Temperature detection method and device for mine engineering tire

Abstract

The invention discloses a temperature detection method and device for a mine engineering tire. The method comprises the following steps: acquiring a target thermal imaging image of a mine target area, and determining position information of each target tire on a target vehicle in the target thermal imaging image by using a target detection model; intercepting a tire area image including the target tire from the thermal imaging image according to the position information of the target tire, and dividing the tire area image into a tread area image and a sidewall area image by using a target segmentation model; the temperature information in the tread area image and the temperature information in the sidewall area image are subjected to statistics, the temperature states of the tread and the sidewall of the target tire are determined according to the statistical result, the temperature states are displayed, and the temperature states are used for reflecting whether the temperature of the tread or the sidewall is abnormal or not and the temperature abnormal area corresponding to the abnormal temperature. The technical problem that a traditional tire temperature detection scheme is not suitable for engineering tires working in the mine environment is solved.

Description

Technical Field

[0001] The present application relates to the technical field of tire detection, and in particular, to a method and device for detecting the temperature of a mine engineering tire. Background Art

[0002] As a key device in mining operations, the performance and health status of the tires of mining transport vehicles are directly related to the stability and operation efficiency of the vehicles. The particularity of the mining environment, such as high temperature, heavy load, many gravels and complex terrain, makes the tires extremely vulnerable to wear and damage during operation. In particular, the abnormal increase in the surface temperature of the tires will accelerate the aging and damage of the tires. Therefore, monitoring and controlling the surface temperature of mining tires to ensure that they work within a safe range is an urgent need in the maintenance of mining transport vehicles.

[0003] Traditional tire temperature detection methods mainly rely on temperature sensors installed on the tires. These sensors can monitor the temperature inside the tires or the contact surface in real time. However, they have obvious limitations. First, the installation process of the sensors is complex and costly, and they need to be accurately arranged on each tire, increasing the maintenance difficulty and operating cost of mining vehicles. Second, the detection range of the sensors is limited, and they can only obtain local temperature information of the tires and cannot achieve comprehensive monitoring of the surface temperature of the entire tire. In the case of uneven surface temperature distribution of the tires, this may lead to missed detection of abnormal temperatures. Third, the mining operation environment is harsh, and the sensors are easily damaged or fail, thus reducing the accuracy and reliability of the detection. In addition, although infrared detectors can perform non-contact temperature detection, their point-type temperature measurement method also cannot cover the entire surface of the tire, and environmental factors such as dust and humidity will affect their temperature measurement accuracy.

[0004] No effective solution has been proposed for the above problems. Summary of the Invention

[0005] Embodiments of the present application provide a method and device for detecting the temperature of a mine engineering tire, so as to at least solve the technical problem that the traditional tire temperature detection scheme is not applicable to the engineering tires working in the mine environment.

[0006] According to one aspect of the embodiments of the present application, a method for detecting the temperature of a mine engineering tire is provided, including: obtaining a target thermal imaging image of a mine target area, and using a pre-trained target detection model to determine the position information of each target tire on a target vehicle in the target thermal imaging image; for each target tire, intercepting a tire area image including the target tire from the target thermal imaging image according to the position information of the target tire, and using a pre-trained target segmentation model to divide the tire area image into a tread area image and a sidewall area image, wherein the target segmentation model includes an efficient channel attention module, and the target segmentation model is trained based on an enhanced intersection over union loss function; respectively statistically analyzing the temperature information in the tread area image and the sidewall area image, determining the temperature states of the tread and the sidewall of the target tire according to the statistical results, and displaying the temperature states, where the temperature states are used to reflect whether there is a temperature anomaly in the tread or the sidewall and the corresponding temperature anomaly area when there is an anomaly.

[0007] Optionally, using a pre-trained target detection model to determine the position information of each target tire on a target vehicle in the target thermal imaging image includes: using the target detection model to analyze the target thermal imaging image, determining a first detection frame corresponding to the target vehicle in the target thermal imaging image, a second detection frame corresponding to each group of side-by-side tires on the target vehicle, and a third detection frame corresponding to each target tire in each group of side-by-side tires; determining the position information of each target tire according to the first detection frame, the second detection frame, and the third detection frame, where the position information includes: wheel position information and position coordinate information.

[0008] Optionally, determining the position information of each target tire according to the first detection frame, the second detection frame, and the third detection frame includes: for each target tire, determining the position coordinate information of the target tire according to the third detection frame corresponding to the target tire; determining the first wheel position information of the target tire in the side-by-side tires to which it belongs according to the positional relationship between the third detection frame corresponding to the target tire and the second detection frame to which the third detection frame belongs; determining the second wheel position information of the side-by-side tires to which the target tire belongs on the target vehicle according to the positional relationship between the second detection frame to which the third detection frame corresponding to the target tire belongs and the first detection frame.

[0009] Optionally, the training process of the target segmentation model includes: constructing an initial segmentation model, where the initial segmentation model is a YOLOv8 segmentation model with an efficient channel attention module added; obtaining multiple historical thermal imaging images collected on the mine that include vehicle tires, and cropping at least one sub-region image that includes a single tire from each historical thermal imaging image; using each sub-region image as a training sample, and annotating the tread region and the sidewall region in the sub-region image as sample labels; dividing the obtained multiple training samples into a training set, a validation set, and a test set according to a preset ratio; iteratively training the initial segmentation model using the training set, validating the model performance after each training cycle using the validation set, taking the trained model as the target segmentation model, and testing the performance of the target segmentation model using the test set.

[0010] Optionally, before dividing the obtained multiple training samples into a training set, a validation set, and a test set according to a preset ratio, the above method further includes: performing sample data augmentation processing on the multiple training samples, where the sample data augmentation processing methods include at least one of the following: random rotation, random scaling, and random flipping.

[0011] Optionally, iteratively training the initial segmentation model using the training set includes: obtaining preset hyperparameters and performing model training based on the hyperparameters, where the hyperparameters include at least one of the following: the number of training cycles, batch size, learning rate, optimizer; during the training process, inputting the multiple training samples corresponding to each training batch into the model, constructing an enhanced intersection over union loss function based on the overlap region, center point difference, width difference, and height difference between the predicted bounding box output by the model and the annotated bounding box in the sample label, and adjusting the model parameters based on the enhanced intersection over union loss function.

[0012] Optionally, respectively statistically analyzing the temperature information in the tread region image and the sidewall region image, and determining the temperature status of the tread and sidewall of the target tire according to the statistical results includes: for the target region image, determining the temperature value of each pixel in the target region image, where the target region image is the tread region image or the sidewall region image; determining the mode of all temperature values, and comparing the temperature value of each pixel with the mode, and determining the pixels whose first difference between the temperature value and the mode is greater than the preset temperature threshold as temperature abnormal pixels; connecting adjacent temperature abnormal pixels to obtain multiple connected regions, and determining the connected regions with an area greater than the preset area threshold as the temperature abnormal regions in the target region image.

[0013] Optionally, the above method further includes: for each temperature anomaly region in the target region image, determining the average value of the temperatures of all pixels in the temperature anomaly region, and determining the second difference between the average value and the mode; determining the target temperature range corresponding to the second difference from a preset temperature anomaly level table, and using the temperature anomaly level corresponding to the target temperature range as the temperature anomaly level of the temperature anomaly region, where the temperature anomaly level table stores the mapping relationships between multiple temperature ranges and multiple temperature anomaly levels.

[0014] Optionally, displaying the temperature status includes: for the target region of each target tire, determining the highest temperature anomaly level corresponding to each temperature anomaly region in the target region, where the target region is the tread region or the sidewall region; determining the first color corresponding to the highest temperature anomaly level from a preset temperature level color table, and displaying the temperature status of the target region based on the first color, and adding temperature anomaly warning information, where the temperature anomaly warning information includes at least one of the following: the position information of the target tire, the position information and area information of each temperature anomaly region, and the highest abnormal temperature.

[0015] Optionally, displaying the temperature status includes: generating a two-dimensional temperature display map of multiple target tires based on the target thermal imaging image, the position information of each target tire on the target vehicle, and the position information of the target region of each target tire, where the target region is the tread region or the sidewall region; for the target region of each target tire, determining the temperature anomaly level of each temperature anomaly region in the target region and the second color corresponding to the normal temperature from a preset temperature level color table, and updating the color of the corresponding position in the two-dimensional temperature display map according to the position information and the corresponding second color of each temperature anomaly region and the temperature normal region.

[0016] According to another aspect of the embodiments of the present application, there is also provided a temperature detection device for a mine engineering tire, including: a tire positioning module, configured to obtain a target thermal imaging image of a mine target region, and use a pre-trained target detection model to determine the position information of each target tire on the target vehicle in the target thermal imaging image; a segmentation module, configured to, for each target tire, intercept a tire region image including the target tire from the target thermal imaging image according to the position information of the target tire, and use a pre-trained target segmentation model to divide the tire region image into a tread region image and a sidewall region image, where the target segmentation model includes an efficient channel attention module, and the target segmentation model is trained based on an enhanced intersection over union loss function; a temperature detection module, configured to respectively count the temperature information in the tread region image and the sidewall region image, determine the temperature status of the tread and the sidewall of the target tire according to the statistical results, and display the temperature status, where the temperature status is used to reflect whether there is a temperature anomaly in the tread or the sidewall and the corresponding temperature anomaly region when an anomaly occurs.

[0017] According to another aspect of the embodiments of the present application, there is also provided a computer program product, which includes: a computer program, wherein when the computer program is executed by a processor, the temperature detection method of the mine engineering tire described above is implemented.

[0018] According to another aspect of the embodiments of the present application, there is also provided an electronic device, which includes: a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to execute the temperature detection method of the mine engineering tire described above through the computer program.

[0019] In the embodiments of the present application, a fixed infrared camera is used to capture a thermal imaging image and perform tire positioning and temperature detection. There is no need to deploy sensors on the vehicle tires, which can reduce the equipment investment cost and subsequent maintenance expenses; when performing tire temperature detection, by segmenting the tire sidewall and the tread of the tire and performing temperature detection separately, it is possible to avoid the influence of the ambient temperature and reduce the influence of the excessive temperature difference between the tread and the sidewall on the result, so as to achieve more accurate temperature detection of the tire. The present application effectively solves the technical problem that the traditional tire temperature detection scheme is not applicable to the engineering tires working in the mine environment. Description of the Drawings

[0020] The drawings described herein are used to provide a further understanding of the present application, and constitute a part of the present application. The schematic embodiments of the present application and their descriptions are used to explain the present application, and do not constitute an improper limitation to the present application. In the drawings:

[0021] Figure 1 is a schematic flowchart of an optional temperature detection method for a mine engineering tire according to an embodiment of the present application;

[0022] Figure 2 is a schematic diagram of an optional mine engineering tire target detection result according to an embodiment of the present application;

[0023] Figure 3 is a flowchart of an optional temperature detection method for a mine engineering tire according to an embodiment of the present application;

[0024] Figure 4 is a schematic structural diagram of an optional temperature detection device for a mine engineering tire according to an embodiment of the present application;

[0025] Figure 5 is a schematic structural diagram of an optional electronic device according to an embodiment of the present application. Detailed Embodiments

[0026] To enable those skilled in the art to better understand the solution of this application, the following will clearly and completely describe the technical solutions in the embodiments of this application in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of this application.

[0027] It should be noted that the terms "first", "second", etc. in the description, claims and drawings of this application are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of this application described here can be implemented in an order other than those illustrated or described here. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or are inherent to these processes, methods, products or devices.

[0028] To better understand the embodiments of this application, the following first translates and explains some of the nouns or terms that appear in the description process of the embodiments of this application:

[0029] ECA (Efficient Channel Attention): Its principle is to avoid the dimensionality reduction operation in the channel attention module. By adopting a local cross-channel interaction strategy, one-dimensional convolution is used to achieve efficient channel attention calculation. This method significantly reduces the complexity of the model while maintaining performance. By adaptively selecting the convolution kernel size, the coverage range of local cross-channel interaction is determined. The ECA module achieves a significant performance improvement with a small number of parameters and low computational cost, and has higher efficiency and better performance compared to other attention modules.

[0030] EIoU (Enhanced Intersection over Union Loss) function: The EIoU loss function is an improved version based on the IoU (Intersection over Union) concept. It is mainly applied in object detection tasks to optimize the matching degree between the predicted bounding box and the ground truth bounding box. IoU is an index to measure the overlapping degree between the predicted bounding box and the ground truth bounding box, and its calculation method is the area of the overlapping region between the two divided by the area of their union. However, IoU only considers the size of the overlapping region and ignores the distance and direction information between the predicted bounding box and the ground truth bounding box. To solve this problem, the EIoU loss function introduces distance loss and direction loss on the basis of IoU, enabling the loss function to simultaneously consider the overlapping region, the distance between the centers of the predicted bounding box and the ground truth bounding box, and the directions of the two boxes, thus more comprehensively measuring the similarity between the predicted bounding box and the ground truth bounding box.

[0031] YOLOv8 segmentation model: It is the latest version of the YOLO (You Only Look Once) series of models, evolved from its predecessor YOLOv7, aiming to further improve the performance and efficiency of object detection and segmentation. The YOLOv8 segmentation model, namely the Segmentation model of YOLOv8, is a computer vision model that focuses on the object segmentation task in images and can identify and mark the precise boundaries of different objects in the image.

[0032] Example 1

[0033] According to the embodiments of the present application, a method for detecting the temperature of a mining engineering tire is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. And although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0034] Figure 1 It is a schematic flowchart of a method for detecting the temperature of a mining engineering tire provided according to the embodiments of the present application. As Figure 1 shown, the method includes the following steps:

[0035] Step S102, obtain a target thermal imaging image of the mining target area, and use a pre-trained object detection model to determine the position information of each target tire on the target vehicle in the target thermal imaging image;

[0036] Step S104: For each target tire, intercept a tire region image including the target tire from the target thermal imaging image according to the position information of the target tire, and use a pre-trained target segmentation model to divide the tire region image into a tread region image and a sidewall region image. The target segmentation model includes an efficient channel attention module and is trained based on an enhanced intersection over union loss function.

[0037] Step S106: Statistically analyze the temperature information in the tread region image and the sidewall region image respectively, determine the temperature status of the tread and sidewall of the target tire according to the statistical results, and display the temperature status. The temperature status is used to reflect whether there is a temperature anomaly in the tread or sidewall and the corresponding temperature anomaly region when there is an anomaly.

[0038] The following describes each step of the temperature detection method for mine engineering tires in combination with a specific implementation process.

[0039] As an optional implementation manner, use a pre-trained target detection model to determine the position information of each target tire on the target vehicle in the target thermal imaging image, which can be specifically implemented by the following method: Analyze the target thermal imaging image using the target detection model to determine the first detection frame corresponding to the target vehicle in the target thermal imaging image, the second detection frame corresponding to each group of side-by-side tires on the target vehicle, and the third detection frame corresponding to each target tire in each group of side-by-side tires; Determine the position information of each target tire according to the first detection frame, the second detection frame, and the third detection frame. The position information includes wheel position information and position coordinate information.

[0040] Optionally, after respectively determining the first detection frame, the second detection frame, and the third detection frame, the position information of each target tire can be determined by the following method: For each target tire, determine the position coordinate information of the target tire according to the third detection frame corresponding to the target tire; Determine the first wheel position information of the target tire in the side-by-side tires to which it belongs according to the positional relationship between the third detection frame corresponding to the target tire and the second detection frame to which the third detection frame belongs; Determine the second wheel position information of the side-by-side tires to which the target tire belongs on the target vehicle according to the positional relationship between the second detection frame to which the third detection frame corresponding to the target tire belongs and the first detection frame.

[0041] Figure 2 Fig. shows a schematic diagram of an optional target detection result of mine engineering tires. A is the detection frame of the vehicle as a whole, B1 and B2 are the detection frames of side-by-side tires, and C1 - C4 are the detection frames of individual tires. Among them, A can be seen as the first detection frame, B1 and B2 can be regarded as the second detection frames, and C1 - C4 can be regarded as the third detection frames. Each third detection frame can uniquely frame a target tire. When determining the specific position of the target tire on the target vehicle, it can be implemented by the following method:

[0042] First, determine the first wheel position information. First, obtain the third detection frame information of the target tire, including the upper left and lower right corner coordinates of the third detection frame, and the category of the detection frame (the detection frame that frames the target tire), and find the second detection frame information associated with the third detection frame, that is, the detection frame information of the side-by-side tires to which the target tire belongs, wherein the second detection frame provides the overall position information of the side-by-side tires. Analyze the relative position relationship between the third detection frame and the second detection frame in the horizontal direction, such as the x-coordinate of the center point of the detection frame. If the x-coordinate of the third detection frame is smaller than the x-coordinate center point of the second detection frame to which it belongs, the target tire is located on the left side of the side-by-side tire group, otherwise, it is located on the right side.

[0043] Secondly, determine the second wheel position information, obtain the second detection frame information of the side-by-side tires to which the target tire belongs, including the coordinates of the upper left corner and the lower right corner of the detection frame, and the category of the detection frame (the detection frame that frames the side-by-side tires), find the first detection frame information associated with the second detection frame, that is, the detection frame information of the entire target vehicle, the first detection frame provides the position information of the entire vehicle, and analyze the relative position relationship between the second detection frame and the first detection frame in the horizontal direction, such as the x-coordinate of the center point of the detection frame. If the x-coordinate of the second detection frame is less than the x-coordinate center point of the first detection frame, it is determined that the side-by-side tires are on the left side of the vehicle; otherwise, the side-by-side tires are on the right side of the vehicle.

[0044] Finally, the first wheel position information and the second wheel position information are combined to jointly determine the specific position of the target tire on the target vehicle, such as if the target tire is the tire on the right side of the side-by-side tires on the left side of the vehicle, and the coordinate information of the target tire is given.

[0045] After obtaining the position information of each target tire on the target vehicle, the tire area image including the target tire can be captured from the target thermal imaging image according to the position information of the target tire, and the tire area image can be divided into a tread area image and a sidewall area image using a pre-trained target segmentation model.

[0046] Training a mature and complete target segmentation model is of key significance for the accurate segmentation of the tread area image and the sidewall area image. As an optional implementation, the training process of the target segmentation model can be implemented by steps S1 to S5. The specific process is as follows:

[0047] Step S1, constructing an initial segmentation model.

[0048] Take the YOLOv8 segmentation model with the added ECA module as the initial segmentation model. Since the core idea of the ECA module is to capture the dependencies between channels through one-dimensional convolution, compared with traditional attention mechanisms, the ECA module avoids complex dimensionality reduction and increase processes, thus achieving efficient and lightweight characteristics. The specific algorithmic ideas include the following steps:

[0049] Step S11: First, perform global average pooling on the feature map with an input size of W×H×C, that is, add up all the pixel values of the feature map and take the average to obtain a value, which is used to represent the corresponding feature map, resulting in a 1×1×C feature map, where W represents the width, H represents the height, and C represents the number of channels.

[0050] Step S12: Adaptively calculate the kernel size k of the one-dimensional convolution according to the number of channels. The calculation formula for the kernel size is as follows:

[0051]

[0052] where C is the number of channels of the input feature, and γ and b are hyperparameters. Taking the absolute value and rounding down to the nearest odd number is to ensure that the kernel size is odd. The above formula is used to calculate the kernel size k of the one-dimensional convolution.

[0053] Step S13: Set a padding to ensure that the number of output channels remains unchanged, then process the 1×1×C feature map with a one-dimensional convolution with the obtained kernel size k, and then pass through a Sigmoid activation function to activate the final 1×1×C attention weights.

[0054] Step S14: Apply the obtained 1×1×C attention weights to the input features to learn the importance of each channel relative to other channels. This process can be represented by the following formula:

[0055] out=Conv1D k (in)

[0056] The above formula means that the input feature in is converted into the output feature out through a one-dimensional convolution operation (kernel size k).

[0057] Step S2: Obtain multiple historical thermal imaging images collected on the mine that include vehicle tires, and intercept at least one sub-region image that includes a single tire from each historical thermal imaging image.

[0058] Among them, the collection and acquisition of multiple historical thermal imaging images ensure the comprehensiveness of the training sample dataset. Diverse historical images cover different lighting, angles, and background conditions, which can effectively enhance the robustness and generalization ability of the model.

[0059] Step S3: Take each sub-region image as a training sample, and label the tread region and sidewall region in the sub-region image as sample labels.

[0060] Specifically, an image segmentation semi-automated annotation tool can be used to perform segmentation and annotation operations on the tire. Since the segmentation operation is performed after detection, the annotated data are all images that only retain the tire region after cropping. The tread label is set to Tread, and the sidewall label is set to Side, generating a corresponding label file in json format.

[0061] Optionally, to further enhance the robustness of the model, data augmentation processing can be performed on the obtained sample data. The sample data augmentation processing methods at least include: random rotation, random scaling, and random flipping.

[0062] Among them, 1) Random rotation can increase the robustness of the model to rotational transformation by randomly rotating the image; 2) Random scaling can simulate objects at different scales and increase the adaptability of the model to scale changes by randomly scaling the image; 3) Random flipping can simulate scenarios such as mirroring and symmetry by randomly flipping the image horizontally or vertically, increasing the robustness of the model to mirror transformation. Enhancement processing is particularly important in a mine environment because vehicles may enter the detection area from various angles, and the size of the tire may also vary depending on the model and inflation status. By operations such as random rotation, scaling, and flipping, the diversity of training samples is increased, enabling the model to maintain good segmentation performance when facing tire images at different angles, sizes, and orientations.

[0063] Step S4: Divide the obtained multiple training samples into a training set, a validation set, and a test set according to a preset ratio.

[0064] Specifically, after data annotation, the json format file with labels and the corresponding source image are saved as a txt label according to the dataset annotation format of YOLOv8. The txt label file includes the target category and the coordinates of the points on the closed curve generated by segmentation. Among the data sets in the data folder, there are two folders named "images" and "labels" respectively, which are used to store images and the label files corresponding to the images. The data of the training set, validation set, and test set are stored in these two folders respectively. The ratio of the training set, validation set, and test set can be set to 8:1:1. The ratio set here is only an example and does not constitute a specific limitation.

[0065] Step S5: Use the training set to iteratively train the initial segmentation model, use the validation set to verify the model performance after each training cycle, take the trained model as the target segmentation model, and use the test set to test the performance of the target segmentation model.

[0066] As an alternative implementation, the iterative training of the initial segmentation model using the training set includes the following processes:

[0067] Step S51, obtain preset hyperparameters and perform model training according to the hyperparameters, where the hyperparameters include at least one of the following: the number of training epochs, batch size, learning rate, and optimizer.

[0068] For example, during the training process, the number of training rounds can be set to 300, the batch size to 8, the initial learning rate to 0.01, the learning rate momentum to 0.937, and the weight decay parameter of the optimizer to 0.0005. The relevant numerical values here are only for illustration and do not constitute specific limitations.

[0069] Step S52, during the training process, input multiple training samples corresponding to each training batch into the model, construct an enhanced intersection over union (IoU) loss function based on the overlap area, center point difference, width difference, and height difference between the predicted bounding box output by the model and the annotated bounding box in the sample label, and adjust the model parameters based on the enhanced IoU loss function.

[0070] Specifically, the enhanced IoU loss function can be expressed by the following formula:

[0071] L EIOU = L IOU + L dis + L asp

[0072] Among them, L IOU represents the IoU loss function, L dis represents the distance loss function, and L asp represents the direction loss function. The enhanced IoU loss function consists of three parts and can retain the effective characteristics of the complete IoU loss function.

[0073] L IOU 、L dis 、L asp can be expressed by the following formulas respectively:

[0074] L IOU = 1 - IOU

[0075]

[0076] Among them, IOU is the intersection over union, which is an index for evaluating the overlap degree between the detection or segmentation result and the true label. w c and h c are the width and height of the smallest bounding box covering the two detection boxes, b and b gt represent the center point coordinates of the predicted bounding box and the true bounding box respectively, and w and w gtrepresent the widths of the predicted bounding box and the ground truth bounding box, and h and h respectively gt represent the heights of the predicted bounding box and the ground truth bounding box respectively. ρ 2 Calculate the Euclidean distance between the embedding vectors.

[0077] When specifically calculating the IOU, first, find the intersection area of the two bounding boxes (the predicted bounding box and the ground truth bounding box), and then calculate the area of this intersection area; next, find the union area of the two boxes and calculate the area of the union area; finally, the IOU is the ratio of the intersection area to the union area. The value of the IOU is between 0 and 1, where 1 means the two boxes completely overlap and 0 means there is no overlap.

[0078] The training process of the above steps S1 - S5 follows the cross - validation principle, which can effectively optimize the model and improve the generalization ability of the model, thus ensuring the reliability of the model. Based on the feedback of the validation set and the test set, researchers can identify the weaknesses of the model, such as poor segmentation effect in certain specific types of objects or backgrounds, and then improve the model accordingly, such as adjusting the number of network layers, introducing new loss functions, optimizing the data augmentation strategy, etc., continuously promoting the improvement of the model performance.

[0079] Use a pre - trained object segmentation model to divide the tire area image into a tread area image and a sidewall area image. After detecting the tread and sidewall temperatures in the target area image, the temperature information of the tread and sidewall can be statistically analyzed in the following ways respectively, and the temperature states of the tread and sidewall of the target tire can be determined based on the statistical results: For the target area image, determine the temperature value of each pixel in the target area image, where the target area image is the tread area image or the sidewall area image; determine the mode of all temperature values, and compare the temperature value of each pixel with the mode, and determine the pixels whose first difference between the temperature value and the mode is greater than the preset temperature threshold as temperature - abnormal pixels; connect adjacent temperature - abnormal pixels to obtain multiple connected regions, and determine the connected regions with an area greater than the preset area threshold as the temperature - abnormal regions in the target area image.

[0080] Specifically, the brightness value of each pixel point can be read from the target area image, and the brightness value can be directly corresponding to the temperature value. Considering that environmental factors may affect the temperature data, the temperature data can be calibrated, including but not limited to operations such as background temperature compensation and atmospheric absorption correction, to improve the accuracy of temperature measurement. Structurally store the temperature value of each pixel to form a temperature matrix or a temperature map, which is convenient for subsequent mode calculation and temperature - abnormal detection.

[0081] Further, first, all the obtained temperature values are statistically analyzed to calculate the mode of the temperature values, that is, the temperature point with the highest occurrence frequency, which is used as the reference standard for normal temperature. Second, for each pixel in the target area image, the difference between its temperature value and the mode is calculated. The magnitude of the difference reflects the degree to which the temperature of this pixel deviates from the normal temperature. A preset temperature threshold is set, and based on the temperature difference, it is determined which pixels can be defined as temperature-abnormal pixels. The setting of the preset temperature threshold needs to consider factors such as the actual working conditions and material properties of the tire to ensure the accuracy of anomaly detection. Finally, adjacent temperature-abnormal pixels are connected to form multiple connected regions. By analyzing the size of the connected regions, the connected regions with an area larger than the preset area threshold are determined as the temperature-abnormal regions in the target area image.

[0082] Specifically, the temperature-abnormal regions can be determined in the following way: Based on the identified temperature-abnormal pixels, a binary image is generated, where the temperature-abnormal pixel points are set to 1 and the normal temperature points are set to 0. Using the connected-region analysis method in image processing technology, such as labeling connected regions, adjacent temperature-abnormal pixel points are connected to form multiple connected regions. A preset area threshold is set, and the area size of each connected region is analyzed. The connected regions with an area larger than the preset area threshold are selected, and these connected regions are the temperature-abnormal regions.

[0083] For each determined temperature-abnormal region, its average temperature, area, shape and other characteristics can be further analyzed to provide information for subsequent determination of abnormal temperature levels and early warnings.

[0084] As an alternative implementation, the temperature-abnormal regions can be graded in the following way: For each temperature-abnormal region in the target area image, the average value of the temperatures of all pixels in the temperature-abnormal region is determined, and the second difference between the average value and the mode is determined; The target temperature range corresponding to the second difference is determined from the preset temperature-abnormal level table, and the temperature-abnormal level corresponding to the target temperature range is used as the temperature-abnormal level of the temperature-abnormal region. Among them, the temperature-abnormal level table stores the mapping relationship between multiple temperature ranges and multiple temperature-abnormal levels.

[0085] Specifically, from the temperature-abnormal pixel map, the pixel points belonging to each temperature-abnormal region and their corresponding temperature values can be extracted. The average value of the extracted temperature values is calculated to obtain the average temperature of the temperature-abnormal region. For each temperature-abnormal region, the difference between its average temperature and the mode is calculated, that is, the second difference. The second difference temperature reflects the temperature fluctuation degree of the abnormal region. Generally speaking, the larger the difference, the more uneven the temperature distribution in this region, and there may be local overheating phenomena.

[0086] Suppose the preset temperature level table is as shown in Table 1.

[0087] Table 1

[0088] Temperature range (0℃, 5℃] Normal temperature Temperature range (5℃, 10℃] Temperature anomaly level 1 Temperature range (10℃, 15℃] Temperature anomaly level 2 Temperature range (15℃, +∞) Temperature anomaly level 3

[0089] As shown in the above table, when the second difference falls within the temperature range (0°C, 5°C], it indicates that the temperature is normal; when the second difference falls within the temperature range (5°C, 10°C], it corresponds to temperature anomaly level 1; when the second difference falls within the temperature range (10°C, 15°C], it corresponds to temperature anomaly level 2; when the second difference falls within the temperature range (15°C, +∞), it corresponds to temperature anomaly level 3.

[0090] As an alternative implementation, the temperature status can be displayed in the following way: for each target area of the target tire, determine the highest temperature anomaly level corresponding to each temperature anomaly area in the target area, where the target area is the tread area or the sidewall area; determine the first color corresponding to the highest temperature anomaly level from the preset temperature level color table, and display the temperature status of the target area based on the first color, and add temperature anomaly warning information, where the temperature anomaly warning information includes at least one of the following: the position information of the target tire, the position information and area information of each temperature anomaly area, and the highest abnormal temperature.

[0091] To facilitate the display of the temperature status, the embodiments of the present application also provide a preset temperature level color table, as shown in Table 2 specifically.

[0092] Table 2

[0093] Normal temperature Green Temperature anomaly level 1 Yellow Temperature anomaly level 2 Orange Temperature anomaly level 3 Red

[0094] Specifically, in all abnormal areas of the target area, select the area with the highest temperature anomaly level as the highest temperature anomaly level of the target area, determine the first color corresponding to the highest temperature anomaly level of the target area from the preset temperature level color table, and display the temperature status of the target area based on the first color. In the visual display of the target area, mark each temperature anomaly area and use different colors to represent it according to its temperature anomaly level to ensure the clear visibility of the abnormal information. In the case of temperature anomaly, temperature anomaly warning information can be added to the mark. The structure of the warning information includes but is not limited to the wheel position information of the target tire, the position coordinates, area, and the value of the abnormal temperature of the abnormal area. The above information can be used to fill the structure of the warning information, and finally it is displayed on the front-end interface. The display of color marks and warning information can intuitively display information such as the position, area, and temperature level of the abnormal area, which is convenient for mine operators to quickly understand and respond.

[0095] In order to more comprehensively display the temperature state distribution of each sub-region in the target area, an optional method is also provided for displaying the temperature state: generate two-dimensional temperature display maps of multiple target tires based on the target thermal imaging image, the position information of each target tire on the target vehicle, and the position information of the target area of each target tire, where the target area is the tread area or the sidewall area; for the target area of each target tire, determine the temperature anomaly level of each temperature anomaly area in the target area and the second color corresponding to the normal temperature from a preset temperature grade color table, and update the color at the corresponding position in the two-dimensional temperature display map according to the position information of each temperature anomaly area and the temperature normal area and the corresponding second color.

[0096] Specifically, each target area is composed of each sub-region. Whether the sub-region is a temperature anomaly area or a temperature normal area, a color corresponding to the current sub-region temperature situation can be selected from the preset temperature grade color table for temperature state display. The display interface is connected to the temperature detection module in real time. Once a change in the temperature state is detected, the color information in the two-dimensional temperature display map is immediately updated to ensure that the mine operation personnel can receive the latest temperature warning information in a timely manner.

[0097] Figure 3 The flowchart of a more complete temperature detection method for mine engineering tires is shown. Next, in combination with Figure 3 the temperature detection method for mine engineering tires will be further described. Figure 3 The process can be specifically summarized into the following steps:

[0098] Step S1, obtain a thermal imaging image.

[0099] The system can first obtain the infrared thermal imaging image of the mine vehicle from the infrared camera deployed in the mine environment. These images carry the specific values of the tire surface temperature.

[0100] Step S2, input the thermal imaging image into the target detection model to determine whether a target is detected.

[0101] Specifically, when the thermal imaging image is input into the target detection model, it is mainly used to detect the target, and the target includes: the vehicle, the side-by-side tires, and the target tires in the side-by-side tires. If the above targets are not detected, it means that there may be deviations in the detection process and the target tires in the side-by-side tires of the vehicle are not correctly identified, and the detection steps need to be repeated. After the target is normally detected, step S3 can be entered.

[0102] Step S3, tire positioning.

[0103] The tire positioning process mainly analyzes the relative relationships among the first detection frame corresponding to the vehicle, the second detection frame corresponding to the side-by-side tires, and the third detection frame corresponding to the target tire among the side-by-side tires to obtain the position information of the final target tire.

[0104] Step S4: Use the pre-trained object segmentation model to divide the tread and the sidewall.

[0105] Specifically, the tire region image including the target tire can be intercepted from the target thermal imaging image according to the position information of the target tire, and the pre-trained object segmentation model is used to divide the tire region image into a tread region image and a sidewall region image. After the segmentation by the segmentation model, it is detected whether the tread and sidewall region images are normally segmented. If the tread and sidewall images cannot be normally segmented, it indicates that there is a problem in the target detection, and it is necessary to go back to step S2 to re-detect the thermal imaging image.

[0106] Step S5: Use the temperature detection module to detect the target region image.

[0107] The target region image includes the tread region image and the sidewall region image. The temperature detection module is used to detect the temperature, distinguish the regions with normal temperature and abnormal temperature, and determine the abnormal level for the regions with abnormal temperature.

[0108] Step S6: Display the temperature status.

[0109] First, it is judged whether there is any abnormality in the temperature of the target region. If there is no abnormality, the corresponding region is displayed in green; if there is an abnormality in the temperature of the target region, the abnormal temperature can be classified according to the preset temperature level table. When the abnormal temperature level of the abnormal temperature region is level 1, it is correspondingly displayed in yellow; when the abnormal temperature level of the abnormal temperature region is level 2, it is correspondingly displayed in orange; when the abnormal temperature level of the abnormal temperature region is level 3, it is correspondingly displayed in red. When there is an abnormal temperature region in the target region, not only can the temperature status of the corresponding region be displayed by colors, but also warning information can be displayed. The warning information includes information such as the tire position, the abnormal region level, the position and the area, etc., so as to facilitate the mine operation personnel to intuitively understand the temperature distribution and the abnormal situation.

[0110] In the above steps, the collected thermal imaging images are detected, reducing the upfront equipment investment and subsequent maintenance expenses. The detected images are divided using a target segmentation model, and the temperature of the divided regions is detected and statistically analyzed respectively. Finally, the temperature distribution of the target tire is displayed in combination with the temperature status map, which can effectively reduce the problem of inaccurate temperature measurement caused by the excessive influence of the temperature difference between the tread and the sidewall. The display of the results can visually show the current temperature status to the operators, facilitating the timely identification of tires with potential risks and the avoidance of dangerous situations.

[0111] Embodiment 2

[0112] According to an embodiment of the present application, there is also provided a temperature detection device for a mine engineering tire for implementing the temperature detection method of the mine engineering tire in Embodiment 1, as Figure 4 shown. The temperature detection device for the mine engineering tire includes at least: a tire positioning module 41, a segmentation module 42, and a temperature detection module 43, where:

[0113] The tire positioning module 41 can obtain the target thermal imaging image of the mine target area and use a pre-trained target detection model to determine the position information of each target tire on the target vehicle in the target thermal imaging image;

[0114] The segmentation module 42, for each target tire, intercepts the tire area image including the target tire from the target thermal imaging image according to the position information of the target tire, and uses a pre-trained target segmentation model to divide the tire area image into a tread area image and a sidewall area image. Among them, the target segmentation model includes an efficient channel attention module, and the target segmentation model is trained based on an enhanced intersection over union loss function;

[0115] The temperature detection module 43 statistically analyzes the temperature information in the tread area image and the sidewall area image respectively, determines the temperature status of the tread and the sidewall of the target tire according to the statistical results, and displays the temperature status, where the temperature status is used to reflect whether there is a temperature anomaly in the tread or the sidewall and the corresponding temperature anomaly area when there is an anomaly.

[0116] The functions of each module of the temperature detection device for the mine engineering tire will be described below in combination with the specific implementation process.

[0117] As an alternative implementation, when the tire positioning module determines the position information of each target tire on the target vehicle in the target thermal imaging image by using a pre-trained object detection model, it can be achieved in the following way: Analyze the target thermal imaging image by using the object detection model to determine the first detection box corresponding to the target vehicle in the target thermal imaging image, the second detection box corresponding to each group of side-by-side tires on the target vehicle, and the third detection box corresponding to each target tire in each group of side-by-side tires; Determine the position information of each target tire based on the first detection box, the second detection box, and the third detection box, where the position information includes: wheel position information and position coordinate information.

[0118] Optionally, the tire positioning module determines the position information of each target tire based on the first detection box, the second detection box, and the third detection box, and specifically can be achieved in the following way: For each target tire, determine the position coordinate information of the target tire based on the third detection box corresponding to the target tire; Determine the first wheel position information of the target tire in the side-by-side tires to which it belongs based on the positional relationship between the third detection box corresponding to the target tire and the second detection box to which the third detection box belongs; Determine the second wheel position information of the side-by-side tires to which the target tire belongs on the target vehicle based on the positional relationship between the second detection box to which the third detection box corresponding to the target tire belongs and the first detection box.

[0119] After the tire positioning module determines the position information of each target tire on the target vehicle in the thermal imaging image, the segmentation module can intercept the tire area image including the target tire from the target thermal imaging image according to the position information of the target tire. After obtaining the tire area image, in order to facilitate more accurate detection of the temperature of the target tire in the later stage, the segmentation module needs to use a pre-trained object segmentation model to divide the tire area image into a tread area image and a sidewall area image.

[0120] As an alternative implementation, the training process of the object segmentation model can be achieved in the following way: Construct an initial segmentation model, where the initial segmentation model is a YOLOv8 segmentation model with an efficient channel attention module added; Obtain a plurality of historical thermal imaging images collected on the mine and including vehicle tires, and intercept at least one sub-region image including a single tire from each historical thermal imaging image; Use each sub-region image as a training sample, and label the tread area and the sidewall area in the sub-region image as sample labels; Divide the obtained multiple training samples into a training set, a validation set, and a test set according to a preset ratio; Use the training set to perform iterative training on the initial segmentation model, use the validation set to verify the model performance after each training cycle, use the trained model as the object segmentation model, and use the test set to test the performance of the object segmentation model.

[0121] To enhance the robustness and generalization ability of the model, optionally, before dividing the obtained multiple training samples into a training set, a validation set, and a test set according to a preset ratio, the multiple training samples can also be processed by sample data augmentation. The ways of sample data augmentation include at least one of the following: random rotation, random scaling, and random flipping.

[0122] As an alternative implementation, when iteratively training the initial segmentation model using the training set, preset hyperparameters can be obtained and the model can be trained based on the hyperparameters. The hyperparameters include at least one of the following: the number of training epochs, batch size, learning rate, and optimizer. During the training process, multiple training samples corresponding to each training batch are input into the model, and an enhanced intersection over union (IoU) loss function is constructed based on the overlapping region, center point difference, width difference, and height difference between the predicted bounding boxes output by the model and the annotated bounding boxes in the sample labels. The model parameters are adjusted based on the enhanced IoU loss function.

[0123] As an alternative implementation, after the segmentation module successfully divides the tread area image and the sidewall area image, the temperature detection module is used to detect the tread area image and the sidewall area image respectively, and the temperature information in the tread area image and the sidewall area image is statistically analyzed. The temperature states of the tread and sidewall of the target tire are determined based on the statistical results. Specifically, it can be achieved through the following steps: for the target area image, determine the temperature value of each pixel in the target area image, where the target area image is the tread area image or the sidewall area image; determine the mode of all temperature values, and compare the temperature value of each pixel with the mode. Determine the pixels with a first difference between the temperature value and the mode greater than the preset temperature threshold as temperature abnormal pixels; connect adjacent temperature abnormal pixels to obtain multiple connected regions, and determine the connected regions with an area greater than the preset area threshold as the temperature abnormal regions in the target area image.

[0124] Optionally, after the temperature detection module determines the temperature abnormal regions in the target area image, it can also classify the temperature abnormal regions into different abnormal levels. The specific classification process can be achieved through the following steps: for each temperature abnormal region in the target area image, determine the average temperature of all pixels in the temperature abnormal region, and determine the second difference between the average value and the mode; determine the target temperature range corresponding to the second difference from the preset temperature abnormal level table, and use the temperature abnormal level corresponding to the target temperature range as the temperature abnormal level of the temperature abnormal region. The temperature abnormal level table stores the mapping relationship between multiple temperature ranges and multiple temperature abnormal levels.

[0125] Optionally, the above device may further include a display module, and the display module can be used to display the temperature state.

[0126] As an alternative embodiment, the display module can display the temperature status in the following manner: for each target area of each target tire, determine the highest temperature anomaly level corresponding to each temperature anomaly area in the target area, where the target area is the tread area or the sidewall area; determine the first color corresponding to the highest temperature anomaly level from a preset temperature level color table, display the temperature status of the target area based on the first color, and add temperature anomaly warning information, where the temperature anomaly warning information includes at least one of the following: the position information of the target tire, the position information and area information of each temperature anomaly area, and the highest anomaly temperature.

[0127] Since the tread area or the sidewall area in the target area can be composed of multiple sub-areas, in order to further comprehensively display the temperature status of each sub-area on the tire, an alternative embodiment is also provided for displaying the temperature status, which can be specifically implemented in the following manner: generate a two-dimensional temperature display map of multiple target tires based on the target thermal imaging image, the position information of each target tire on the target vehicle, and the position information of the target area of each target tire, where the target area is the tread area or the sidewall area; for the target area of each target tire, determine the temperature anomaly level of each temperature anomaly area in the target area and the second color corresponding to the normal temperature from a preset temperature level color table, and update the color of the corresponding position in the two-dimensional temperature display map according to the position information and the corresponding second color of each temperature anomaly area and the temperature normal area.

[0128] It should be noted that each module in the temperature detection device of the mine engineering tire in the embodiments of the present application corresponds one by one to each implementation step of the temperature detection method of the mine engineering tire in Embodiment 1. Since detailed descriptions have been made in Embodiment 1, some details not shown in this embodiment can be referred to Embodiment 1 and will not be elaborated here.

[0129] Embodiment 3

[0130] According to the embodiments of the present application, a computer program product is also provided. The computer program product includes a computer program, where when the computer program is executed by a processor, it implements the temperature detection method of the mine engineering tire in Embodiment 1.

[0131] According to the embodiments of the present application, a non-volatile storage medium is also provided. The non-volatile storage medium includes a stored computer program, where the device where the non-volatile storage medium is located executes the temperature detection method of the mine engineering tire in Embodiment 1 by running the computer program.

[0132] According to the embodiments of the present application, a processor is also provided. The processor is used to run a computer program, where when the computer program runs, it executes the temperature detection method of the mine engineering tire in Embodiment 1.

[0133] According to an embodiment of the present application, an electronic device is further provided. The electronic device includes: a memory and a processor. Among them, a computer program is stored in the memory, and the processor is configured to execute the temperature detection method of the mine engineering tire in Embodiment 1 through the computer program.

[0134] Specifically, when the computer program runs, it executes the following steps: obtaining a target thermal imaging image of a mine target area, and using a pre-trained target detection model to determine the position information of each target tire on the target vehicle in the target thermal imaging image; for each target tire, intercepting a tire area image including the target tire from the target thermal imaging image according to the position information of the target tire, and using a pre-trained target segmentation model to divide the tire area image into a tread area image and a sidewall area image. Among them, the target segmentation model includes an efficient channel attention module, and the target segmentation model is trained based on an enhanced intersection over union loss function; respectively statistically analyzing the temperature information in the tread area image and the sidewall area image, determining the temperature states of the tread and the sidewall of the target tire according to the statistical results, and displaying the temperature states, where the temperature states are used to reflect whether there is a temperature anomaly in the tread or the sidewall and the corresponding temperature anomaly area when there is an anomaly.

[0135] As an optional implementation manner, the above electronic device may exist in the form of a mobile terminal, a computer terminal, or a similar computing device. Figure 5 The hardware structure block diagram of an electronic device for implementing the temperature detection method of the mine engineering tire is shown. As Figure 5 shown, the electronic device 50 may include one or more processors 502 (shown as 502a, 502b,..., 502n in the figure) (the processor 502 may include, but is not limited to, a processing device such as a microprocessor MCU or a field programmable gate array FPGA), a memory 504 for storing data, and a transmission device 506 for communication functions. In addition, it may further include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply, and / or a camera. Those of ordinary skill in the art can understand that Figure 5 the structure shown is only schematic and does not limit the structure of the above electronic device. For example, the electronic device 50 may further include more or fewer components than Figure 5 shown, or have a different configuration from Figure 5 shown.

[0136] It should be noted that one or more of the above-mentioned processors 502 and / or other data processing circuits can generally be referred to as "data processing circuits" herein. The data processing circuit can be embodied in software, hardware, firmware, or any combination thereof, in whole or in part. In addition, the data processing circuit can be a single independent processing module, or be incorporated in whole or in part into any one of the other components in the electronic device 50. As involved in the embodiments of the present application, the data processing circuit is a kind of processor control (such as the selection of a variable resistance terminal path connected to an interface).

[0137] The memory 504 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the temperature detection method of mine engineering tires in the embodiments of the present application. The processor 502 executes various functional applications and data processing by running the software programs and modules stored in the memory 504, that is, implements the vulnerability detection method of the above-mentioned application program. The memory 504 can include a high-speed random access memory, and can also include a non-volatile memory, such as one or more magnetic storage devices, flash memories, or other non-volatile solid-state memories. In some instances, the memory 504 can further include a memory remotely set relative to the processor 502, and these remote memories can be connected to the electronic device 50 through a network. Examples of the above-mentioned network include but are not limited to the Internet, intranet, local area network, mobile communication network, and combinations thereof.

[0138] The transmission device 506 is used to receive or send data via a network. Specific examples of the above-mentioned network can include the wireless network provided by the communication provider of the electronic device 50. In one instance, the transmission device 506 includes a network interface controller (NIC), which can be connected to other network devices through a base station and thus can communicate with the Internet. In one instance, the transmission device 506 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0139] The display can be, for example, a touch-screen liquid crystal display (LCD), which enables the user to interact with the user interface of the electronic device 50.

[0140] The above-mentioned embodiment numbers are only for description and do not represent the advantages or disadvantages of the embodiments.

[0141] In the above embodiments of the present application, the descriptions of each embodiment have their own focuses. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0142] In several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are merely illustrative. For example, the division of units can be a logical function division. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the couplings or direct couplings or communication connections shown or discussed with each other can be through some interfaces, and the indirect couplings or communication connections of units or modules can be in electrical or other forms.

[0143] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they can be located in one place or distributed to multiple units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0144] In addition, in each embodiment of this application, the functional units can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0145] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in each embodiment of this application. The foregoing storage medium includes: various media such as USB flash drives, read-only memories (ROMs), random access memories (RAMs), mobile hard disks, magnetic disks, or optical discs that can store program codes.

[0146] The above is only the preferred embodiment of this application. It should be noted that for those of ordinary skill in the art, without departing from the principle of this application, several improvements and refinements can still be made, and these improvements and refinements should also be regarded as the protection scope of this application.

Claims

1. A temperature detection method for mining engineering tires, characterized in that: include: Acquire a target thermal imaging image of a target area of ​​a mine, and determine position information of each target tire on a target vehicle in the target thermal imaging image using a pre-trained target detection model; For each target tire, a tire region image including the target tire is intercepted from the target thermal imaging image according to the position information of the target tire, and the tire region image is divided into a tread region image and a sidewall region image using a pre-trained target segmentation model, wherein the target segmentation model includes an efficient channel attention module, and the target segmentation model is trained based on an enhanced intersection-over-union loss function; The temperature information in the tread area image and the sidewall area image are statistically analyzed respectively, and the temperature status of the tread and sidewall of the target tire is determined based on the statistical results, and the temperature status is displayed, wherein the temperature status is used to reflect whether there is a temperature abnormality on the tread or sidewall and the corresponding temperature abnormality area when an abnormality occurs.

2. The method according to claim 1, characterized in that Determining the position information of each target tire on the target vehicle in the target thermal imaging image using a pre-trained target detection model includes: Analyzing the target thermal imaging image using the target detection model to determine a first detection frame corresponding to the target vehicle in the target thermal imaging image, a second detection frame corresponding to each group of side-by-side tires on the target vehicle, and a third detection frame corresponding to each target tire in each group of side-by-side tires; The position information of each target tire is determined according to the first detection frame, the second detection frame and the third detection frame, wherein the position information includes: wheel position information and position coordinate information.

3. The method according to claim 2, characterized in that Determining the position information of each target tire according to the first detection frame, the second detection frame, and the third detection frame includes: For each target tire, determining the position coordinate information of the target tire according to the third detection frame corresponding to the target tire; Determining first wheel position information of the target tire among the parallel tires to which it belongs according to a positional relationship between a third detection frame corresponding to the target tire and a second detection frame to which the third detection frame belongs; The second wheel position information of the side-by-side tires to which the target tire belongs on the target vehicle is determined according to the positional relationship between the first detection frame and the second detection frame to which the third detection frame corresponding to the target tire belongs.

4. The method according to claim 1, characterized in that: The training process of the target segmentation model includes: Constructing an initial segmentation model, wherein the initial segmentation model is a YOLOv8 segmentation model with an efficient channel attention module added; Acquire a plurality of historical thermal imaging images including vehicle tires collected in the mine, and intercept at least one sub-region image including a single tire from each of the historical thermal imaging images; Taking each of the sub-region images as a training sample, and marking the tread region and the sidewall region in the sub-region image as sample labels; Divide the obtained multiple training samples into a training set, a validation set, and a test set according to a preset ratio; The initial segmentation model is iteratively trained using the training set, the model performance after each training cycle is verified using the verification set, the trained model is used as the target segmentation model, and the performance of the target segmentation model is tested using the test set.

5. The method according to claim 4, characterized in that Before dividing the obtained multiple training samples into a training set, a validation set and a test set according to a preset ratio, the method further includes: Performing sample data enhancement processing on the plurality of training samples, wherein the sample data enhancement processing comprises at least one of the following methods: random rotation, random scaling, and random flipping.

6. The method according to claim 4, characterized in that Iteratively training the initial segmentation model using the training set includes: Obtaining preset hyperparameters, and performing model training according to the hyperparameters, wherein the hyperparameters include at least one of the following: number of training cycles, batch size, learning rate, and optimizer; During the training process, multiple training samples corresponding to each training batch are input into the model, and an enhanced intersection-over-union loss function is constructed based on the overlapping area, center point difference, width difference and height difference between the prediction box output by the model and the annotation box in the sample label, and the model parameters are adjusted based on the enhanced intersection-over-union loss function.

7. The method according to claim 1, characterized in that The temperature information in the tread region image and the sidewall region image are respectively counted, and the temperature state of the tread and the sidewall of the target tire is determined according to the statistical results, including: For a target area image, determining a temperature value of each pixel in the target area image, wherein the target area image is the tread area image or the sidewall area image; Determine the mode of all temperature values, and compare the temperature value of each pixel with the mode, and determine that a pixel whose first difference between the temperature value and the mode is greater than a preset temperature threshold is a temperature abnormality pixel; Adjacent temperature abnormality pixels are connected to obtain a plurality of connected domains, and a connected domain having an area greater than a preset area threshold is determined as a temperature abnormality region in the target region image.

8. The method according to claim 7, characterized in that The method further comprises: For each temperature abnormality region in the target region image, determining an average value of the temperatures of all pixels in the temperature abnormality region, and determining a second difference between the average value and the mode; A target temperature interval corresponding to the second difference is determined from a preset temperature anomaly level table, and the temperature anomaly level corresponding to the target temperature interval is used as the temperature anomaly level of the temperature anomaly area, wherein the temperature anomaly level table stores a mapping relationship between multiple temperature intervals and multiple temperature anomaly levels.

9. The method according to claim 8, characterized in that Displaying the temperature status includes: For a target area of ​​each target tire, determining a maximum temperature abnormality level corresponding to each temperature abnormality area in the target area, wherein the target area is a tread area or a sidewall area; Determine a first color corresponding to the highest temperature abnormality level from a preset temperature level color table, display the temperature status of the target area based on the first color, and add temperature abnormality alarm information, wherein the temperature abnormality alarm information includes at least one of the following: position information of the target tire, position information and area information of each temperature abnormality area, and the highest abnormal temperature.

10. The method according to claim 8, characterized in that Displaying the temperature status includes: Generate a two-dimensional temperature display diagram of multiple target tires based on the target thermal imaging image, the position information of each target tire on the target vehicle, and the position information of the target area of ​​each target tire, wherein the target area is a tread area or a sidewall area; For the target area of ​​each target tire, the temperature abnormality level of each temperature abnormality area in the target area and the second color corresponding to the normal temperature are determined from a preset temperature level color table, and the color of the corresponding position in the two-dimensional temperature display diagram is updated according to the position information of each temperature abnormality area and the normal temperature area and the corresponding second color.

11. A temperature detection device for mining engineering tires, characterized in that: include: A tire positioning module is used to obtain a target thermal imaging image of a target area of ​​a mine, and determine the position information of each target tire on a target vehicle in the target thermal imaging image using a pre-trained target detection model; A segmentation module, for each target tire, according to the position information of the target tire, intercepting a tire region image including the target tire from the target thermal imaging image, and dividing the tire region image into a tread region image and a sidewall region image using a pre-trained target segmentation model, wherein the target segmentation model includes an efficient channel attention module, and the target segmentation model is trained based on an enhanced intersection-over-union loss function; The temperature detection module is used to collect statistics on the temperature information in the tread area image and the sidewall area image respectively, determine the temperature status of the tread and sidewall of the target tire based on the statistical results, and display the temperature status, wherein the temperature status is used to reflect whether there is a temperature abnormality on the tread or sidewall and the corresponding temperature abnormality area when the abnormality occurs.

12. A computer program product, characterized in that include: A computer program, wherein when the computer program is executed by a processor, the temperature detection method of the mining engineering tire according to any one of claims 1 to 10 is implemented.

13. An electronic device, characterized in that: include: A memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the temperature detection method of the mining engineering tire according to any one of claims 1 to 10 through the computer program.

Images