A method and apparatus for evaluating the loading quality of a mine car

Abstract

The application discloses a mine car loading quality evaluation method and device, a storage medium and an electronic device. The method comprises the following steps: performing mine car feature recognition processing on a target image to be evaluated to obtain a plurality of first images containing mine car features; performing gap feature recognition processing on a box region and a cargo region in the first image to obtain a second image not containing a gap feature; performing image segmentation processing on the second image to obtain a target image corresponding to the second image; and evaluating the loading quality of the mine car based on the target image to obtain an evaluation result. The mine car loading quality evaluation method can avoid the problems of low efficiency and inaccuracy of traditional manual image observation for loading quality.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and in particular to a method, apparatus, storage medium and electronic device for assessing the loading quality of mining trucks. Background Technology

[0002] Identifying whether the loading quality of mining trucks is up to standard primarily involves determining whether the amount of cargo loaded on each truck meets the required standards. In mining operations, the efficiency of cargo transportation by trucks is crucial, and improving efficiency requires ensuring that each truck carries the required amount of cargo. Currently, the technical methods used in mining areas to determine whether the loading amount of cargo on a truck is up to standard include: Weighbridge weighing: This method involves unloading the cargo from the truck and weighing it on a weighbridge. Manual judgment: This method uses image data collected by cameras and visual inspection to determine whether the load is up to standard. Weighbridge weighing is costly, inefficient, and inconvenient to move. Therefore, mining areas rarely use weighbridge weighing to determine whether the loading is up to standard; most judgment is done manually, mainly by reviewing image data acquired by cameras. However, manual judgment leads to a lack of objectivity in the identification results and also does not align with the direction of intelligent mining development. Summary of the Invention

[0003] In view of this, the present invention provides a method, apparatus, storage medium and electronic device for assessing the loading quality of mine cars, the main purpose of which is to solve the problem that the current method of assessing the loading quality of mine cars in open-pit mines, which relies on manual judgment and image data collected by cameras, is inefficient and inaccurate in terms of human visual observation of whether the data is qualified.

[0004] To address the above problems, this application provides a method for assessing the loading quality of mining trucks, comprising:

[0005] The target image to be evaluated is processed by mine car feature recognition to obtain several first images containing mine car features;

[0006] Perform gap feature recognition processing on the cargo box area and cargo area in the first image to obtain a second image that does not contain gap features;

[0007] The second image is segmented to obtain a target image corresponding to the second image;

[0008] The loading quality of the mining truck is evaluated based on the target image, and the evaluation results are obtained.

[0009] Optionally, before performing mine car feature recognition on the target image to be evaluated to obtain several first images containing mine car features, the method further includes: acquiring each of the images to be evaluated, specifically including:

[0010] Image acquisition was performed on the target mining area to obtain several raw images;

[0011] Each of the original acquired images is filtered to obtain a first original acquired image whose clarity meets the preset conditions, and a second original acquired image whose clarity does not meet the preset conditions.

[0012] Based on the image attributes of each second original acquired image, the second original acquired image is processed using an image enhancement processing method corresponding to the image attributes to obtain the first enhanced image after image enhancement;

[0013] The first original acquired image and the first enhanced image are combined to obtain the target image to be evaluated.

[0014] Optionally, the step of processing the second original acquired images based on their image attributes using an image enhancement method corresponding to those attributes to obtain the enhanced first image specifically includes:

[0015] The histogram equalization algorithm is used to enhance the images in the second original acquired image that do not meet the preset sharpness conditions and whose image attributes are daytime acquired images, to obtain daytime enhanced images;

[0016] The Laplacian operator algorithm is used to perform image enhancement processing on the images in the second original acquired image that do not meet the preset sharpness conditions in the night acquired image, so as to obtain the night enhanced image;

[0017] The enhanced daytime image and the enhanced nighttime image are combined to obtain the first enhanced image.

[0018] Optionally, the method further includes:

[0019] The pre-acquired sample images are processed to obtain a first target sample image containing the features of a mining truck;

[0020] The first target sample image is annotated to obtain first annotated images with annotation information;

[0021] Based on each of the first labeled images, the initial first image recognition model is trained to obtain the target first image recognition model;

[0022] The step of performing mine truck feature recognition on the target image to be evaluated to obtain several first images containing mine truck features specifically includes:

[0023] Based on the target first image recognition model, the mine car features of each target image to be evaluated are identified to obtain a first image containing mine car features.

[0024] Optionally, the method further includes:

[0025] The first target sample image is processed to obtain a second target sample image containing gap features and a third target sample image not containing gap features;

[0026] The second target sample image is annotated to obtain each second annotated image with annotation information;

[0027] The initial second image recognition model is trained based on each of the second labeled images to obtain the target second image recognition model;

[0028] Perform gap feature recognition on the cargo box area and cargo area in the first image to obtain a second image that does not contain gap features, specifically including:

[0029] Based on the target second image recognition model, gap feature recognition is performed on each of the first images to obtain a second image that does not contain gap features.

[0030] Optionally, the method further includes:

[0031] The pre-acquired third target sample image is subjected to image segmentation processing to obtain a label image corresponding to the third target sample image, which includes cargo area features and container area features;

[0032] The initial segmentation model is trained based on the third target sample image and the label image to obtain the target segmentation model;

[0033] The step of performing image segmentation processing on the second image to obtain the target image corresponding to the second image specifically includes:

[0034] Based on the target segmentation model, each of the second images is segmented to obtain the target image corresponding to the second image.

[0035] Optionally, the step of evaluating the loading quality of the mining truck based on the target image to obtain the evaluation result specifically includes:

[0036] The loading quality assessment value is calculated based on the pixel size of the cargo area and the pixel size of the container area in the target image.

[0037] The loading quality assessment result of the mining car is obtained based on the assessment value of the loading quality.

[0038] To address the aforementioned problems, this application provides a mine car loading quality assessment device, comprising:

[0039] Mining truck feature recognition module: used to recognize mining truck features in the target image to be evaluated, and obtain several first images containing mining truck features;

[0040] Gap feature recognition module: used to perform gap feature recognition on the cargo box area and cargo area in the first image, and obtain a second image that does not contain gap features;

[0041] Image segmentation module: used to perform image segmentation processing on the second image to obtain a target image corresponding to the second image;

[0042] Evaluation module: Used to evaluate the loading quality of the mining truck based on the target image and obtain the evaluation result.

[0043] To address the aforementioned problems, this application provides a storage medium storing a computer program that, when executed by a processor, implements the steps of the aforementioned mine car loading quality assessment method.

[0044] To address the aforementioned problems, this application provides an electronic device, comprising at least a memory and a processor. The memory stores a computer program, and the processor, when executing the computer program in the memory, implements the steps of the aforementioned mine car loading quality assessment method.

[0045] This application constructs an image recognition model to identify gaps between the mine car and the cargo and the cargo box, and constructs an image segmentation model to segment the cargo area and cargo box area of ​​the mine car. The loading quality assessment value is calculated based on the number of pixels in the cargo area and cargo box area to obtain the assessment result of the loading quality of the mine car, which can reduce the assessment cost and make the assessment result more accurate.

[0046] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0047] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0048] Figure 1 This is a flowchart illustrating a method for assessing the loading quality of a mining truck, as described in an embodiment of this application.

[0049] Figure 2This is a flowchart of a method for assessing the loading quality of a mining truck, which is another embodiment of this application. Detailed Implementation

[0050] Various embodiments and features of this application are described herein with reference to the accompanying drawings.

[0051] It should be understood that various modifications can be made to the embodiments described herein. Therefore, the above description should not be considered as limiting, but merely as an example of embodiments. Other modifications within the scope and spirit of this application will be apparent to those skilled in the art.

[0052] The accompanying drawings, which are included in and form part of this specification, illustrate embodiments of the present application and, together with the general description of the present application given above and the detailed description of the embodiments given below, serve to explain the principles of the present application.

[0053] These and other features of this application will become apparent from the following description of preferred forms of embodiments given as non-limiting examples, with reference to the accompanying drawings.

[0054] It should also be understood that although this application has been described with reference to some specific examples, those skilled in the art can certainly implement many other equivalent forms of this application.

[0055] The above and other aspects, features and advantages of this application will become more apparent when taken in conjunction with the accompanying drawings and in view of the following detailed description.

[0056] Specific embodiments of this application are described thereafter with reference to the accompanying drawings; however, it should be understood that the claimed embodiments are merely examples of this application, which can be implemented in various ways. Well-known and / or repeated functions and structures are not described in detail to avoid unnecessary or redundant details that could obscure the application. Therefore, the specific structural and functional details claimed herein are not intended to be limiting, but merely serve as the basis and representative basis for the claims to teach those skilled in the art to use this application in a variety of substantially any suitable detailed structures.

[0057] This specification may use the phrases “in one embodiment,” “in another embodiment,” “in yet another embodiment,” or “in other embodiments,” all of which may refer to one or more of the same or different embodiments according to this application.

[0058] This application provides a method for assessing the loading quality of mining trucks, such as... Figure 1 As shown, it includes the following steps:

[0059] Step S101: Perform mine car feature recognition on the target image to be evaluated to obtain several first images containing mine car features;

[0060] In this step, images of the target mining area are captured using a camera, resulting in several original images. These original images are then filtered to obtain a first original image whose clarity meets a preset condition, and a second original image whose clarity does not meet the preset condition. A histogram equalization algorithm is then used to enhance the images in the original images acquired during the day that do not meet the preset clarity condition, resulting in a daytime enhanced image. Similarly, a Laplacian operator algorithm is used to enhance the images in the second original images acquired at night that do not meet the preset clarity condition, resulting in a nighttime enhanced image. The daytime and nighttime enhanced images are combined to form the first enhanced image. By enhancing the unclear data, the image contrast is improved, resulting in richer edge details, which helps the convolutional neural network extract more useful features. The first original image meeting the clarity requirement and the first enhanced image constitute the target image to be evaluated. Then, a first image recognition model is constructed to identify mining vehicles in the target image to be evaluated, resulting in the first image containing mining vehicle features.

[0061] Step S102: Perform gap feature recognition on the cargo box area and cargo area in the first image to obtain a second image that does not contain gap features.

[0062] In the specific implementation process, this step mainly involves constructing a second image recognition model. This model identifies the gap between the cargo box area and the cargo area in the first image. Images containing gap features are filtered out and deleted, as the presence of gaps indicates that the loading quality of the mining car is unqualified. At the same time, a second image without gap features is identified. The loading quality of the second image without gaps identified in this step is not necessarily qualified either, thus laying the foundation for subsequent segmentation based on the second image.

[0063] Step S103: Perform image segmentation processing on the second image to obtain the target image corresponding to the second image;

[0064] In this step, a target segmentation model is constructed. This model primarily uses the semantic segmentation model DeepLabv3 for image segmentation. The cargo and container regions of the second image are semantically segmented to obtain a target image containing three distinct pixel regions: the cargo region, the container region, and the background region. This lays the foundation for subsequent loading quality assessment based on the target image.

[0065] Step S104: Evaluate the loading quality of the mining truck based on the target image to obtain the evaluation result.

[0066] In this step, the loading quality assessment value is calculated based on the number of pixels in the cargo area and the container area of ​​the target image. The calculation process for the loading quality assessment value is as follows: First, the sum of the number of pixels in the container area and the number of pixels in the cargo area is calculated to obtain the total number of pixels in the container area and the cargo area; then, the ratio of the number of pixels in the cargo area to the total number of pixels in the container area and the cargo area is calculated to obtain the loading quality assessment value; finally, the loading quality of the mining car is assessed based on the loading quality assessment value to obtain the assessment result.

[0067] This application constructs an image recognition model to identify gaps between the mine car and the cargo and the cargo box, and constructs an image segmentation model to segment the cargo area and cargo box area of ​​the mine car. The loading quality assessment value is calculated based on the number of pixels in the cargo area and cargo box area to obtain the assessment result of the loading quality of the mine car, which can reduce the assessment cost and make the assessment result more accurate.

[0068] Another embodiment of this application provides a method for assessing the loading quality of a mining truck, such as... Figure 2 As shown, it includes the following steps:

[0069] Step S201: Obtain a first target sample image, and train an initial first image recognition model based on the first target sample image to obtain a target first image recognition model;

[0070] In this step, the pre-acquired sample images are first processed using a target detection method. The YOLOv5 target detection algorithm can be used to detect the sample images. Specifically, target recognition processing is performed on the mining trucks in the sample images to obtain first target sample images containing mining truck features. Images where mining trucks cannot be identified are filtered out. Mining trucks in the first target sample images are selected and labeled, resulting in first labeled images with label information. The label category can be set to "truck". Each labeled image corresponds to a tag file. The tag file format includes the mining truck type, the normalized coordinates of the mining truck's center point, and the normalized width and height of the mining truck's center point. For example, the tag file format is: id xywh, where id represents the type. When there is only one mining truck type, id is 0; when there are 10 mining truck types, id is ordered from 0 to 9. x and y are the normalized coordinates of the mining truck's center point in the YOLOv5 target detection software, and w and h are the normalized width and height of the mining truck's center point in the YOLOv5 target detection software. The initial first image recognition model is trained based on the first labeled image to obtain the first label information corresponding to the first labeled image. The parameters of the initial first image recognition model are adjusted based on each first labeled image and the first label information. After continuous iteration until the model converges, the target first image recognition model is obtained.

[0071] Step S202: Perform gap recognition processing on the first target sample image to obtain a second target sample image containing gap features and a third target sample image not containing gap features. Train the initial second image recognition model on the second target sample image to obtain the target second image recognition model.

[0072] In this step, the first step involves using a target detection method to identify the first target sample image. The YOLOv5 target detection algorithm can be used. Specifically, target recognition processing is performed on the gaps between the cargo box area and the cargo area in the first target sample image, resulting in a second target sample image containing gap features and a third target sample image not containing gap features. The gaps between the cargo box area and the cargo area in the second target sample image are then selected and labeled. To reduce the error rate of gap recognition, cargo box corner information is added to the labeled area. This results in second labeled images with labeled information. The label category can be set to "crack". Each labeled image corresponds to a tag file, and the tag file format includes the gap area type, the normalized coordinates of the gap area center point, and the normalized width and height of the gap area center point. The initial second image recognition model is trained based on the second labeled image to obtain the second label information corresponding to the second labeled image. The parameters of the initial second image recognition model are adjusted based on each second labeled image and the second label information. After continuous iteration until the model converges, the target second image recognition model is obtained.

[0073] Step S203: Train the initial image segmentation model based on the third target sample image to obtain the target segmentation model;

[0074] In this step, the DeepLabv3 semantic segmentation model can be used to segment the third target sample image, obtaining a label image corresponding to the third target sample image, containing cargo region features and container region features. The label image includes three pixel regions: a cargo pixel region, a container pixel region, and a background pixel region, with each region having a different pixel size. Based on the label image, the initial segmentation model is trained to obtain a first label image corresponding to the first label image. The parameters of the initial segmentation model are adjusted based on the label image and the first label image, iterating continuously until the model converges, thus obtaining the target segmentation model.

[0075] Step S204: Obtain the image to be evaluated;

[0076] In this step, the image acquisition equipment in the mining area is first used to acquire images of the target mining area, obtaining several original images. These original images are then filtered to obtain a first original image whose clarity meets predetermined conditions, and a second original image whose clarity does not meet these conditions. The first original image includes images acquired during the day that meet the preset clarity conditions and images acquired at night that meet the preset clarity conditions. The second original image includes images acquired during the day that do not meet the preset clarity conditions and images acquired at night that do not meet the preset clarity conditions. The images acquired during the day that do not meet the preset clarity conditions are processed using a histogram equalization algorithm to obtain an enhanced daytime image. The images acquired at night that do not meet the preset clarity conditions are processed using a Laplacian operator to obtain an enhanced nighttime image. The enhanced daytime image and the enhanced nighttime image constitute the first enhanced image. The first original image and the first enhanced image are combined to obtain the image to be evaluated. This image enhancement process improves data contrast, thereby increasing the recognition rate and accuracy of subsequent mine car identification and gap identification, laying the foundation for improving the accuracy of the evaluation results.

[0077] Step S205: Based on the target first image recognition model, perform mining truck feature recognition processing on the target image to be evaluated to obtain several first images containing mining truck features;

[0078] In this step, the target image to be evaluated is subjected to mine car feature recognition based on the first target image recognition model to obtain a first image containing mine car features. Images that do not contain mine cars are then filtered and deleted from the target image.

[0079] Step S206: Based on the target second image recognition model, perform gap feature recognition processing on the cargo box area and cargo area in the first image to obtain a second image that does not contain gap features;

[0080] In this step, based on the target second image recognition model, gap features are identified in the cargo box area and cargo area of ​​the first image to obtain a second image that does not contain gap features. Images in the first image containing gap features are filtered out, as these indicate unqualified loading quality. The loading quality can be effectively determined through simple image recognition. However, the loading quality of the second image without gap features is not necessarily qualified; therefore, further evaluation is needed to obtain a more accurate assessment result.

[0081] Step S207: Perform image segmentation processing on the second image based on the target segmentation model to obtain the target image corresponding to the second image;

[0082] In this step, the second image is segmented based on the target segmentation model to obtain a target image corresponding to the second image, which includes a cargo pixel region, a cargo box pixel region, and a background pixel region. The pixel size of each pixel region in the target image is different, which lays the foundation for subsequent evaluation of loading quality based on the pixel size of the cargo region and the pixel size of the cargo box region in the target image.

[0083] Step S208: Evaluate the loading quality of the mining truck based on the target image to obtain the evaluation result.

[0084] In this step, a loading quality assessment value is calculated based on the pixel size of the cargo area and cargo box area of ​​the target image. The calculation process for the loading quality assessment value is as follows: First, the sum of the number of pixels in the cargo box area and the number of pixels in the cargo area is calculated to obtain the total number of pixels in the cargo box area and the cargo area; then, the ratio of the number of pixels in the cargo area to the total number of pixels in the cargo box area and the cargo area is calculated to obtain the loading quality assessment value; finally, the loading quality of the mining car is assessed based on the loading quality assessment value to obtain the assessment result. Specifically, when the loading quality assessment value is greater than or equal to a preset threshold, the loading quality of the mining car is qualified; when the loading quality assessment value is less than the preset threshold, the loading quality of the mining car is unqualified. The preset threshold can be adjusted according to actual needs and different models.

[0085] This application constructs an image recognition model to identify gaps between the mine car and the cargo and the cargo box, and constructs an image segmentation model to segment the cargo area and cargo box area of ​​the mine car. The loading quality assessment value is calculated based on the number of pixels in the cargo area and cargo box area to obtain the assessment result of the loading quality of the mine car, which can reduce the assessment cost and make the assessment result more accurate.

[0086] Another embodiment of this application provides a mine car loading quality assessment device, comprising:

[0087] Mining truck feature recognition module 1: used to perform mining truck feature recognition on the target image to be evaluated, and obtain several first images containing mining truck features;

[0088] Gap feature recognition module 2: used to perform gap feature recognition on the cargo box area and cargo area in the first image, and obtain a second image that does not contain gap features;

[0089] Image segmentation module 3: used to perform image segmentation processing on the second image to obtain a target image corresponding to the second image;

[0090] Evaluation module 4: Used to evaluate the loading quality of the mining truck based on the target image and obtain the evaluation result.

[0091] In this embodiment, the mine car loading quality assessment method device further includes: an acquisition module, which is used to acquire each of the images to be evaluated. Specifically, the acquisition module is used to: acquire images of the target mining area, obtain several original acquired images, filter each of the original acquired images to obtain a first original acquired image whose clarity meets a predetermined condition, and a second original acquired image whose clarity does not meet the predetermined condition; based on the image attributes of each second original acquired image, process the second original acquired image using an image enhancement processing method corresponding to the image attributes to obtain a first enhanced image; and combine the first original acquired image and the first enhanced image to obtain the target image to be evaluated.

[0092] Specifically, the acquisition module is further configured to: perform image enhancement processing on the image in the second original acquired image whose image attribute is a daytime acquired image that does not meet the preset sharpness condition using a histogram equalization algorithm, to obtain a daytime enhanced image;

[0093] The Laplacian operator algorithm is used to perform image enhancement processing on the images in the second original acquired image that do not meet the preset sharpness conditions in the night acquired image, so as to obtain the night enhanced image;

[0094] The enhanced daytime image and the enhanced nighttime image are combined to obtain the first enhanced image.

[0095] Specifically, the mine car loading quality assessment device further includes: a first image recognition model establishment module, which is specifically used for: performing recognition processing on a pre-acquired sample image to obtain a first target sample image containing mine car features; annotating the first target sample image to obtain each first annotated image with annotation information; and training an initial first image recognition model based on each first annotated image to obtain a target first image recognition model.

[0096] Specifically, the mining truck feature recognition module is used to: perform mining truck feature recognition on each of the target images to be evaluated based on the target first image recognition model, and obtain a first image containing mining truck features.

[0097] Specifically, the mine truck loading quality assessment device further includes: a target second image recognition model establishment module, which is specifically used for: performing recognition processing on the first target sample image to obtain a second target sample image containing gap features and a third target sample image not containing gap features; annotating the second target sample image to obtain each second annotated image with annotation information; and training the initial second image recognition model based on each second annotated image to obtain the target second image recognition model.

[0098] Specifically, the gap feature recognition module 2 is used to: perform gap feature recognition on each of the first images based on the target second image recognition model to obtain a second image that does not contain gap features.

[0099] Specifically, the mining truck loading quality assessment device further includes a target segmentation model establishment module, which is specifically used to: perform image segmentation processing on the pre-acquired third target sample image to obtain a label image corresponding to the third target sample image, containing cargo area features and cargo box area features; and train an initial segmentation model based on the third target sample image and the label image to obtain a target segmentation model.

[0100] Specifically, the image segmentation module 3 is used to: perform image segmentation processing on each of the second images based on the target segmentation model to obtain the target image corresponding to the second image.

[0101] Specifically, the evaluation module 4 is used to: calculate and obtain an evaluation value of the loading quality based on the pixel size of the cargo area and the pixel size of the cargo box area in the target image; and obtain the loading quality evaluation result of the mining car based on the evaluation value of the loading quality.

[0102] This application constructs an image recognition model to identify gaps between mining trucks, cargo, and cargo boxes, excluding invalid images that do not contain mining trucks and images containing gap areas indicating substandard loading. By constructing an image segmentation model, the cargo area and cargo box area of ​​the mining truck are segmented in images that do not contain gap features. The loading quality assessment value is calculated based on the number of pixels in the cargo area and cargo box area to obtain the assessment result of the mining truck loading quality. This can effectively reduce the assessment cost and make the assessment result more accurate.

[0103] Another embodiment of this application provides a storage medium storing a computer program, which, when executed by a processor, implements the following method steps:

[0104] Step 1: Perform mining truck feature recognition processing on the target image to be evaluated to obtain several first images containing mining truck features;

[0105] Step 2: Perform gap feature recognition processing on the cargo box area and cargo area in the first image to obtain a second image that does not contain gap features;

[0106] Step 3: Perform image segmentation processing on the second image to obtain the target image corresponding to the second image;

[0107] Step 4: Evaluate the loading quality of the mining truck based on the target image to obtain the evaluation results.

[0108] This application constructs an image recognition model to identify gaps between mining trucks, cargo, and cargo boxes, excluding invalid images that do not contain mining trucks and images containing gap areas indicating substandard loading. By constructing an image segmentation model, the cargo area and cargo box area of ​​the mining truck are segmented in images that do not contain gap features. The loading quality assessment value is calculated based on the number of pixels in the cargo area and cargo box area to obtain the assessment result of the mining truck loading quality. This can effectively reduce the assessment cost and make the assessment result more accurate.

[0109] Another embodiment of this application provides an electronic device, including at least a memory and a processor. The memory stores a computer program, and the processor, when executing the computer program in the memory, performs the following method steps:

[0110] Step 1: Perform mining truck feature recognition processing on the target image to be evaluated to obtain several first images containing mining truck features;

[0111] Step 2: Perform gap feature recognition processing on the cargo box area and cargo area in the first image to obtain a second image that does not contain gap features;

[0112] Step 3: Perform image segmentation processing on the second image to obtain the target image corresponding to the second image;

[0113] Step 4: Evaluate the loading quality of the mining truck based on the target image to obtain the evaluation results.

[0114] This application constructs an image recognition model to identify gaps between mining trucks, cargo, and cargo boxes, excluding invalid images that do not contain mining trucks and images containing gap areas indicating substandard loading. By constructing an image segmentation model, the cargo area and cargo box area of ​​the mining truck are segmented in images that do not contain gap features. The loading quality assessment value is calculated based on the number of pixels in the cargo area and cargo box area to obtain the assessment result of the mining truck loading quality. This can effectively reduce the assessment cost and make the assessment result more accurate.

[0115] The above embodiments are merely exemplary embodiments of this application and are not intended to limit this application. The scope of protection of this application is defined by the claims. Those skilled in the art can make various modifications or equivalent substitutions to this application within its substance and scope of protection, and such modifications or equivalent substitutions should also be considered to fall within the scope of protection of this application.

Claims

1. A method for assessing the loading quality of mining cars, characterized in that, include: Based on the target first image recognition model trained using the YOLOv5 algorithm, the target image to be evaluated in the target mining area is processed to identify the mine car features, and several first images containing mine car features are obtained. Based on the target second image recognition model trained using the YOLOv5 algorithm, gap feature recognition processing is performed on the cargo box area and cargo area in the first image to obtain a second image that does not contain gap features. The second image is segmented based on the target segmentation model trained using the deeplabv3 semantic segmentation algorithm to obtain the target image corresponding to the second image. The loading quality of the mining truck is evaluated based on the target image, and the evaluation result is obtained. The evaluation of the loading quality of the mining truck based on the target image, to obtain the evaluation result, specifically includes: Calculate the sum of the number of pixels in the cargo box area and the number of pixels in the cargo area to obtain the total number of pixels in both areas. The loading quality assessment value is obtained by calculating the ratio of the number of pixels in the cargo area to the total number of pixels in the container area and the cargo area. The loading quality of the mining truck is evaluated based on the loading quality assessment value to obtain the evaluation result; The pre-acquired sample images are processed to obtain a first target sample image containing the features of a mining truck; The first target sample image is annotated to obtain first annotated images with annotation information; Based on each of the first labeled images, the initial first image recognition model is trained to obtain the target first image recognition model; The step of performing mine truck feature recognition on the target image to be evaluated to obtain several first images containing mine truck features specifically includes: Based on the target first image recognition model, mine truck feature recognition is performed on each of the target images to be evaluated to obtain a first image containing mine truck features; The first target sample image is processed to obtain a second target sample image containing gap features and a third target sample image not containing gap features; The second target sample image is annotated to obtain each second annotated image with annotation information; The initial second image recognition model is trained based on each of the second labeled images to obtain the target second image recognition model; Perform gap feature recognition on the cargo box area and cargo area in the first image to obtain a second image that does not contain gap features, specifically including: Based on the target second image recognition model, gap feature recognition is performed on each of the first images to obtain a second image that does not contain gap features.

2. The method as described in claim 1, characterized in that, Before performing mine car feature recognition processing on the target image to be evaluated to obtain several first images containing mine car features, the method further includes: acquiring each of the images to be evaluated, specifically including: Image acquisition was performed on the target mining area to obtain several raw images; Each of the original acquired images is filtered to obtain a first original acquired image whose clarity meets the preset conditions, and a second original acquired image whose clarity does not meet the preset conditions. Based on the image attributes of each second original acquired image, the second original acquired image is processed using an image enhancement processing method corresponding to the image attributes to obtain the first enhanced image after image enhancement; The first original acquired image and the first enhanced image are combined to obtain the target image to be evaluated.

3. The method as described in claim 2, characterized in that, The step of processing the second original acquired images based on their image attributes using an image enhancement method corresponding to those attributes to obtain an enhanced first image specifically includes: The histogram equalization algorithm is used to enhance the images in the second original acquired image that do not meet the preset sharpness conditions and whose image attributes are daytime acquired images, to obtain daytime enhanced images; The Laplacian operator algorithm is used to perform image enhancement processing on the images in the second original acquired image that do not meet the preset sharpness conditions in the night acquired image, so as to obtain the night enhanced image; The enhanced daytime image and the enhanced nighttime image are combined to obtain the first enhanced image.

4. The method as described in claim 1, characterized in that, The method further includes: The pre-acquired third target sample image is subjected to image segmentation processing to obtain a label image corresponding to the third target sample image, which includes cargo area features and container area features; The initial segmentation model is trained based on the third target sample image and the label image to obtain the target segmentation model; The step of performing image segmentation processing on the second image to obtain the target image corresponding to the second image specifically includes: Based on the target segmentation model, each of the second images is segmented to obtain the target image corresponding to the second image.

5. The method as described in claim 1, characterized in that, The evaluation of the loading quality of the mining truck based on the target image, to obtain the evaluation result, specifically includes: The loading quality assessment value is calculated based on the pixel size of the cargo area and the pixel size of the container area in the target image. The loading quality assessment result of the mining car is obtained based on the assessment value of the loading quality.

6. A mine car loading quality assessment device, used to implement any one of the mine car loading quality assessment methods as described in claims 1 to 5, characterized in that, include: The mine car feature recognition module is used to perform mine car feature recognition on the target image to be evaluated in the target mining area based on the target first image recognition model trained with the YOLOv5 algorithm, and obtain several first images containing mine car features. Gap feature recognition module: used to perform gap feature recognition on the cargo box area and cargo area in the first image based on the target second image recognition model trained with YOLOv5 algorithm, and obtain a second image that does not contain gap features; Image segmentation module: used to perform image segmentation processing on the second image based on the target segmentation model trained using the deeplabv3 semantic segmentation algorithm, and obtain the target image corresponding to the second image; Evaluation module: Used to evaluate the loading quality of the mining truck based on the target image and obtain the evaluation result. Specifically, it calculates the sum of the number of pixels in the cargo box area and the number of pixels in the cargo area to obtain the total number of pixels in the cargo box area and the cargo area; calculates the ratio of the number of pixels in the cargo area to the total number of pixels in the cargo box area and the cargo area to obtain the loading quality evaluation value; and evaluates the loading quality of the mining truck based on the loading quality evaluation value to obtain the evaluation result.

7. A storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the steps of the mining truck loading quality assessment method according to any one of claims 1-5.

8. An electronic device, characterized in that, It includes at least a memory and a processor, the memory The memory stores a computer program, and the processor, when executing the computer program in the memory, implements the steps of the mining truck loading quality assessment method according to any one of claims 1-5.

Images