It should be noted that since the collected images are RGB color images with three-channel color space, their image feature information is extremely rich. If the color images are directly processed and analyzed, there are disadvantages of complex operation and time-consuming calculation. In addition, color cannot help in obtaining frame coding. In order to improve the real-time performance of the system detection, the color images need to be grayed out.

[0033] Specifically, the maximum value method or the average value method can usually be used to grayscale a color image. The maximum value method compares the pixel values ​​of the three channels of the color image, and the maximum pixel value is used as the grayscale value of the grayscale image. The average value method averages the pixel values ​​of the three channels of the color image, and the average pixel value is used as the grayscale value of the grayscale image.

[0034] In a possible implementation, the present invention provides a new method for graying a frame image, which can characterize the features and contrast information of a color image as much as possible with less information, and compared with the maximum value method or the average value method, the brightness of the image processed by the graying of the present invention is more balanced, and there will be no excessive brightness or excessive brightness, which is more consistent and natural visually. S102 specifically includes sub-steps S1021 and S1022:

[0035] S1021: Obtain the pixel value of each pixel point of the frame image in the R channel, the G channel and the B channel.

[0036] S1022: grayscale the frame image using the following formula:

[0037] Gray(i,j)=0.299×R(i,j)+0.578×G(i,j)+0.123×B(i,j)

[0038] Among them, (i, j) represents the coordinates of the pixel point, Gray(i, j) represents the grayscale value of the pixel point with coordinates (i, j), R(i, j) represents the R channel pixel value of the pixel point with coordinates (i, j), G(i, j) represents the G channel pixel value of the pixel point with coordinates (i, j), and B(i, j) represents the B channel pixel value of the pixel point with coordinates (i, j).

[0039] The weight of the R channel pixel value is set to 0.299, the weight of the G channel pixel value is set to 0.578, and the weight of the B channel pixel value is set to 0.123, which is determined based on the human eye's perception sensitivity to different color channels. The human eye has a relatively high perception of green and a relatively low perception of blue. Therefore, the use of weight distribution can better simulate the human eye's perception of images, making grayscale images more visually consistent and natural. Weight distribution can reduce redundant information between different channels. In some cases, different channels may contain similar information, and the use of weighted average can merge redundant information to make the grayscale image more representative.

[0040] In the present invention, the weighted average grayscale processing method can better simulate the perception of the human eye, so that the brightness of the image processed by the grayscale of the present invention is more balanced, without too low brightness or too high brightness, and it is more consistent and natural visually.

[0041] It should be noted that the frame code is easily interfered by image noise, dust particles and printing dots. Therefore, it is necessary to filter and reduce noise on the frame image after grayscale processing.

[0042] Specifically, mean filtering, Gaussian filtering, median filtering, etc. may be used to perform filtering and noise reduction processing on the grayscale processed frame image.

[0043] In a possible implementation, S102 specifically further includes sub-step S1023:

[0044] S1023: Filter and reduce noise on the grayscale frame image according to the following formula:

[0045]

[0046] Among them, f m (x, y) represents the pixel value after filtering and denoising at the pixel point (x, y), x represents the horizontal coordinate of the pixel point, y represents the vertical coordinate of the pixel point, σ represents the standard deviation of the pixel value in the neighborhood, and e represents the natural logarithm.

[0047] It should be noted that Gaussian filtering performs weighted averaging on pixels in the neighborhood, and the weight gradually decreases as the distance increases, which means that farther pixels have less influence on the current pixel, and closer pixels have a greater influence. This local weighted average allows the filtered image to maintain edge clarity while being smooth.

[0048] In the present invention, Gaussian filtering can effectively eliminate high-frequency noise in the image while retaining the edge information of the image. Gaussian filtering performs well in blurring images, reducing noise and smoothing images, so it is suitable for noise and interference problems that may exist in frame images.

[0049] It should be noted that brightness equalization can remap the pixel values ​​in an image so that the brightness range is more evenly distributed, thereby highlighting the details in the image and improving the visual effect of the image.

[0050] In a possible implementation, S102 specifically further includes sub-step S1024:

[0051] S1024: Perform brightness equalization on the frame image after the noise reduction filtering process according to the following formula:

[0052] I out =255c·F γ

[0053]

[0054] γ=log2(mFm )

[0055] Among them, I out Indicates the brightness value after brightness equalization processing, I in represents the brightness value before brightness equalization processing, c represents the proportional coefficient, F represents the normalized brightness value, γ represents the gamma correction index, and F m represents the median value of the brightness before brightness equalization processing, m represents the target brightness mean, and m can be obtained in the following way:

[0056]

[0057] in, It represents the mean of the brightness values ​​before brightness equalization processing, and round(.) represents the rounding function.

[0058] In the present invention, brightness equalization can remap an image with uneven brightness distribution to a more uniform distribution, which helps to highlight the details in the image. The details of both dark and bright parts can be displayed more clearly, thereby improving the visual effect of the image. In the image preprocessing process, brightness equalization can provide a better basis for subsequent image processing steps. For example, in recognition or analysis tasks, the image after brightness equalization can better show the target features, thereby improving the accuracy of subsequent processing.

[0059] It should be noted that contrast equalization can make the brightness range of the image more widely distributed in the entire grayscale range, thereby highlighting the details of the image and enhancing the visual effect of the image.

[0060] Specifically, contrast equalization processing may be performed on the frame image after brightness equalization processing by contrast transformation.

[0061] In a possible implementation manner, in order to achieve a better contrast equalization effect, the present invention provides a new contrast equalization method, and S102 specifically further includes sub-steps S1025 to S1028:

[0062] S1025: Construct energy change function:

[0063]

[0064] X = (x, y, t)

[0065] W = (u, v, l)

[0066] Where E(W) represents the energy change function, X represents the gray value of the pixel (x, y) at time t, W represents the displacement vector, u represents the displacement in the x direction, v represents the displacement in the y direction, and I represents the displacement in the t direction. Represents a three-dimensional gradient symbol, Represents the consistency parameter of gray value, represents the gradient of the displacement vector W, ||*||2 represents the Euclidean norm symbol, Represents the smoothness parameter of the gray value, f t (x, y) represents the pixel value after brightness equalization at the pixel point (x, y).

[0067] The energy change function is usually calculated by comparing the pixel values ​​at different positions and time points in the image. The energy change function can be used to describe image features such as moving objects, texture changes, edge extraction, and illumination changes.

[0068] S1026: Construct the Euler Lagrange equation based on the energy change function:

[0069]

[0070] in,

[0071] S1027: Solve the Euler Lagrange equation and calculate the horizontal component u(x, y) and vertical component v(x, y) of the energy change function in the spatial domain:

[0072]

[0073] in, represents the average value of u, u represents the displacement in the x direction, represents the average value of v, v represents the displacement in the y direction, F x represents the partial derivative of F with respect to x, F y represents the partial derivative of F with respect to y, F t represents the partial derivative of F with respect to t.

[0074] S1028: According to the horizontal component u(x, y) and the vertical component v(x, y) of the energy change function in the spatial domain, contrast equalization is performed on the frame image after brightness equalization:

[0075]

[0076] Among them, f d (x, y) represents the pixel value after contrast equalization at the pixel point (x, y).

[0077] It should be noted that by solving the energy change function and performing contrast equalization operations, the visual effect of the image can be improved, edge information and details can be enhanced, and the quality and observation ability of the image can be improved. Contrast equalization is of great significance for the frame coding acquisition task.

[0078] In the present invention, graying, filtering and noise reduction, brightness balance and contrast balance can be used to uniformly standardize the frame images collected under different environments, which has high applicability and is convenient for the promotion and popularization of temperature label management technology.

[0079] S103: Process the frame image by using an image recognition algorithm to obtain a frame code.

[0080] Specifically, framework encoding can be obtained through machine learning and deep learning algorithms.

[0081] In a possible implementation, S103 specifically includes sub-steps S1031 to S1035:

[0082] S1031: Based on the YOLO algorithm, a frame encoding recognition model is constructed.

[0083] Among them, YOLO (You Only Look Once) is a popular target detection algorithm, and its main feature is that it can simultaneously locate and classify objects in one forward pass (only one network inference is required). This makes it excellent in real-time performance and suitable for scenarios that require fast and accurate detection, such as real-time video analysis, autonomous driving, object tracking, etc.

[0084] In the present invention, the YOLO algorithm is used to locate a detection frame containing a frame code from a frame image.

[0085] S1032: Determine the loss function of the framework encoding recognition model.

[0086] In a possible implementation, the loss function is specifically:

[0087]

[0088] Among them, L represents the loss function, L kpts represents the key point loss, K represents the number of output feature maps, k = 1, 2, ..., K, α k represents the weight of each output feature map, s 2 Indicates the number of grids, i = 1, 2, ..., s 2 , B represents the number of prediction boxes on each grid, j = 1, 2, ..., B, L box represents the prediction box loss, β box represents the weight of the prediction box loss, L obj represents the confidence loss, β obj Represents the weight of confidence loss, L cls represents the category loss, β cls Represents the weight of the class loss.

[0089] Among them, the key point loss is mainly used to process the accuracy of the key points on the target object. The key point loss measures the error between the predicted key point position and the actual position to enable the model to predict the key point more accurately.

[0090] Among them, the prediction box loss is used to ensure the accuracy of the detection box. In object detection, each detection box is represented by its bounding box coordinates (position and size). The prediction box loss evaluates the gap between the bounding box coordinates of the predicted box and the true bounding box coordinates so that the model can better locate the object.

[0091] Among them, confidence loss refers to whether each detection box predicted by the model contains the target. Confidence loss is used to evaluate the prediction accuracy of the model on confidence. For boxes containing targets, the model needs to predict high confidence, and for boxes that do not contain targets, the model needs to predict low confidence.

[0092] Among them, the category loss is used to ensure the accuracy of the model in the classification task. Each detection box needs to predict which category of object it belongs to. The category loss measures the error between the model's prediction of the target category and the actual category to help the model better classify.

[0093] Those skilled in the art can set the weight α of each output feature map according to actual conditions. k , the weight of the prediction box loss β box , the weight of confidence loss β obj and the weight of the class loss β cls The present invention does not limit the size.

[0094] Furthermore, the loss function of each part has corresponding weights, which can be adjusted according to the importance of the task. This allows the model to more flexibly adjust the importance of different tasks, thereby better adapting to actual needs.

[0095] In the present invention, the loss functions of different aspects such as key point loss, prediction box loss, confidence loss and category loss are integrated together, which helps to improve the comprehensive performance, multi-task learning ability and flexibility of the framework code recognition model, thereby bringing better results for the framework code recognition tasks in practical applications.

[0096] S1033: With the goal of minimizing the loss function, the frame encoding recognition model is trained through training samples.

[0097] S1034: Segmenting a frame coding detection frame through the trained frame coding recognition model.

[0098] S1035: Identify the frame code in the frame code detection frame through a convolutional recurrent neural network.

[0099] In the present invention, the frame code and the temperature label code are accurately obtained through image recognition technology, which improves the efficiency of code entry, saves time, reduces labor costs, reduces manual entry errors, and improves the accuracy of temperature label management.

[0100] In a possible implementation, sub-step S1035 specifically includes sub-steps S10351 to S10356:

[0101] S10351: Extracting coding features in the frame coding detection frame.

[0102] S10352: Construct a recurrent neural network, which includes: an input layer, a state layer, an attention layer, and an output layer.

[0103] S10353: In the input layer, input the encoded feature sequence [x1, x2, ..., x m ], where m represents the total number of license plate character features, x m represents the mth encoded feature.

[0104] S10354: In the state layer, calculate the hidden state at time t:

[0105]

[0106]

[0107]

[0108] Among them, h t represents the hidden state at time t, represents the previous state of the forward loop, represents the previous state of the backward loop, GRU() represents the nonlinear calculation after the recurrent neural network, x t represents the encoding feature at time t, u t express The weight coefficient, v t express The weight coefficient, b t Represents the bias term of the hidden state at time t.

[0109] Hidden states refer to states that are not directly visible in a model or algorithm. These states carry hidden information about the data and are usually used by the model to represent the intrinsic features or potential structure of the input data. Hidden states may have different meanings in different contexts, but they are the core components of the model in many cases.

[0110] S10355: In the attention layer, assign weights to each overall performance value and add them up to get the hidden state of the current attention layer:

[0111]

[0112] Among them, s represents the hidden state of the current attention layer, γ t represents the weight at time t, h t represents the hidden state at time t.

[0113] S10356: In the output layer, the frame code is identified based on the hidden state of the current attention layer.

[0114] In the present invention, through RNN, attention mechanism and other methods, the system can better understand the contextual relationship of the encoded features in the sequence, so that the system can make more accurate predictions based on the information of the entire sequence, rather than just based on a single feature. Further, the combination of recurrent neural network, attention mechanism and other technologies enables the system to better understand the relationship in the sequence data, thereby improving the accuracy and robustness of frame encoding recognition.

[0115] S104: Scan the temperature label attached to the frame to obtain the temperature label code.

[0116] Specifically, a QR code can be set on the temperature label, and the temperature label code can be obtained by scanning the QR code with a camera.

[0117] S105: Receive the sensor signal sent by the temperature tag, and obtain the received sensor signal strength value.

[0118] Among them, the sensor signal can be a Bluetooth signal, an NFC signal, a UWB signal, etc.

[0119] S106: When the sensor signal strength value is less than a preset value, it is determined that there is a potential risk of the temperature label falling off.

[0120] Specifically, the preset value may be -60dbm. Those skilled in the art may set the preset value according to actual conditions, and the present invention does not limit this.

[0121] S107: A prompt message is issued to inform the staff to go to the temperature label where there is a potential risk of falling off according to the frame code and the temperature label code for manual confirmation and correction.

[0122] In the present invention, whether there is a potential risk of falling off is determined based on the sensor signal strength value emitted by the temperature tag. When the risk of falling off occurs, the system can immediately issue a prompt message to notify the staff to go and deal with it, thereby avoiding the expansion and delay of the problem. Real-time monitoring of the risk of falling off can be performed, and timely alarms and intervention measures can be provided to ensure a quick response when the risk of temperature tag falling off occurs.

[0123] In a possible implementation, the physical quantity position information management method further includes step S108:

[0124] S108: Adjust the preset value according to the result of manual confirmation by the staff at the temperature label where there is a potential risk of falling off.

[0125] It should be noted that the actual confirmation by the staff can verify whether the risk of falling off predicted by the system through the Bluetooth signal is accurate. The staff can check whether there is really a risk of falling off, thereby providing actual feedback information. Based on the confirmation results of the staff, it is possible to evaluate whether the preset value is appropriate. If the preset value is set too conservatively, frequent false alarms may occur. On the contrary, if the setting is too aggressive, false alarms may be missed. By constantly adjusting the preset value, it can be made closer to the actual situation, thereby improving the prediction accuracy of the system. This process of adjusting the preset value makes the system adaptive. The system is able to optimize according to the actual situation, so as to adapt to different environments and application scenarios and provide better performance and effects.

[0126] Compared with the prior art, the present invention has at least the following beneficial technical effects:

[0127] (1) In the present invention, the frame code and the temperature label code are accurately obtained through image recognition technology, which improves the efficiency of code entry, saves time, reduces labor costs, and can reduce manual entry errors and improve the accuracy of temperature label management.

[0128] (2) In the present invention, whether there is a potential risk of falling off is determined based on the sensor signal strength value emitted by the temperature tag. When the risk of falling off occurs, the system can immediately issue a prompt message to notify the staff to go and deal with it, thereby avoiding the expansion and delay of the problem. Real-time monitoring of the risk of falling off can be carried out, and timely alarms and intervention measures can be provided. It can ensure that a quick response is made when the risk of temperature label falling off occurs, realize continuous monitoring of the temperature of the goods, and reduce the accident rate.

[0129] Example 2

[0130] In one embodiment, the present invention provides a temperature label management system for executing the physical quantity position information management method in Embodiment 1.

[0131] A temperature label management system provided by the present invention can implement the steps and effects of the physical quantity position information management method in the above-mentioned embodiment 1. To avoid repetition, the present invention will not go into details.

[0132] Compared with the prior art, the present invention has at least the following beneficial technical effects:

[0133] (1) In the present invention, the frame code and the temperature label code are accurately obtained through image recognition technology, which improves the efficiency of code entry, saves time, reduces labor costs, and can reduce manual entry errors and improve the accuracy of temperature label management.

[0134] (2) In the present invention, whether there is a potential risk of falling off is determined based on the sensor signal strength value emitted by the temperature tag. When the risk of falling off occurs, the system can immediately issue a prompt message to notify the staff to go and deal with it, thereby avoiding the expansion and delay of the problem. Real-time monitoring of the risk of falling off can be carried out, and timely alarms and intervention measures can be provided. It can ensure that a quick response is made when the risk of temperature label falling off occurs, realize continuous monitoring of the temperature of the goods, and reduce the accident rate.

[0135] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0136] The above embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for those of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A method for managing physical quantity position information, characterized in that: The physical quantity includes temperature, and the physical quantity position information management method includes: S101: Acquire a frame image with a temperature label attached; S102: Preprocessing the frame image, the preprocessing including: grayscale, filtering and noise reduction, brightness balance and contrast balance; S103: Processing the frame image by using an image recognition algorithm to obtain a frame code; S104: Scan the temperature label attached to the frame to obtain the temperature label code; S105: receiving a sensor signal sent by the temperature tag, and obtaining a received sensor signal strength value; S106: When the sensor signal strength value is less than a preset value, it is determined that the temperature label has a potential risk of falling off; S107: Send a prompt message to inform the staff to go to the temperature label where there is a potential risk of falling off according to the frame code and the temperature label code for manual confirmation and correction; S108: adjusting the preset value according to the result of manual confirmation by a staff member at a temperature tag with a potential risk of falling off; The S103 specifically includes: S1031: Build a frame encoding recognition model based on the YOLO algorithm; S1032: Determine the loss function of the framework encoding recognition model; S1033: training the framework encoding recognition model through training samples with the goal of minimizing the loss function; S1034: Segmenting a frame code detection frame using the trained frame code recognition model; S1035: identifying the frame code in the frame code detection frame through a convolutional recurrent neural network; The loss function is specifically: Among them, L represents the loss function, L kpts represents the key point loss, K represents the number of output feature maps, k = 1, 2, ..., K, α k represents the weight of each output feature map, s 2 Indicates the number of grids, i=1,2,…,s 2 , B represents the number of prediction boxes on each grid, j = 1, 2, ..., B, L box represents the prediction box loss, β box represents the weight of the prediction box loss, L obj represents the confidence loss, β obj Represents the weight of confidence loss, L cls represents the category loss, β cls Represents the weight of the class loss; The S1035 specifically includes: S10351: extracting coding features in the framework coding detection frame; S10352: Construct a recurrent neural network, wherein the recurrent neural network includes: an input layer, a state layer, an attention layer, and an output layer; S10353: In the input layer, input the coded feature sequence [x1, x2, ..., x m ], where m represents the total number of license plate character features, x m represents the mth encoding feature; S10354: In the state layer, the hidden state at time t is calculated: Among them, h t represents the hidden state at time t, represents the previous state of the forward loop, represents the previous state of the backward loop, GRU() represents the nonlinear calculation after the recurrent neural network, x t represents the encoding feature at time t, u t express The weight coefficient, v t express The weight coefficient, b t Represents the bias term of the hidden state at time t; S10355: In the attention layer, assign a weight to each overall performance value, and add them up to obtain the current hidden state of the attention layer: Among them, s represents the hidden state of the current attention layer, γ t represents the weight at time t, h t represents the hidden state at time t; S10356: In the output layer, the frame code is identified according to the hidden state of the current attention layer.

2. The physical quantity position information management method according to claim 1, characterized in that: The S102 specifically includes: S1023: Filter and reduce noise on the grayscale frame image according to the following formula: Among them, f m (x, y) represents the pixel value after filtering and denoising at the pixel point (x, y), x represents the horizontal coordinate of the pixel point, y represents the vertical coordinate of the pixel point, σ represents the standard deviation of the pixel value in the neighborhood, and e represents the natural logarithm.

3. The physical quantity position information management method according to claim 2, characterized in that: The S102 specifically includes: S1024: Perform brightness equalization on the frame image after the noise reduction filtering process according to the following formula: I out =255c·Fγ γ=log2(mF m ) Among them, I out Indicates the brightness value after brightness equalization processing, I in represents the brightness value before brightness equalization processing, c represents the proportional coefficient, F represents the normalized brightness value, γ represents the gamma correction index, and F m represents the median value of the brightness before brightness equalization processing, m represents the target brightness mean, and m can be obtained in the following way: in, represents the mean of the brightness values ​​before brightness equalization processing, and round(·) represents the rounding function.

4. A temperature label management system, characterized in that: Used to execute the physical quantity position information management method described in any one of claims 1 to 3.