Non-contact intelligent real-time detection method and device for machine-made sand and gravel powder content

Abstract

The embodiment of the present application provides a non-contact intelligent real-time detection method and device for the content of manufactured sand and stone powder, the method comprising: obtaining target manufactured sand image data; inputting the target manufactured sand image data into a trained stone powder content detection model, and determining the stone powder content detection result data in the manufactured sand corresponding to the target manufactured sand image data based on the output of the stone powder content detection model. The present application can realize non-contact real-time detection of the content of manufactured sand and stone powder, can effectively improve the degree of intelligence, automation, convenience and efficiency of the process of detecting the content of manufactured sand and stone powder, and can effectively improve the accuracy and real-time performance of the detection results of the content of manufactured sand and stone powder.

Description

Technical Field

[0001] The present application relates to the technical field of artificial sand composition detection, and specifically to a non-contact intelligent real-time detection method and device for artificial sand and stone powder content. Background Art

[0002] As a local resource, natural sand is non-renewable in the short term. Due to the limitation of natural sand resources, the growing demand for concrete can no longer be met, especially in the southwest of my country. For the purpose of environmental protection, the use of machine-made sand instead of natural sand has become a sustainable development trend in the concrete industry. Machine-made sand refers to rock particles with a particle size of less than 4.75 mm made by mechanical crushing and screening. It is significantly different from natural sand in terms of particle shape and gradation. Particles with a particle size of less than 0.075 mm in machine-made sand are called stone powder, which is one of the important differences between machine-made sand and natural sand. An appropriate amount of stone powder can improve the cohesion and water retention of the mixture, improve the interface performance of concrete, optimize the gradation of concrete particles, improve the compactness of concrete, and improve the comprehensive performance of concrete. However, a higher content of stone powder can easily lead to adverse phenomena such as bleeding and honeycombing in concrete. Therefore, in the production and construction process, it is necessary to strictly control the stone powder content of machine-made sand.

[0003] At present, the detection method of stone powder content in machine-made sand adopts physical or chemical methods. The current sand and gravel specification "Construction Sand" (GB / T14684-2011) has specific provisions for the determination method of stone powder content, and professional instruments are required to complete the relevant steps. Among them, the physical method is screening or water washing. The test requires the use of sand and gravel standard sieves or vibrating sieves, beakers, glass rods and other experimental instruments to perform screening, washing, drying and other steps; the chemical method mainly uses methylene blue reagent to detect the stone powder content. Because light and air can denature the methylene blue solution, it is necessary to prepare the test solution with methylene blue powder before the test, which takes about 2 hours or more. The test requires the use of water baths, impeller mixers, beakers, glass rods, test papers and other experimental instruments to prepare methylene blue solution, stir, timed quantitative sampling and titration and other steps. The test results are qualitative analysis and cannot accurately quantify the stone powder content.

[0004] However, the current methods for testing the content of artificial sand and gravel powder all have problems such as sample collection requiring manual participation, complex and time-consuming testing methods, and reliance on manual experience and the accuracy of testing equipment, resulting in poor testing accuracy. Summary of the invention

[0005] In response to the problems in the prior art, the present application provides a non-contact intelligent real-time detection method and device for the content of artificial sand and gravel powder, which can realize non-contact detection of the content of artificial sand and gravel powder, effectively improve the intelligence, automation, convenience and efficiency of the detection process of the content of artificial sand and gravel powder, and effectively improve the accuracy and real-time nature of the detection results of the content of artificial sand and gravel powder.

[0006] In order to solve the above technical problems, this application provides the following technical solutions:

[0007] In the first aspect, the present application provides a non-contact intelligent real-time detection method for machine-made sand and gravel powder content, comprising:

[0008] Acquire target machine-made sand image data;

[0009] The target machine-made sand image data is input into a trained stone powder content detection model, and the stone powder content detection result data in the machine-made sand corresponding to the target machine-made sand image data is determined based on the output of the stone powder content detection model.

[0010] Furthermore, before inputting the target machine-made sand image data into the trained stone powder content detection model, the method further includes:

[0011] Acquire multiple historical machine-made sand image data and labels corresponding to each of the historical machine-made sand image data, wherein the acquisition area of ​​the historical machine-made sand image data is different from that of the target machine-made sand image data; and the label is used to indicate the stone powder content data in the machine-made sand corresponding to the historical machine-made sand image data;

[0012] A preset machine learning model is trained based on the labels corresponding to each of the historical machine-made sand image data to obtain a stone powder content detection model for detecting the stone powder content in machine-made sand.

[0013] Furthermore, the step of obtaining a plurality of historical machine-made sand image data and labels corresponding to each of the historical machine-made sand image data includes:

[0014] Obtain samples of manufactured sand;

[0015] The manufactured sand sample is doped with stone powder so that the mud content of the manufactured sand sample reaches a preset target mud content, and then an image acquisition device is used to acquire an image of the manufactured sand sample that reaches the target mud content to obtain corresponding original image data of the manufactured sand;

[0016] Various historical machine-made sand image data are extracted from the machine-made sand original image data, and each of the historical machine-made sand image data is labeled according to the target mud content to obtain a label corresponding to each of the historical machine-made sand image data.

[0017] Furthermore, before inputting the target machine-made sand image data into the trained stone powder content detection model, the method further includes:

[0018] The target machine-made sand image data is magnified based on a preset image magnification factor and then the magnified target machine-made sand image data is collected.

[0019] Furthermore, the target machine-made sand image data is input into a trained stone powder content detection model, and the stone powder content detection result data in the machine-made sand corresponding to the target machine-made sand image data is determined based on the output of the stone powder content detection model, including:

[0020] Inputting the plurality of target machine-made sand image data belonging to the same batch of target machine-made sand into the trained stone powder content detection model respectively, so as to obtain the stone powder content detection result data corresponding to each of the target machine-made sand image data respectively output by the stone powder content detection model;

[0021] The stone powder content detection result data corresponding to each of the target machine-made sand image data are averaged to obtain the corresponding stone powder content detection average data;

[0022] The average data of the stone powder content detection is used as the stone powder content detection result data of the target machine-made sand.

[0023] Furthermore, the extracting of each historical machine-made sand image data from the machine-made sand original image data includes:

[0024] Performing region extraction on each original image in the machine-made sand original image data to form at least one region map corresponding to each original image;

[0025] Performing image processing on each of the region images to form sample images corresponding to each of the region images, wherein the image processing includes: at least one of format conversion, size change and histogram equalization processing;

[0026] Each of the sample images is stored as historical machine-made sand image data.

[0027] Furthermore, the preset machine learning model is trained based on the labels corresponding to each of the historical machine-made sand image data to obtain a stone powder content detection model for detecting the stone powder content in the machine-made sand, including:

[0028] Dividing the plurality of historical machine-made sand image data with the labels into a training set, a test set and a validation set respectively;

[0029] Training a preset deep learning model based on the training set to obtain an initial model for detecting the stone powder content in machine-made sand;

[0030] The test set and the validation set are applied to generate the stone powder content detection model according to the initial model.

[0031] Furthermore, the obtaining of target machine-made sand image data includes:

[0032] receiving a surface image of machine-made sand in a target area acquired by at least one image acquisition device;

[0033] The surface image is pre-processed to form at least one target manufactured sand image data.

[0034] In a second aspect, the present application provides a non-contact intelligent real-time detection device for machine-made sand and stone powder content, comprising:

[0035] A target data acquisition module is used to acquire target machine-made sand image data;

[0036] The prediction model application module is used to input the target machine-made sand image data into a trained stone powder content detection model, and determine the stone powder content detection result data in the machine-made sand corresponding to the target machine-made sand image data based on the output of the stone powder content detection model.

[0037] Furthermore, it also includes:

[0038] A historical data acquisition module, used to acquire a plurality of historical machine-made sand image data and labels corresponding to each of the historical machine-made sand image data, wherein the acquisition area of ​​the historical machine-made sand image data is different from that of the target machine-made sand image data; and the labels are used to indicate the stone powder content data in the machine-made sand corresponding to the historical machine-made sand image data;

[0039] The prediction model training module is used to train a preset machine learning model based on the labels corresponding to each of the historical machine-made sand image data to obtain a stone powder content detection model for detecting the stone powder content in the machine-made sand.

[0040] Furthermore, the historical data acquisition module is used to perform the following:

[0041] Obtain samples of manufactured sand;

[0042] The manufactured sand sample is doped with stone powder so that the mud content of the manufactured sand sample reaches a preset target mud content, and then an image acquisition device is used to acquire an image of the manufactured sand sample that reaches the target mud content to obtain corresponding original image data of the manufactured sand;

[0043] Various historical machine-made sand image data are extracted from the machine-made sand original image data, and each of the historical machine-made sand image data is labeled according to the target mud content to obtain a label corresponding to each of the historical machine-made sand image data.

[0044] In a third aspect, the present application provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the non-contact intelligent real-time detection method for the content of artificial sand and gravel is implemented.

[0045] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the non-contact intelligent real-time detection method for the content of machine-made sand and gravel powder.

[0046] It can be seen from the above technical scheme that the present application provides a non-contact intelligent real-time detection method and device for artificial sand and stone powder content. The method obtains target artificial sand image data; inputs the target artificial sand image data into a trained stone powder content detection model, and determines the stone powder content detection result data in the artificial sand corresponding to the target artificial sand image data based on the output of the stone powder content detection model. Through image acquisition and artificial intelligence technology, non-contact detection of artificial sand and stone powder content can be achieved, which can effectively improve the intelligence, automation, convenience and efficiency of the artificial sand and stone powder content detection process, and can effectively improve the accuracy and realism of the artificial sand and stone powder content detection results. It is timely, thus realizing real-time adjustment of machine-made sand in the production process and rapid verification work of engineering personnel in scenarios without testing instruments and experimental reagents, avoiding the problems of poor accuracy caused by unrepresentative sampling and large consumption of manpower and material resources required for repeated experiments. It is also suitable for a variety of application scenarios. For example, in the machine-made sand sales scenario, there is no need for testing personnel to be present. The machine-made sand detection results can be provided to buyers based on the machine-made sand images collected non-contactly. For another example, in the machine-made sand production scenario, production personnel only need to provide production personnel with the adjustment of stone powder content in the machine-made sand or the quality inspection of stone powder content based on the machine-made sand images collected non-contactly. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0048] Figure 1It is a first flow chart of the non-contact intelligent real-time detection method of machine-made sand and gravel powder content in the embodiment of the present application.

[0049] Figure 2 It is a second flow chart of the non-contact intelligent real-time detection method of machine-made sand and gravel powder content in the embodiment of the present application.

[0050] Figure 3 It is a specific flow chart of step 010 in the non-contact intelligent real-time detection method of machine-made sand and gravel powder content in the embodiment of the present application.

[0051] Figure 4 It is a specific flow chart of step 013 in the non-contact intelligent real-time detection method of machine-made sand and gravel powder content in the embodiment of the present application.

[0052] Figure 5 It is a specific flow chart of step 020 in the non-contact intelligent real-time detection method of machine-made sand and gravel powder content in the embodiment of the present application.

[0053] Figure 6 It is a specific flow chart of step 100 in the non-contact intelligent real-time detection method of machine-made sand and gravel powder content in an embodiment of the present application.

[0054] Figure 7 It is the first processing logic diagram of the non-contact intelligent real-time detection method of machine-made sand and gravel powder content in the application example of this application.

[0055] Figure 8 This is the second processing logic diagram of the non-contact intelligent real-time detection method for machine-made sand and gravel powder content in the application example of this application.

[0056] Fig. 9 It is a first structural schematic diagram of a non-contact intelligent real-time detection device for machine-made sand and gravel powder content in an embodiment of the present application.

[0057] Fig.10 This is a second structural schematic diagram of the non-contact intelligent real-time detection device for machine-made sand and gravel powder content in the embodiment of the present application.

[0058] Fig.11 It is a schematic diagram of the structure of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0060] In the actual detection process, the existing test methods for the content of machine-made sand and gravel powder have the following shortcomings:

[0061] (1) To ensure the representativeness of the sampling, it is necessary to take samples from multiple locations and repeat the testing steps multiple times to make the test results more representative.

[0062] (2) A single test can take 2 to 4 hours. It is impossible to obtain data in real time, and it is even more difficult to make real-time adjustments. The overall test time is long and the detection process is time-consuming and labor-intensive.

[0063] (3) It is highly dependent on the experience of engineering personnel. The mass loss caused by screening and weighing directly leads to human errors, and its accuracy cannot be well controlled.

[0064] (4) The test requires the use of testing instruments and reagents. The instruments need to be cleaned and used repeatedly, and the reagents need to be stored in good containers and in a dry, light-proof environment.

[0065] Based on this, the present application provides a non-contact intelligent real-time detection method for machine-made sand and stone powder content, a non-contact intelligent real-time detection device for machine-made sand and stone powder content, an electronic device and a computer-readable storage medium, by acquiring target machine-made sand image data; inputting the target machine-made sand image data into a trained stone powder content detection model, and determining the stone powder content detection result data in the machine-made sand corresponding to the target machine-made sand image data based on the output of the stone powder content detection model. Through image acquisition and artificial intelligence technology, non-contact detection of machine-made sand and stone powder content can be achieved, and the intelligence, automation, convenience and efficiency of the machine-made sand and stone powder content detection process can be effectively improved, and the accuracy and real-time nature of the machine-made sand and stone powder content detection results can be effectively improved, thereby realizing real-time adjustment of machine-made sand in the production process, as well as rapid verification work by engineering personnel in scenarios without detection instruments and experimental reagents, avoiding the problems of poor accuracy caused by unrepresentative sampling and large consumption of manpower and material resources required for repeated experiments.

[0066] Based on the above content, the present application also provides a non-contact intelligent real-time detection device for machine-made sand and stone powder content, which is used to implement the non-contact intelligent real-time detection method for machine-made sand and stone powder content provided in one or more embodiments of the present application. The non-contact intelligent real-time detection device for machine-made sand and stone powder content can also be communicated with at least one client device to send the stone powder content detection result data in the machine-made sand corresponding to the target machine-made sand image data to at least one client device when or after obtaining the stone powder content detection result data in the machine-made sand corresponding to the target machine-made sand image data, so that the user can adjust the content of the machine-made sand or decide whether to purchase it in a timely and targeted manner according to the stone powder content detection result data in the machine-made sand corresponding to the target machine-made sand image data displayed on the client device.

[0067] It is understandable that the non-contact intelligent real-time detection device for machine-made sand and stone powder content can be a server. The part of performing non-contact intelligent real-time detection of machine-made sand and stone powder content can be executed at the server node where the non-contact intelligent real-time detection device for machine-made sand and stone powder content as described above is located.

[0068] In another practical application scenario, the part of performing non-contact intelligent real-time detection of machine-made sand and stone powder content can be executed at the server node where the non-contact intelligent real-time detection device for machine-made sand and stone powder content as described above is located, or all operations can be completed in the client device. The specific selection can be based on the processing capacity of the client device and the limitations of the user's usage scenario. This application is not limited to this. If all operations are completed in the client device, the client device may also include a processor for specific processing of the non-contact intelligent real-time detection of machine-made sand and stone powder content.

[0069] It is understandable that the client device may include a smart phone, a tablet electronic device, a network set-top box, a portable computer, a desktop computer, a personal digital assistant (PDA), a vehicle-mounted device, a smart wearable device, etc. Among them, the smart wearable device may include smart glasses, a smart watch, a smart bracelet, etc.

[0070] The above-mentioned client device may have a communication module (i.e., a communication unit) that can communicate with a remote server to achieve data transmission with the server. The communication unit can also receive the prediction results returned by the server. The server may include a server on the side of the task scheduling center, and other implementation scenarios may also include a server of an intermediate system, such as a server of a third-party server system that has a communication link with the task scheduling center server. The server may include a single computer device, or a server cluster consisting of multiple servers, or a server structure of a distributed device.

[0071] The server and the client device may communicate with each other using any suitable network protocol, including network protocols that have not yet been developed on the date of filing this application. The network protocols may include, for example, TCP / IP, UDP / IP, HTTP, HTTPS, etc. Of course, the network protocols may also include, for example, RPC (Remote Procedure Call Protocol) and REST (Representational State Transfer) protocols used on top of the above protocols.

[0072] In one or more embodiments of the present application, the present application proposes a method that can quickly, conveniently and accurately identify the sand and gravel powder content, and uses a computational model based on artificial intelligence to perform image processing on the image information of artificial sand, thereby realizing real-time adjustment of artificial sand in the production process, as well as rapid verification work by engineering personnel in scenarios where there are no detection instruments or experimental reagents, thereby avoiding the inaccuracy caused by unrepresentative sampling and the manpower and material resources required for repeated experiments.

[0073] The details are described in detail through the following multiple embodiments.

[0074] In one or more embodiments of the present application, the stone powder content is for artificial sand. The non-contact intelligent real-time detection method for artificial sand and stone powder content in the present application is also applicable to the detection of mud content. The mud content specifically refers to the mud content of natural sand. That is to say, the non-contact intelligent real-time detection method for artificial sand and stone powder content in the present application is also applicable to non-contact intelligent real-time detection of mud content of natural sand, etc.

[0075] In order to solve the problems of the existing test methods for the content of artificial sand and gravel powder, such as the sample collection method requires manual participation, the detection method is complicated and time-consuming, and relies on manual experience and the accuracy of the detection equipment, the present application provides an embodiment of a non-contact intelligent real-time detection method for the content of artificial sand and gravel powder, see Figure 1 The non-contact intelligent real-time detection method for the content of machine-made sand and gravel powder specifically includes the following contents:

[0076] Step 100: Acquire target machine-made sand image data.

[0077] In step 100, the stacked machine-made sand can be directly imaged at any angle and distance using an image acquisition device. The image acquisition device includes but is not limited to a mobile phone, a camera, a video camera, a camera, a monitoring device, and the like.

[0078] It can be understood that the target machine-made sand image data refers to the image of the machine-made sand pile in the target area, and the image can be collected from any angle.

[0079] Step 200: input the target machine-made sand image data into a trained stone powder content detection model, and determine the stone powder content detection result data in the machine-made sand corresponding to the target machine-made sand image data based on the output of the stone powder content detection model.

[0080] In step 200, the surface image information of the machine-made sand can be input into a model trained by machine learning, and the stone powder content of the machine-made sand can be automatically calculated. The stone powder content detection model is pre-trained based on a preset machine learning model, and the machine learning model can specifically be a deep learning model, such as a convolutional neural network.

[0081] Specifically, the application of the solutions provided in the above steps 100 to 200 does not require the use of detection instruments and reagents for testing, and adopts non-contact image acquisition and artificial intelligence calculation methods. The image acquisition equipment is portable and the image information sampling is convenient. The artificial sand and stone powder content is obtained in real time. After collecting the image information of the artificial sand surface, the target information can be obtained in real time, which can realize real-time adjustment in the artificial sand production process and rapid verification of stone powder content in construction scenarios, saving time and effort. It does not rely on the experience of engineering personnel, eliminates human errors, and can better control the accuracy of the output results.

[0082] From the above description, it can be seen that the non-contact intelligent real-time detection method for the content of artificial sand and stone powder provided in the embodiment of the present application can realize non-contact detection of the content of artificial sand and stone powder through image acquisition and artificial intelligence technology, and can effectively improve the intelligence, automation, convenience and efficiency of the detection process of artificial sand and stone powder content, and can effectively improve the accuracy and real-time of the detection results of artificial sand and stone powder content, thereby realizing real-time adjustment of artificial sand in the production process, and rapid verification work of engineering personnel in scenarios without detection instruments and experimental reagents, avoiding the problems of poor accuracy caused by unrepresentative sampling and large consumption of manpower and material resources required for repeated experiments, and is suitable for a variety of application scenarios. For example, in the artificial sand sales scenario, there is no need for detection personnel to be present, and the artificial sand detection results can be provided to buyers based on the artificial sand images collected non-contactly. For example, in the artificial sand production scenario, the production personnel only need to provide the production personnel with the adjustment of the stone powder content in the artificial sand or the quality inspection of the stone powder content based on the artificial sand images collected non-contactly. Provide the artificial sand detection results.

[0083] In order to effectively improve the automation, intelligence and accuracy of the non-contact intelligent real-time detection results of the mechanism sand and stone powder content, in one embodiment of the non-contact intelligent real-time detection method of mechanism sand and stone powder content provided in the present application, see Figure 2 The non-contact intelligent real-time detection method for the content of machine-made sand and gravel powder includes the following contents before step 100:

[0084] Step 010: Acquire multiple historical machine-made sand image data and labels corresponding to each of the historical machine-made sand image data, wherein the collection area of ​​the historical machine-made sand image data is different from the target machine-made sand image data; and the label is used to represent the stone powder content data in the machine-made sand corresponding to the historical machine-made sand image data.

[0085] Specifically, when machine learning is performed, inputting historical machine-made sand image data plays a decisive role in the output stone powder content detection model. By using a large amount of image information to support the accuracy of the model, improving the quality and quantity of historical machine-made sand image data can directly improve the efficiency of machine learning. The atlas used for machine learning in this application is rich in samples, containing machine-made sand produced by various gradations, various stone powder contents, various parent rocks, and various crushers. The samples are more uniform, and the stone powder content detection model obtained after inputting it into the preset machine learning model for calculation has good robustness.

[0086] Step 020: Train a preset machine learning model based on the labels corresponding to each of the historical machine-made sand image data to obtain a stone powder content detection model for detecting the stone powder content in machine-made sand.

[0087] In order to effectively improve the application reliability and accuracy of historical machine-made sand image data, in one embodiment of the non-contact intelligent real-time detection method for machine-made sand and stone powder content provided in this application, see Figure 3 Step 010 in the non-contact intelligent real-time detection method for the content of machine-made sand and gravel powder further specifically includes the following contents:

[0088] Step 011: Get the manufactured sand sample.

[0089] Step 012: The manufactured sand sample is doped with stone powder so that the mud content of the manufactured sand sample reaches a preset target mud content, and then an image acquisition device is used to acquire an image of the manufactured sand sample that has reached the target mud content to obtain corresponding original image data of the manufactured sand.

[0090] Step 013: extracting each historical machine-made sand image data from the machine-made sand original image data, and labeling each historical machine-made sand image data according to the target mud content to obtain a label corresponding to each historical machine-made sand image data.

[0091] In order to effectively improve the effectiveness and efficiency of preprocessing the original image of the machine-made sand surface, in one embodiment of the non-contact intelligent real-time detection method for the content of machine-made sand and stone powder provided in the present application, see Figure 4 Step 013 in the non-contact intelligent real-time detection method for the content of machine-made sand and gravel powder further specifically includes the following contents:

[0092] Step 0131: performing region extraction on each original image in the machine-made sand original image data to form at least one region map corresponding to each original image.

[0093] Step 0132: Perform image processing on each of the area images to form sample images corresponding to each of the area images, wherein the image processing includes at least one of format conversion, size change and histogram equalization processing.

[0094] Step 0133: Store each of the sample images as historical machine-made sand image data.

[0095] Step 0134: Label each of the historical machine-made sand image data according to the target mud content to obtain a label corresponding to each of the historical machine-made sand image data.

[0096] Specifically, the original images in the pre-collected machine-made sand surface image atlas can be preprocessed to establish a dataset of machine-made sand feature images. The image preprocessing process mainly includes extracting the region of interest, converting the RGB image into a grayscale image, shrinking the transformation, histogram equalization, and adding labels. Image preprocessing, the original image data is input for preprocessing. One original image can generate multiple sample images of the same size with labels, which can be used for image data of machine learning models.

[0097] In order to effectively improve the application reliability and accuracy of the target machine-made sand image data, in one embodiment of the non-contact machine-made sand and stone powder content intelligent real-time detection method provided in the present application, the non-contact machine-made sand and stone powder content intelligent real-time detection method further specifically includes the following contents between step 100 and step 200:

[0098] Step 101: amplifying the target machine-made sand image data based on a preset image magnification factor and collecting the amplified target machine-made sand image data.

[0099] The specific image magnification may be 2 to 20 times, preferably 4 times.

[0100] In order to further improve the accuracy of the detection of the content of artificial sand and stone powder, in one embodiment of the non-contact intelligent real-time detection method of artificial sand and stone powder content provided in the present application, step 200 in the non-contact intelligent real-time detection method of artificial sand and stone powder content specifically includes the following contents:

[0101] Step 210: inputting the plurality of target machine-made sand image data belonging to the same batch of target machine-made sand into the trained stone powder content detection model respectively, so as to obtain the stone powder content detection result data corresponding to each of the target machine-made sand image data respectively output by the stone powder content detection model;

[0102] Step 220: performing an average calculation process on the stone powder content detection result data corresponding to each of the target machine-made sand image data to obtain corresponding stone powder content detection average data;

[0103] Step 230: Using the stone powder content detection average data as the stone powder content detection result data of the target machine-made sand.

[0104] In order to effectively improve the accuracy and reliability of machine learning model training, in one embodiment of the non-contact intelligent real-time detection method for machine-made sand and stone powder content provided in this application, see Figure 5 , step 020 in the non-contact intelligent real-time detection method for the content of machine-made sand and gravel powder further specifically includes the following contents:

[0105] Step 021: Divide the plurality of historical machine-made sand image data with the labels into a training set, a test set and a validation set respectively.

[0106] In one example, 60% of the historical machine-made sand image data with the labels can be randomly selected as a training set, another 20% of the historical machine-made sand image data with the labels can be used as a verification set, and the remaining 20% ​​of the historical machine-made sand image data with the labels can be used as a test set to test the accuracy of the machine learning model.

[0107] Step 022: Train a preset deep learning model based on the training set to obtain an initial model for detecting the stone powder content in machine-made sand.

[0108] Step 023: Apply the test set and the validation set to generate the stone powder content detection model according to the initial model.

[0109] In step 023, the data set is randomly divided into a training set, a validation set and a test set, and then the training set is used to train an initial model, and the validation set is used to validate the initial model. The initial model is continuously adjusted according to the situation, and the best model is selected. A final model is then trained using the training set and validation set data, and finally the final model is evaluated using the test set to form the final stone powder content detection model.

[0110] Specifically, during machine learning, inputting an atlas of machine-made sand surface images plays a decisive role in the output stone powder content detection model. By supporting the accuracy of the model with a large amount of image information, improving the quality and quantity of the atlas can directly improve the efficiency of machine learning. The atlas used for machine learning in this application is a unique machine-made sand atlas for this application. It has rich samples and contains machine-made sand produced by various gradations, various stone powder contents, various parent rocks, and various crushers. The samples are more uniform, and the stone powder content detection model obtained after inputting it into the machine learning model for calculation is more robust.

[0111] The process of building the model is as follows: In order to use machine learning to identify the stone powder content of artificial sand, the surface image of artificial sand with known characteristic values ​​is used as a sample input into the machine learning model, and the machine learning model is trained and verified.

[0112] The principle of implementing the stone powder content detection model is to input samples of known information into a preset machine learning model. Therefore, before training, the information of the image sample is first extracted and its features are predicted. Then, a relationship is established between it and the given known label. The error between the two is fed back to the model, and the machine learning model adjusts the parameters. Finally, the label of the sample can be identified, that is, the stone powder content of the sample is obtained.

[0113] In order to effectively improve the convenience and accuracy of obtaining the target machine-made sand image data, in one embodiment of the non-contact intelligent real-time detection method for machine-made sand and stone powder content provided in the present application, see Figure 6 The step 100 in the non-contact intelligent real-time detection method of machine-made sand and gravel powder content specifically includes the following contents:

[0114] Step 110: Receive a surface image of machine-made sand in a target area acquired by at least one image acquisition device.

[0115] Step 120: Pre-process the surface image to form at least one target manufactured sand image data.

[0116] Specifically, if the ratio of the environment occupied by artificial sand in the surface image is greater than or equal to a first threshold value, the surface image is respectively subjected to region extraction to form at least one target region map corresponding to the surface image; and, at least one of the target region maps is respectively subjected to image processing to form at least one target sample map corresponding to the surface image, so as to determine at least one of the target sample maps as target artificial sand image data.

[0117] In addition, if there are two or more image acquisition devices that capture the surface images of the same artificial sand pile in the target area, this problem can be solved by setting up two or more image acquisition devices. The image information of the same part captured by the two image acquisition devices can be compared to restore the actual size of the image information. Place a reference object of known size on the surface of the sand pile, such as a box, reference mark, etc. The information processing device can directly restore the surface information image of the artificial sand to its actual size through perspective transformation, affine transformation or other transformation methods. Automatically extract the area of ​​interest. After restoring the image, the information processing device will automatically extract the area of ​​interest in the image within the size reference frame as the basis for calculating the stone powder content.

[0118] Based on the above content, in order to further illustrate this solution, this application also provides a specific application example of a non-contact intelligent real-time detection method for machine-made sand and gravel powder content, see Figure 7 , by acquiring the machine-made sand image. The image acquisition device is used to collect the image information of the machine-made sand surface as the basis for calculating the stone powder content; the stone powder content is calculated. The information processing device automatically uses the artificial intelligence trained model to analyze the extracted area of ​​interest, and automatically calculates the stone powder content of the machine-made sand; and outputs the result. The display connected to the information processing device will display the stone powder content calculation result output by the information processing device in real time.

[0119] For further information, see Figure 8 After obtaining the machine-made sand image, the original image can also be rotated and corrected, and the information processing equipment can automatically extract the area of ​​interest and then calculate the stone powder content.

[0120] Specifically, the specific application examples of the non-contact intelligent real-time detection method of machine-made sand and gravel powder content are realized through the following contents:

[0121] Step 1: Get the image of machine-made sand. For the stacked machine-made sand, you can directly use image acquisition equipment to take images at any angle and distance. Image acquisition equipment includes but is not limited to mobile phones, cameras, video cameras, cameras, monitoring, etc.

[0122] Step 2: Calculate the stone powder content. The information processing equipment will input the surface image information of the machine-made sand into the model trained by machine learning, and automatically calculate the stone powder content of the machine-made sand.

[0123] When machine learning, inputting an atlas of machine-made sand surface images plays a decisive role in the output computing model. By supporting the accuracy of the model with a large amount of image information, improving the quality and quantity of the atlas can directly improve the efficiency of machine learning. The atlas used for machine learning in this application is a unique machine-made sand atlas for this application. It has rich samples and contains machine-made sand produced by various gradations, various stone powder contents, various parent rocks, and various crushers. The samples are more uniform, and the computing model obtained after inputting it into the machine learning calculation has good robustness.

[0124] The process of building the model is as follows: In order to use machine learning to identify the stone powder content of artificial sand, the surface image of artificial sand with known characteristic values ​​is used as a sample input into the machine learning model, and the machine learning model is trained and verified.

[0125] The principle of implementing the machine learning model is to input samples of known information into the computing model. Therefore, before training, the information of the image sample is extracted and its features are predicted. Then, a relationship is established between it and the given known label. The error between the two is fed back to the model, and the computing model adjusts the parameters. Finally, the label of the sample can be identified, that is, the stone powder content of the sample is obtained.

[0126] The original images in the pre-collected machine-made sand surface image atlas are preprocessed to establish a dataset of machine-made sand feature images. The image preprocessing process mainly includes extracting the area of ​​interest, converting the RGB image into a grayscale image, reducing the transformation, histogram equalization, and adding labels. Image preprocessing, the original image data is input for preprocessing. One original image can generate multiple sample images of the same size with labels, which can be used for image data of the machine learning model. In the atlas, 80% are randomly selected as the training atlas, and the remaining 20% ​​are used as the test atlas to test the accuracy of the machine learning model.

[0127] The method for extracting image features is as follows: a convolutional neural network can be used to extract the features of machine-made sand image data, and the surface image of the machine-made sand image is input into the convolutional neural network and extracted into several matrices, which are then visualized through image processing filters to form different feature maps corresponding to the relevant features represented by the corresponding images, such as the contour, area and proportion of the machine-made sand particles.

[0128] The specific process is:

[0129] (1) The preprocessed sample set is input into the machine learning model, and the output is a label representing the stone powder content. The convolutional neural network extracts spatial features by stacking convolutional layers and pooling layers.

[0130] (2) In a single convolutional layer, multiple local convolution kernels are used to calculate the input image to generate a feature map. After the convolution and pooling steps, the high-dimensional image matrix and the one-dimensional vector that actually represents the physical parameters of the sand particles are flattened using a flattening layer. The flattening layer is located between the convolutional neural network for feature learning and the fully connected layer.

[0131] (3) After flattening, a fully connected layer is used for regression calculation, and the data results of the above feature learning are output to the fully connected layer, so that the physical parameters of the sand particles can be fitted, and finally the calculation model of the machine-made sand and gravel powder content can be output.

[0132] After obtaining the calculation model of the artificial sand and stone powder content, the test set of pictures was used to test the accuracy of the model calculation. After testing, the model had a high accuracy rate and the time to identify a group of sand individually was about 0.3 seconds, which was much faster than the traditional recognition method and could achieve the effect of real-time output of stone powder content.

[0133] The non-contact intelligent real-time detection method for the content of machine-made sand and gravel powder may also include the following contents:

[0134] (1) Restore size:

[0135] Due to the angle of the image acquisition device during shooting, the actual size of each point in the captured image cannot be estimated intuitively. The positional relationship between the image acquisition device and the machine-made sand is determined through technical means.

[0136] Specifically, this problem can be solved by setting up two or more image acquisition devices. By comparing the image information of the same part acquired by the two image acquisition devices, the actual size of the image information can be restored. Reference objects of known size can also be placed on the surface of the sand pile, such as boxes, reference marks, etc. The information processing equipment can directly restore the surface information image of the machine-made sand to its actual size through perspective transformation, affine transformation or other transformation methods.

[0137] Automatically extract the area of ​​interest. After restoring the image, the information processing device will automatically extract the area of ​​interest in the image within the size reference frame as the basis for calculating the stone powder content.

[0138] (2) Obtaining high-definition images:

[0139] The particle size of stone powder in artificial sand is less than 0.075mm, and it is difficult to capture clear image information that can be used for feature calculation using traditional mobile phones, cameras and other equipment.

[0140] In order to obtain high-definition machine-made sand image information, the method that can be used specifically is to use high-definition image acquisition equipment, such as high-definition cameras, high-definition video cameras, etc. The image information obtained by high-definition equipment has high resolution, and there are more pixels on the same size of image, which is conducive to processing and analysis by information processing equipment.

[0141] Among them, a microscope or other equipment can be used as an auxiliary device of the image acquisition device, and image information can be collected through a magnifying glass to obtain clearer image information of the machine-made sand surface. The specific image magnification can be 2 to 20 times, preferably 4 times.

[0142] You can also use infrared cameras to eliminate the influence of environment and light and obtain clear image information.

[0143] When identifying the stone powder content in stored machine-made sand, due to the large volume of sand accumulation, in order to make the calculation results more representative, image information sampling can be performed at multiple locations on the sand pile, and values ​​with a difference of more than 5% can be eliminated, and the remaining values ​​can be averaged.

[0144] When identifying the stone powder content in the artificial sand in the production line, an image acquisition device with a fixed position and angle can be added to the production line, and information processing equipment and a display can be set up. In this way, the stone powder content in each link of the artificial sand production line can be obtained intuitively and in real time.

[0145] From the above description, it can be seen that the non-contact intelligent real-time detection method for the content of artificial sand and stone powder provided in the application example of this application does not require the use of detection instruments and reagents for the test. It adopts non-contact image acquisition and artificial intelligence calculation methods. The image acquisition equipment is portable and the image information sampling is convenient. The content of artificial sand and stone powder can be obtained in real time. After collecting the image information of the surface of artificial sand, the target information can be obtained in real time. It can realize real-time adjustment in the production process of artificial sand and rapid verification of stone powder content in construction scenarios, saving time and effort. It does not rely on the experience of engineering personnel, eliminates human errors, and can better control the accuracy of the output results.

[0146] Therefore, the application example of this application provides a fast, convenient, and accurate non-contact intelligent real-time detection method for the content of artificial sand and stone powder based on artificial intelligence and image processing. After the image acquisition device acquires the image information of the artificial sand, the information processing device uses the calculation model derived from artificial intelligence to process the image information of the artificial sand, calculates and analyzes the content of the artificial sand and stone powder, and the display displays the content of the artificial sand and stone powder in real time. The application can be used to obtain the stone powder content of the artificial sand in real time and accurately, so as to help realize the real-time detection and adjustment of the stone powder content of the artificial sand during the production process, as well as the rapid detection and verification of the content of the artificial sand and stone powder by engineering personnel in the absence of detection instruments and experimental conditions, timely obtain the content of the artificial sand and stone powder, realize real-time data output, greatly improve the efficiency of the intelligent real-time detection of the content of the non-contact artificial sand and stone powder, reduce time and labor costs, avoid complex detection test procedures and manual errors, and facilitate the production and user to grasp and adjust the quality of the artificial sand.

[0147] From the software level, in order to solve the problems of the existing artificial sand and stone powder content testing method, such as the sample collection method requires manual participation, the detection method is complicated and time-consuming, and relies on manual experience and the accuracy of the detection equipment, the present application provides an embodiment of a non-contact artificial sand and stone powder content intelligent real-time detection device for executing all or part of the contents of the non-contact artificial sand and stone powder content intelligent real-time detection method, see Fig. 9 The non-contact intelligent real-time detection device for machine-made sand and gravel powder content specifically includes the following contents:

[0148] The target data acquisition module 10 is used to acquire target machine-made sand image data.

[0149] In the target data acquisition module 10, the stacked machine-made sand can be directly imaged at any angle and distance using an image acquisition device. The image acquisition device includes but is not limited to a mobile phone, a camera, a video camera, a camera, a monitoring device, and the like.

[0150] It can be understood that the target machine-made sand image data refers to the image of the machine-made sand pile in the target area, and the image can be collected from any angle.

[0151] The prediction model application module 20 is used to input the target machine-made sand image data into a trained stone powder content detection model, and determine the stone powder content detection result data in the machine-made sand corresponding to the target machine-made sand image data based on the output of the stone powder content detection model.

[0152] In the prediction model application module 20, the machine-made sand surface image information can be input into the model trained by machine learning, and the stone powder content of the machine-made sand can be automatically calculated. The stone powder content detection model is pre-trained based on a preset machine learning model, and the machine learning model can specifically be a deep learning model, such as a convolutional neural network.

[0153] Specifically, the test does not require the use of testing instruments and reagents, and adopts non-contact image acquisition and artificial intelligence calculation methods. The image acquisition equipment is portable and the image information sampling is convenient. The artificial sand and stone powder content can be obtained in real time. After collecting the image information of the artificial sand surface, the target information can be obtained in real time. It can realize real-time adjustment in the artificial sand production process and rapid verification of stone powder content in construction scenarios, saving time and effort. It does not rely on the experience of engineering personnel, eliminates human errors, and can better control the accuracy of the output results.

[0154] The embodiment of the non-contact intelligent real-time detection device for artificial sand and gravel powder content provided in the present application can be specifically used to execute the processing flow of the embodiment of the non-contact intelligent real-time detection method for artificial sand and gravel powder content in the above-mentioned embodiment. Its functions will not be repeated here, and reference can be made to the detailed description of the above-mentioned method embodiment.

[0155] From the above description, it can be seen that the non-contact intelligent real-time detection device for artificial sand and stone powder content provided in the embodiment of the present application can realize non-contact detection of artificial sand and stone powder content through image acquisition and artificial intelligence technology, and can effectively improve the intelligence, automation, convenience and efficiency of the artificial sand and stone powder content detection process, and can effectively improve the accuracy and real-time of the artificial sand and stone powder content detection results, thereby realizing real-time adjustment of artificial sand in the production process, and rapid verification work of engineering personnel in scenarios without detection instruments and experimental reagents, avoiding the problems of poor accuracy caused by unrepresentative sampling and large consumption of manpower and material resources required for repeated experiments, and is suitable for a variety of application scenarios. For example, in the artificial sand sales scenario, there is no need for detection personnel to be present, and the artificial sand detection results can be provided to buyers based on the artificial sand images collected non-contactly. For example, in the artificial sand production scenario, the production personnel only need to provide the production personnel with the adjustment of the stone powder content in the artificial sand or the quality inspection of the stone powder content based on the artificial sand images collected non-contactly. Provide artificial sand detection results.

[0156] In order to effectively improve the automation, intelligence and accuracy of the non-contact intelligent real-time detection results of the mechanism sand and stone powder content, in one embodiment of the non-contact intelligent real-time detection device for mechanism sand and stone powder content provided in the present application, see Fig.10 The non-contact intelligent real-time detection device for the content of machine-made sand and stone powder also specifically includes the following contents:

[0157] The historical data acquisition module 01 is used to acquire multiple historical machine-made sand image data and labels corresponding to each of the historical machine-made sand image data, wherein the collection area of ​​the historical machine-made sand image data is different from the target machine-made sand image data; and the label is used to represent the stone powder content data in the machine-made sand corresponding to the historical machine-made sand image data.

[0158] Specifically, when machine learning is performed, inputting historical machine-made sand image data plays a decisive role in the output stone powder content detection model. By using a large amount of image information to support the accuracy of the model, improving the quality and quantity of historical machine-made sand image data can directly improve the efficiency of machine learning. The atlas used for machine learning in this application is rich in samples, containing machine-made sand produced by various gradations, various stone powder contents, various parent rocks, and various crushers. The samples are more uniform, and the stone powder content detection model obtained after inputting it into the preset machine learning model for calculation has good robustness.

[0159] The prediction model training module 02 is used to train a preset machine learning model based on the labels corresponding to each of the historical machine-made sand image data to obtain a stone powder content detection model for detecting the stone powder content in the machine-made sand.

[0160] In order to effectively improve the application reliability and accuracy of historical machine-made sand image data, in one embodiment of the non-contact machine-made sand and stone powder content intelligent real-time detection device provided in the present application, the historical data acquisition module 01 in the non-contact machine-made sand and stone powder content intelligent real-time detection device is also specifically used to perform the following contents:

[0161] Step 011: Get the manufactured sand sample.

[0162] Step 012: The manufactured sand sample is doped with stone powder so that the mud content of the manufactured sand sample reaches a preset target mud content, and then an image acquisition device is used to acquire an image of the manufactured sand sample that has reached the target mud content to obtain corresponding original image data of the manufactured sand.

[0163] Step 013: extracting each historical machine-made sand image data from the machine-made sand original image data, and labeling each historical machine-made sand image data according to the target mud content to obtain a label corresponding to each historical machine-made sand image data.

[0164] In order to effectively improve the effectiveness and efficiency of preprocessing the original image of the machine-made sand surface, in one embodiment of the non-contact intelligent real-time detection device for machine-made sand and stone powder content provided in the present application, the historical data acquisition module 01 in the non-contact intelligent real-time detection device for machine-made sand and stone powder content is also specifically used to perform the following contents:

[0165] Step 0131: performing region extraction on each original image in the machine-made sand original image data to form at least one region map corresponding to each original image.

[0166] Step 0132: Perform image processing on each of the area maps to form sample maps corresponding to each of the area maps, wherein the image processing includes at least one of format conversion, size change and histogram equalization processing.

[0167] Step 0133: Store each of the sample images as historical machine-made sand image data.

[0168] Step 0134: Label each of the historical machine-made sand image data according to the target mud content to obtain a label corresponding to each of the historical machine-made sand image data.

[0169] Specifically, the original images in the pre-collected machine-made sand surface image atlas can be preprocessed to establish a dataset of machine-made sand feature images. The image preprocessing process mainly includes extracting the region of interest, converting the RGB image into a grayscale image, shrinking the transformation, histogram equalization, and adding labels. Image preprocessing, the original image data is input for preprocessing. One original image can generate multiple sample images of the same size with labels, which can be used for image data of machine learning models.

[0170] In order to effectively improve the application reliability and accuracy of the target machine-made sand image data, in one embodiment of the non-contact machine-made sand and stone powder content intelligent real-time detection device provided in the present application, the non-contact machine-made sand and stone powder content intelligent real-time detection device is also used to perform the following contents:

[0171] Step 101: amplifying the target machine-made sand image data based on a preset image magnification factor and collecting the amplified target machine-made sand image data.

[0172] The specific image magnification may be 2 to 20 times, preferably 4 times.

[0173] In order to further improve the accuracy of machine-made sand and stone powder content detection, in one embodiment of the non-contact machine-made sand and stone powder content intelligent real-time detection device provided in the present application, the prediction model application module 20 in the non-contact machine-made sand and stone powder content intelligent real-time detection device is specifically used to perform the following contents:

[0174] Step 210: inputting the plurality of target machine-made sand image data belonging to the same batch of target machine-made sand into the trained stone powder content detection model respectively, so as to obtain the stone powder content detection result data corresponding to each of the target machine-made sand image data respectively output by the stone powder content detection model;

[0175] Step 220: performing an average calculation process on the stone powder content detection result data corresponding to each of the target machine-made sand image data to obtain corresponding stone powder content detection average data;

[0176] Step 230: Using the stone powder content detection average data as the stone powder content detection result data of the target machine-made sand.

[0177] In order to effectively improve the accuracy and reliability of machine learning model training, in one embodiment of the non-contact intelligent real-time detection device for machine-made sand and stone powder content provided in the present application, the prediction model training module 02 in the non-contact intelligent real-time detection device for machine-made sand and stone powder content is also specifically used to perform the following contents:

[0178] Step 021: Divide the plurality of historical machine-made sand image data with the labels into a training set, a test set and a validation set respectively.

[0179] In one example, 60% of the historical machine-made sand image data with the labels can be randomly selected as a training set, another 20% of the historical machine-made sand image data with the labels can be used as a verification set, and the remaining 20% ​​of the historical machine-made sand image data with the labels can be used as a test set to test the accuracy of the machine learning model.

[0180] Step 022: Train a preset deep learning model based on the training set to obtain an initial model for detecting the stone powder content in machine-made sand.

[0181] Step 023: Apply the test set and the validation set to generate the stone powder content detection model according to the initial model.

[0182] In step 023, the data set is randomly divided into a training set, a validation set and a test set, and then the training set is used to train an initial model, and the validation set is used to validate the initial model. The initial model is continuously adjusted according to the situation, and the best model is selected. A final model is then trained using the training set and validation set data, and finally the final model is evaluated using the test set to form the final stone powder content detection model.

[0183] Specifically, during machine learning, inputting an atlas of machine-made sand surface images plays a decisive role in the output stone powder content detection model. By supporting the accuracy of the model with a large amount of image information, improving the quality and quantity of the atlas can directly improve the efficiency of machine learning. The atlas used for machine learning in this application is a unique machine-made sand atlas for this application. It has rich samples and contains machine-made sand produced by various gradations, various stone powder contents, various parent rocks, and various crushers. The samples are more uniform, and the stone powder content detection model obtained after inputting it into the machine learning model for calculation is more robust.

[0184] The process of building the model is as follows: In order to use machine learning to identify the stone powder content of artificial sand, the surface image of artificial sand with known characteristic values ​​is used as a sample input into the machine learning model, and the machine learning model is trained and verified.

[0185] The principle of implementing the stone powder content detection model is to input samples of known information into a preset machine learning model. Therefore, before training, the information of the image sample is first extracted and its features are predicted. Then, a relationship is established between it and the given known label. The error between the two is fed back to the model, and the machine learning model adjusts the parameters. Finally, the label of the sample can be identified, that is, the stone powder content of the sample is obtained.

[0186] In order to effectively improve the convenience and accuracy of acquiring the target machine-made sand image data, in one embodiment of the non-contact machine-made sand and stone powder content intelligent real-time detection device provided in the present application, the target data acquisition module 10 in the non-contact machine-made sand and stone powder content intelligent real-time detection device specifically includes the following contents:

[0187] Step 110: Receive a surface image of machine-made sand in a target area acquired by at least one image acquisition device.

[0188] Step 120: Pre-process the surface image to form at least one target manufactured sand image data.

[0189] Specifically, if the ratio of the environment occupied by artificial sand in the surface image is greater than or equal to a first threshold value, the surface image is respectively subjected to region extraction to form at least one target region map corresponding to the surface image; and, at least one of the target region maps is respectively subjected to image processing to form at least one target sample map corresponding to the surface image, so as to determine at least one of the target sample maps as target artificial sand image data.

[0190] In addition, if there are two or more image acquisition devices that capture the surface images of the same artificial sand pile in the target area, this problem can be solved by setting up two or more image acquisition devices. The image information of the same part captured by the two image acquisition devices can be compared to restore the actual size of the image information. Place a reference object of known size on the surface of the sand pile, such as a box, reference mark, etc. The information processing device can directly restore the surface information image of the artificial sand to its actual size through perspective transformation, affine transformation or other transformation methods. Automatically extract the area of ​​interest. After restoring the image, the information processing device will automatically extract the area of ​​interest in the image within the size reference frame as the basis for calculating the stone powder content.

[0191] From the hardware level, in order to solve the problems of the existing artificial sand and stone powder content testing method, such as the sample collection method requires manual participation, the detection method is complex and time-consuming, and relies on manual experience and the accuracy of the detection equipment, the present application provides an embodiment of an electronic device for realizing all or part of the contents of the non-contact artificial sand and stone powder content intelligent real-time detection method, and the electronic device specifically includes the following contents:

[0192] Fig.11 FIG. 9 is a schematic block diagram of the system structure of the electronic device 9600 according to an embodiment of the present application. Fig.11 As shown, the electronic device 9600 may include a central processor 9100 and a memory 9140; the memory 9140 is coupled to the central processor 9100. It is worth noting that Fig.11 is exemplary; other types of structures may also be used to supplement or replace this structure to implement telecommunication functions or other functions.

[0193] In one embodiment, the non-contact intelligent real-time detection function of the mechanical sand and stone powder content can be integrated into the central processing unit. The central processing unit can be configured to perform the following control:

[0194] Step 100: Acquire target machine-made sand image data.

[0195] In step 100, the stacked machine-made sand can be directly imaged at any angle and distance using an image acquisition device. The image acquisition device includes but is not limited to a mobile phone, a camera, a video camera, a camera, a monitoring device, and the like.

[0196] It can be understood that the target machine-made sand image data refers to the image of the machine-made sand pile in the target area, and the image can be collected from any angle.

[0197] Step 200: input the machine-made sand image data into a trained stone powder content detection model, and determine the stone powder content detection result data in the machine-made sand corresponding to the target machine-made sand image data based on the output of the stone powder content detection model.

[0198] In step 200, the surface image information of the machine-made sand can be input into a model trained by machine learning, and the stone powder content of the machine-made sand can be automatically calculated. The stone powder content detection model is pre-trained based on a preset machine learning model, and the machine learning model can specifically be a deep learning model, such as a convolutional neural network.

[0199] Specifically, the application of the solutions provided in the above steps 100 to 200 does not require the use of detection instruments and reagents for testing, and adopts non-contact image acquisition and artificial intelligence calculation methods. The image acquisition equipment is portable and the image information sampling is convenient. The artificial sand and stone powder content is obtained in real time. After collecting the image information of the artificial sand surface, the target information can be obtained in real time, which can realize real-time adjustment in the artificial sand production process and rapid verification of stone powder content in construction scenarios, saving time and effort. It does not rely on the experience of engineering personnel, eliminates human errors, and can better control the accuracy of the output results.

[0200] From the above description, it can be seen that the electronic device provided in the embodiment of the present application can realize non-contact detection of the content of artificial sand and stone powder through image acquisition and artificial intelligence technology, and can effectively improve the intelligence, automation, convenience and efficiency of the detection process of artificial sand and stone powder content, and can effectively improve the accuracy and real-time of the detection results of artificial sand and stone powder content, thereby realizing real-time adjustment of artificial sand in the production process, and rapid verification work of engineering personnel in scenarios without detection instruments and experimental reagents, avoiding the problems of poor accuracy caused by unrepresentative sampling and large consumption of manpower and material resources required for repeated experiments, and is suitable for a variety of application scenarios. For example, in the artificial sand sales scenario, there is no need for detection personnel to be present, and the artificial sand detection results can be provided to buyers based on the artificial sand images collected non-contactly. For example, in the artificial sand production scenario, the production personnel only need to provide the production personnel with the adjustment of the stone powder content in the artificial sand or the quality inspection of the stone powder content based on the artificial sand images collected non-contactly. Provide artificial sand detection results.

[0201] In another embodiment, the non-contact intelligent real-time detection device for machine-made sand and gravel powder content can be configured separately from the central processor 9100. For example, the non-contact intelligent real-time detection device for machine-made sand and gravel powder content can be configured as a chip connected to the central processor 9100, and the non-contact intelligent real-time detection function of machine-made sand and gravel powder content can be realized through the control of the central processor.

[0202] like Fig.11As shown, the electronic device 9600 may also include: a communication module 9110, an input unit 9120, an audio processor 9130, a display 9160, and a power supply 9170. It is worth noting that the electronic device 9600 does not necessarily have to include Fig.11 In addition, the electronic device 9600 may also include Fig.11 For components not shown, reference may be made to the prior art.

[0203] like Fig.11 As shown, the central processing unit 9100 is sometimes also referred to as a controller or an operation control, and may include a microprocessor or other processor device and / or logic device. The central processing unit 9100 receives input and controls the operation of various components of the electronic device 9600.

[0204] The memory 9140 may be, for example, one or more of a cache, a flash memory, a hard drive, a removable medium, a volatile memory, a non-volatile memory or other suitable devices. The above-mentioned information related to the failure may be stored, and a program for executing the relevant information may also be stored. The CPU 9100 may execute the program stored in the memory 9140 to implement information storage or processing, etc.

[0205] The input unit 9120 provides input to the central processing unit 9100. The input unit 9120 is, for example, a key or a touch input device. The power supply 9170 is used to provide power to the electronic device 9600. The display 9160 is used to display display objects such as images and texts. The display may be, for example, an LCD display, but is not limited thereto.

[0206] The memory 9140 may be a solid-state memory, such as a read-only memory (ROM), a random access memory (RAM), a SIM card, etc. It may also be a memory that saves information even when the power is off, can be selectively erased, and is provided with more data, examples of which are sometimes referred to as EPROMs, etc. The memory 9140 may also be some other type of device. The memory 9140 includes a buffer memory 9141 (sometimes referred to as a buffer). The memory 9140 may include an application / function storage unit 9142, which is used to store application programs and function programs or processes for executing the operation of the electronic device 9600 through the central processor 9100.

[0207] The memory 9140 may also include a data storage unit 9143 for storing data, such as contacts, digital data, pictures, sounds, and / or any other data used by the electronic device. The driver storage unit 9144 of the memory 9140 may include various drivers for communication functions of the electronic device and / or for executing other functions of the electronic device (such as messaging applications, address book applications, etc.).

[0208] The communication module 9110 is a transmitter / receiver 9110 that sends and receives signals via an antenna 9111. The communication module (transmitter / receiver) 9110 is coupled to the central processor 9100 to provide input signals and receive output signals, which may be the same as the case of a conventional mobile communication terminal.

[0209] Based on different communication technologies, multiple communication modules 9110 may be provided in the same electronic device, such as a cellular network module, a Bluetooth module and / or a wireless LAN module, etc. The communication module (transmitter / receiver) 9110 is also coupled to a speaker 9131 and a microphone 9132 via an audio processor 9130 to provide an audio output via the speaker 9131 and receive an audio input from the microphone 9132, thereby realizing a common telecommunication function. The audio processor 9130 may include any suitable buffer, decoder, amplifier, etc. In addition, the audio processor 9130 is also coupled to the central processor 9100, so that recording can be performed on the local machine through the microphone 9132, and the sound stored on the local machine can be played through the speaker 9131.

[0210] The embodiments of the present application also provide a computer-readable storage medium capable of implementing all the steps in the non-contact intelligent real-time detection method for the content of machine-made sand and stone powder in the above embodiments. The computer-readable storage medium stores a computer program. When the computer program is executed by a processor, all the steps of the non-contact intelligent real-time detection method for the content of machine-made sand and stone powder in the above embodiments are implemented. For example, when the processor executes the computer program, the following steps are implemented:

[0211] Step 100: Acquire target machine-made sand image data.

[0212] In step 100, the stacked machine-made sand can be directly imaged at any angle and distance using an image acquisition device. The image acquisition device includes but is not limited to a mobile phone, a camera, a video camera, a camera, a monitoring device, and the like.

[0213] It can be understood that the target machine-made sand image data refers to the image of the machine-made sand pile in the target area, and the image can be collected from any angle.

[0214] Step 200: input the target machine-made sand image data into a trained stone powder content detection model, and determine the stone powder content detection result data in the machine-made sand corresponding to the target machine-made sand image data based on the output of the stone powder content detection model.

[0215] In step 200, the surface image information of the machine-made sand can be input into a model trained by machine learning, and the stone powder content of the machine-made sand can be automatically calculated. The stone powder content detection model is pre-trained based on a preset machine learning model, and the machine learning model can specifically be a deep learning model, such as a convolutional neural network.

[0216] Specifically, the application of the solutions provided in the above steps 100 to 200 does not require the use of detection instruments and reagents for testing, and adopts non-contact image acquisition and artificial intelligence calculation methods. The image acquisition equipment is portable and the image information sampling is convenient. The artificial sand and stone powder content is obtained in real time. After collecting the image information of the artificial sand surface, the target information can be obtained in real time, which can realize real-time adjustment in the artificial sand production process and rapid verification of stone powder content in construction scenarios, saving time and effort. It does not rely on the experience of engineering personnel, eliminates human errors, and can better control the accuracy of the output results.

[0217] From the above description, it can be seen that the computer-readable storage medium provided in the embodiment of the present application can realize non-contact detection of the content of artificial sand and stone powder through image acquisition and artificial intelligence technology, and can effectively improve the intelligence, automation, convenience and efficiency of the detection process of artificial sand and stone powder content, and can effectively improve the accuracy and real-time of the detection results of artificial sand and stone powder content, thereby realizing real-time adjustment of artificial sand in the production process, and rapid verification work of engineering personnel in scenarios without detection instruments and experimental reagents, avoiding the problems of poor accuracy caused by unrepresentative sampling and large consumption of manpower and material resources required for repeated experiments, and is suitable for a variety of application scenarios. For example, in the artificial sand sales scenario, there is no need for detection personnel to be present, and the artificial sand detection results can be provided to buyers based on the artificial sand images collected non-contactly. For example, in the artificial sand production scenario, the production personnel only need to provide the production personnel with the adjustment of the stone powder content in the artificial sand or the quality inspection of the stone powder content based on the artificial sand images collected non-contactly. Provide artificial sand detection results.

[0218] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, devices, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0219] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (apparatus), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0220] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0221] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0222] The present invention uses specific embodiments to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A non-contact intelligent real-time detection method for the content of machine-made sand and gravel powder, characterized in that: include: Acquire target machine-made sand image data; Input the target machine-made sand image data into a trained stone powder content detection model, and determine the stone powder content detection result data in the machine-made sand corresponding to the target machine-made sand image data based on the output of the stone powder content detection model; Before inputting the target machine-made sand image data into the trained stone powder content detection model, the method further includes: Acquire multiple historical machine-made sand image data and labels corresponding to each of the historical machine-made sand image data, wherein the acquisition area of ​​the historical machine-made sand image data is different from that of the target machine-made sand image data; and the label is used to indicate the stone powder content data in the machine-made sand corresponding to the historical machine-made sand image data; Based on the labels corresponding to each of the historical machine-made sand image data, a preset machine learning model is trained to obtain a stone powder content detection model for detecting the stone powder content in the machine-made sand; The acquiring of a plurality of historical machine-made sand image data and labels corresponding to each of the historical machine-made sand image data comprises: Obtain samples of manufactured sand; The manufactured sand sample is doped with stone powder so that the mud content of the manufactured sand sample reaches a preset target mud content, and then an image acquisition device is used to acquire an image of the manufactured sand sample that reaches the target mud content to obtain corresponding original image data of the manufactured sand; Extracting each historical machine-made sand image data from the machine-made sand original image data, and labeling each historical machine-made sand image data according to the target mud content, so as to obtain a label corresponding to each historical machine-made sand image data; The extracting of each historical machine-made sand image data from the machine-made sand original image data, and labeling each historical machine-made sand image data according to the target mud content to obtain a label corresponding to each historical machine-made sand image data, includes: Performing region extraction on each original image in the machine-made sand original image data to form at least one region map corresponding to each original image; Performing image processing on each of the region images to form sample images corresponding to each of the region images, wherein the image processing includes: at least one of format conversion, size change and histogram equalization processing; storing each of the sample images as historical machine-made sand image data; Labeling each of the historical machine-made sand image data according to the target mud content to obtain a label corresponding to each of the historical machine-made sand image data; The step of obtaining target machine-made sand image data comprises: receiving a surface image of machine-made sand in a target area acquired by at least one image acquisition device; Preprocessing the surface image to form at least one target manufactured sand image data; If the ratio of the environment occupied by artificial sand in the surface image is greater than or equal to a first threshold value, the surface image is respectively subjected to region extraction to form at least one target region map corresponding to the surface image; and, at least one of the target region maps is respectively subjected to image processing to form at least one target sample map corresponding to the surface image, so as to determine at least one of the target sample maps as target artificial sand image data.

2. The non-contact intelligent real-time detection method for machine-made sand and stone powder content according to claim 1 is characterized in that: Before inputting the target machine-made sand image data into the trained stone powder content detection model, the method further includes: The target machine-made sand image data is magnified based on a preset image magnification factor and then the magnified target machine-made sand image data is collected.

3. The non-contact intelligent real-time detection method for machine-made sand and stone powder content according to claim 1 is characterized in that: The step of inputting the target machine-made sand image data into a trained stone powder content detection model, and determining the stone powder content detection result data in the machine-made sand corresponding to the target machine-made sand image data based on the output of the stone powder content detection model, comprises: Inputting the plurality of target machine-made sand image data belonging to the same batch of target machine-made sand into the trained stone powder content detection model respectively, so as to obtain the stone powder content detection result data corresponding to each of the target machine-made sand image data respectively output by the stone powder content detection model; The stone powder content detection result data corresponding to each of the target machine-made sand image data are averaged to obtain the corresponding stone powder content detection average data; The average data of the stone powder content detection is used as the stone powder content detection result data of the target machine-made sand.

4. The non-contact intelligent real-time detection method for machine-made sand and stone powder content according to claim 1 is characterized in that: The preset machine learning model is trained based on the labels corresponding to each of the historical machine-made sand image data to obtain a stone powder content detection model for detecting the stone powder content in the machine-made sand, including: Dividing the plurality of historical machine-made sand image data with the labels into a training set, a test set and a validation set respectively; Training a preset deep learning model based on the training set to obtain an initial model for detecting the stone powder content in machine-made sand; The test set and the validation set are applied to generate the stone powder content detection model according to the initial model.

5. A non-contact intelligent real-time detection device for machine-made sand and stone powder content, characterized in that: include: A target data acquisition module is used to acquire target machine-made sand image data; A prediction model application module is used to input the target machine-made sand image data into a trained stone powder content detection model, and determine the stone powder content detection result data in the machine-made sand corresponding to the target machine-made sand image data based on the output of the stone powder content detection model; Also includes: A historical data acquisition module, used to acquire a plurality of historical machine-made sand image data and labels corresponding to each of the historical machine-made sand image data, wherein the acquisition area of ​​the historical machine-made sand image data is different from that of the target machine-made sand image data; and the labels are used to indicate the stone powder content data in the machine-made sand corresponding to the historical machine-made sand image data; A prediction model training module, used to train a preset machine learning model based on the labels corresponding to each of the historical machine-made sand image data, so as to obtain a stone powder content detection model for detecting the stone powder content in the machine-made sand; The historical data acquisition module is used to perform the following: Obtain samples of manufactured sand; The manufactured sand sample is doped with stone powder so that the mud content of the manufactured sand sample reaches a preset target mud content, and then an image acquisition device is used to acquire an image of the manufactured sand sample that reaches the target mud content to obtain corresponding original image data of the manufactured sand; Extracting each historical machine-made sand image data from the machine-made sand original image data, and labeling each historical machine-made sand image data according to the target mud content, so as to obtain a label corresponding to each historical machine-made sand image data; The extracting of each historical machine-made sand image data from the machine-made sand original image data, and labeling each historical machine-made sand image data according to the target mud content to obtain a label corresponding to each historical machine-made sand image data, includes: Performing region extraction on each original image in the machine-made sand original image data to form at least one region map corresponding to each original image; Performing image processing on each of the region images to form sample images corresponding to each of the region images, wherein the image processing includes: at least one of format conversion, size change and histogram equalization processing; storing each of the sample images as historical machine-made sand image data; Labeling each of the historical machine-made sand image data according to the target mud content to obtain a label corresponding to each of the historical machine-made sand image data; The target data acquisition module is specifically used for: receiving a surface image of machine-made sand in a target area acquired by at least one image acquisition device; Preprocessing the surface image to form at least one target manufactured sand image data; If the ratio of the environment occupied by artificial sand in the surface image is greater than or equal to a first threshold value, the surface image is respectively subjected to region extraction to form at least one target region map corresponding to the surface image; and, at least one of the target region maps is respectively subjected to image processing to form at least one target sample map corresponding to the surface image, so as to determine at least one of the target sample maps as target artificial sand image data.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the non-contact intelligent real-time detection method for the content of machine-made sand and gravel powder as described in any one of claims 1 to 4 is implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the non-contact intelligent real-time detection method for the content of machine-made sand and gravel powder as described in any one of claims 1 to 4 is implemented.

Images