Method and device for intelligently preparing sand surface image-mud content relationship data set

Abstract

The present invention discloses a method and device for intelligently preparing a sand surface image-mud content relationship data set, the method comprising: obtaining a prepared sand sample; doping the prepared sand sample with stone powder so that the mud content of the prepared sand sample reaches a preset target mud content, and then performing image acquisition on the prepared sand sample that reaches the target mud content to obtain a plurality of sand surface image data; annotating the sand surface image data according to the target mud content to obtain a sand surface image data set. The present invention realizes efficient preparation of sand surface image data with various mud contents, which helps to improve the robustness and recognition accuracy of the mud content recognition model.

Description

Technical Field

[0001] The present invention relates to the technical field of machine learning training data set preparation, and in particular to a method and device for intelligently preparing a sand surface image-mud content relationship data set. Background Art

[0002] The particle characteristics of sand are important factors affecting the construction and working performance of concrete. Among them, the mud content of sand has a great influence on the performance and quality of concrete. Too high mud content can easily lead to honeycombing in concrete. Therefore, it is necessary to test the mud content of sand.

[0003] With the development of artificial intelligence (AI) technology, a method for detecting the mud content of sand through machine learning has been proposed. Machine learning is efficient and accurate, which can greatly improve the detection efficiency, realize non-contact real-time acquisition of sand characteristics, and reduce labor costs and the impact of human factors. Machine learning requires a large amount of input data as support. The sand surface image dataset plays a decisive role in the output calculation model. The richer the samples, the better the robustness of the machine learning calculation model, and it can also perform better on unfamiliar samples. Improving the quality and quantity of image datasets can directly improve the efficiency and accuracy of machine learning.

[0004] At present, the general process of making sand surface image datasets is: first find a suitable sand sample, detect the mud content of the sand sample, collect a large amount of sand surface image data, annotate the collected sand surface image data according to the mud content of the detected sand sample, and finally form a sand surface image dataset. However, the sand mud content detection process has strict regulations in the current specifications, and professional equipment is required to complete the relevant steps. There are many detection items and a large workload. The detection process is time-consuming and usually takes a day. This leads to the inefficiency of the current method of making sand surface image datasets. Summary of the invention

[0005] In order to solve the technical problem of low efficiency of the current method for preparing sand surface image data sets, the present invention proposes a method and device for intelligently preparing sand surface image-mud content relationship data sets.

[0006] In order to achieve the above object, according to one aspect of the present invention, a method for intelligently preparing a sand surface image-mud content relationship data set is provided, the method comprising:

[0007] Get samples of the produced sand;

[0008] The sand samples are doped with stone powder so that the mud content of the sand samples reaches a preset target mud content, and then the sand samples reaching the target mud content are imaged to obtain a plurality of sand surface image data;

[0009] The sand surface image data is annotated according to the target mud content to obtain a sand surface image data set.

[0010] Optionally, the target mud content is multiple; the mud content of the prepared sand sample is made to reach a preset target mud content by adding stone powder to the prepared sand sample, and then the prepared sand sample that reaches the target mud content is imaged to obtain multiple sand surface image data, including:

[0011] The doping and image acquisition processes are performed on each target mud content among multiple target mud contents in turn according to the numerical value, so as to obtain the sand surface image data corresponding to each target mud content, wherein, in each doping and image acquisition process, the mud content of the manufactured sand sample reaches the target mud content by doping the manufactured sand sample with stone powder, and then the image of the manufactured sand sample that reaches the target mud content is acquired.

[0012] Optionally, doping is performed on the manufactured sand sample when the doping and image acquisition process is performed for the first time, and doping is performed on the doped manufactured sand sample obtained by the previous doping and image acquisition process each time the doping and image acquisition process is performed thereafter.

[0013] Optionally, the prepared sand sample that reaches the target mud content is imaged to obtain a plurality of sand surface image data, including:

[0014] The surface characteristics of the sand sample prepared to achieve the target mud content are changed multiple times, and image acquisition is performed after each change of the surface characteristics to obtain multiple sand surface image data.

[0015] Optionally, the prepared sand sample that reaches the target mud content is imaged to obtain a plurality of sand surface image data, including:

[0016] Collecting image data of the sand sample that has reached the target mud content in a sampler, wherein the sampler vibrates according to a preset frequency to change the surface characteristics of the sand sample that has reached the target mud content;

[0017] A plurality of sand surface image data are generated according to the collected image data.

[0018] Optionally, the collecting of image data of the prepared sand sample that reaches the target mud content in a sampler includes:

[0019] Collecting video image data of the sand sample that reaches the target mud content in the sampler by means of a video acquisition device;

[0020] The method of generating a plurality of sand surface image data according to the collected image data comprises:

[0021] A plurality of sand surface image data are generated by extracting frames from the video image data, wherein the frequency of the extraction of frames is the same as the vibration frequency of the sampler.

[0022] Optionally, the collecting of image data of the prepared sand sample that reaches the target mud content in a sampler includes:

[0023] The image data of the sand sample that reaches the target mud content in the sampler is collected by continuous photography, wherein the frequency of continuous photography is the same as the vibration frequency of the sampler.

[0024] Optionally, the target mud content is a plurality of target mud contents generated according to a preset mud content range and a preset gradient.

[0025] Optionally, the method for intelligently preparing a sand surface image-mud content relationship data set further includes:

[0026] The initial sand sample is screened, washed and dried in sequence to obtain the prepared sand sample.

[0027] Optionally, the mud content of the prepared sand sample is 0.

[0028] Optionally, the produced sand samples are of multiple types, including natural sand, machine-made sand and a mixture of natural sand and machine-made sand.

[0029] Optionally, the method for intelligently preparing a sand surface image-mud content relationship data set further includes:

[0030] A mud content recognition model is trained based on the sand surface image data set using a preset machine learning model, so as to identify the mud content of the sand sample based on the mud content recognition model.

[0031] In order to achieve the above object, according to another aspect of the present invention, a device for intelligently preparing a sand surface image-mud content relationship data set is provided, the device comprising:

[0032] A sand sample obtaining unit, used for obtaining the prepared sand sample;

[0033] A doping and image acquisition unit, configured to dope the prepared sand sample with stone powder so that the mud content of the prepared sand sample reaches a preset target mud content, and then perform image acquisition on the prepared sand sample that reaches the target mud content to obtain a plurality of sand surface image data;

[0034] The data annotation unit is used to annotate the sand surface image data according to the target mud content to obtain a sand surface image data set.

[0035] The beneficial effects of the present invention are as follows: the embodiment of the present invention adds stone powder to the manufactured sand sample so that the mud content of the manufactured sand sample reaches a preset target mud content, and then collects images of the manufactured sand sample that reaches the target mud content to obtain a plurality of sand surface image data, and finally annotates the sand surface image data according to the target mud content to obtain a sand surface image data set, thereby achieving the beneficial effect of efficiently producing sand surface image data with a variety of mud contents, and helping to improve the robustness and recognition accuracy of the mud content recognition model. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0037] Figure 1 It is a flow chart of a method for intelligently preparing a sand surface image-mud content relationship data set according to an embodiment of the present invention;

[0038] Figure 2 is a flow chart of collecting sand surface image data according to an embodiment of the present invention;

[0039] Figure 3 is a flow chart of collecting sand surface image data according to another embodiment of the present invention;

[0040] Figure 4 It is a structural block diagram of a device for intelligently preparing a sand surface image-mud content relationship data set according to an embodiment of the present invention. DETAILED DESCRIPTION

[0041] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0042] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, 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.

[0043] It should be noted that the terms "including" and "having" and any variations thereof in the specification and claims of the present invention and the above-mentioned drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or apparatus comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or apparatus.

[0044] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0045] It should be noted that the mud content of the present invention can also be called powder content in the art. Mud content is generally used to refer to particles smaller than 0.075 mm in natural sand, and powder content refers to particles smaller than 0.075 mm in machine-made sand. The two have the same meaning. The present invention uses mud content to refer to both of them, that is, powder content also belongs to the protection scope of the present invention and should also be protected.

[0046] Figure 1 Flow chart of a method for intelligently preparing a sand surface image-mud content relationship data set according to an embodiment of the present invention. Figure 1 As shown, the method for intelligently preparing a sand surface image-mud content relationship data set in this embodiment includes steps S101 to S103.

[0047] Step S101, obtaining a prepared sand sample.

[0048] In one embodiment of the present invention, the prepared sand sample is obtained by sieving, washing and drying the initial sand sample in sequence. Specifically, the initial sand sample can be sieved through a standard sand and gravel sieve, the sieved stone powder is dried and reserved for standby use, and then the sieved initial sand sample is rinsed with clean water to remove the stone powder attached to the surface of the sand particles, and then dried to obtain the prepared sand sample. In one embodiment of the present invention, the initial sand sample can be natural sand, machine-made sand produced by different parent rocks, machine-made sand mixed with multiple contents of the same original rock powder, machine-made sand mixed with multiple contents of original rock powder, and river sand mixed with multiple contents of stone powder, which can be obtained by direct purchase, and the mud content of the initial sand sample purchased is usually less than 10%. The initial sand sample can be a variety of gradations, including fine sand (fineness modulus 1.8-2.2), medium sand (fineness modulus 2.3-2.7), and coarse sand (fineness modulus 2.8-3.3).

[0049] In one embodiment of the present invention, the mud content of the prepared sand sample is 0.

[0050] In one embodiment of the present invention, the manufactured sand samples are multiple types, and the manufactured sand samples include: natural sand, machine-made sand, and a mixture of natural sand and machine-made sand.

[0051] Step S102, adding stone powder to the prepared sand sample to make the mud content of the prepared sand sample reach a preset target mud content, and then collecting images of the prepared sand sample that reaches the target mud content to obtain a plurality of sand surface image data.

[0052] In the embodiment of the present invention, in order to enrich the sample image data set, the present invention sets a plurality of target mud contents, and then generates a large amount of sand surface image data for each target mud content.

[0053] In one embodiment of the present invention, the target mud content is a plurality of target mud contents generated according to a preset mud content range and a preset gradient. The preset mud content range may be 0% to 30%, and the preset gradient may be 0.01%.

[0054] In one aspect of the present invention, this step may specifically include: changing the surface characteristics of the sand sample that has reached the target mud content for multiple times, and performing image acquisition after each change of the surface characteristics to obtain multiple sand surface image data. Specifically, the spatial structure of the sand particles may be randomly changed by manually stirring the sand sample or stirring the sand sample with a tool, that is, randomly changing the surface characteristics of the sand sample.

[0055] In one embodiment of the present invention, this step may be performed to collect images at different shooting angles and distances, and multiple groups of sand sample image data may be obtained based on the same sand sample.

[0056] In one embodiment of the present invention, this step can also adjust the pixels and size of the collected image to generate more sand surface image data. Specifically, the acquired image data can be cropped according to a fixed size, so that multiple copies of sand sample image data that meet the requirements can be output from one picture, effectively expanding the sample size of the mine image data set.

[0057] In one embodiment of the present invention, the surface of the sand sample is optically magnified before the image is collected, with the magnification being 2-20 times, preferably 4 times. And the magnification should be consistent for the same data set.

[0058] In one embodiment of the present invention, the stone powder can be any existing stone powder.

[0059] Step S103: annotating the sand surface image data according to the target mud content to obtain a sand surface image data set.

[0060] In an embodiment of the present invention, after obtaining the sand surface image data corresponding to the target mud content, the target mud content is marked as a label on the sand surface image data, and finally a sand surface image data set is formed. The finally formed sand surface image data set may include sand surface image data of multiple types of sand and sand surface image data corresponding to multiple mud contents.

[0061] After obtaining the sand surface image data set, the present invention can also train a mud content recognition model based on the sand surface image data set using a preset machine learning model to identify the mud content of the sand sample based on the mud content recognition model. Since the sand surface image data set obtained by the present invention contains sand surface image data of various mud contents, it is helpful to improve the recognition accuracy of the mud content recognition model.

[0062] In one embodiment of the present invention, the target mud content is multiple. The above step S102 is specifically as follows:

[0063] The doping and image acquisition processes are performed on each target mud content among multiple target mud contents in turn according to the numerical value, so as to obtain the sand surface image data corresponding to each target mud content, wherein, in each doping and image acquisition process, the mud content of the manufactured sand sample reaches the target mud content by doping the manufactured sand sample with stone powder, and then the image of the manufactured sand sample that reaches the target mud content is acquired.

[0064] In one embodiment of the present invention, doping is performed on the manufactured sand sample when the doping and image acquisition process is performed for the first time, and doping is performed on the doped manufactured sand sample obtained by the previous doping and image acquisition process each time the doping and image acquisition process is performed thereafter.

[0065] The present invention performs doping and image acquisition processes for each target mud content in a plurality of target mud contents in sequence according to the numerical values. When doping is performed, the prepared sand sample is first doped to the minimum target mud content, and then the prepared sand sample is doped to the second minimum target mud content based on the last doping, and the process continues until the prepared sand sample is doped to the maximum target mud content. Image acquisition is performed each time the doping reaches the target mud content, and a plurality of sand sample image data are obtained.

[0066] Figure 2 FIG. 1 is a flow chart of collecting sand surface image data according to an embodiment of the present invention. Figure 2 The above step S102 of collecting images of the prepared sand sample that reaches the target mud content to obtain a plurality of sand surface image data specifically includes step S201 and step S202.

[0067] Step S201, collecting image data of the manufactured sand sample that reaches the target mud content in a sampler, wherein the sampler vibrates according to a preset frequency to change the surface characteristics of the manufactured sand sample that reaches the target mud content.

[0068] In an embodiment of the present invention, this step can sample the sand sample and put it into a sampler that can vibrate at a set frequency. The sampler vibrates according to the preset frequency to change the surface characteristics of the sand sample, and then the sand sample in the sampler is imaged to obtain multiple image data. In an optional embodiment of the present invention, the sampler can vibrate at a frequency of 5 times per second.

[0069] Step S202: Generate a plurality of sand surface image data according to the collected image data.

[0070] In the embodiment of the present invention, the sand sample in the sampler can be directly doped so that the sand sample reaches a plurality of preset target mud contents in sequence, and image data of the sand sample vibrating in the sampler is collected after each doping. The specific doping process can be seen in the above embodiment.

[0071] Figure 3 FIG. 1 is a flow chart of another embodiment of the present invention for collecting sand surface image data. Figure 3 The above step S102 of collecting images of the sand sample that has reached the target mud content to obtain a plurality of sand surface image data specifically includes step S301 and step S302.

[0072] Step S301, collecting video image data of the sand sample that has reached the target mud content in a sampler through a video acquisition device.

[0073] In the present invention, due to the vibration of the sampler, the spatial structure of sand particles and stone powder is constantly changing, and hundreds of picture image information can be output through a piece of video image information.

[0074] Step S302, generating a plurality of sand surface image data by extracting frames from the video image data, wherein the frequency of the extraction of frames is the same as the vibration frequency of the sampler.

[0075] In one embodiment of the present invention, the image data of the sand sample produced to reach the target mud content in the sampler in step S201 may be specifically:

[0076] The image data of the sand sample that reaches the target mud content in the sampler is collected by continuous photography, wherein the frequency of continuous photography is the same as the vibration frequency of the sampler.

[0077] It can be seen from the above embodiments that the method for intelligently preparing a sand surface image-mud content relationship data set of the present invention achieves at least the following beneficial effects:

[0078] 1. The sample size is rich, and the model obtained by machine learning training has good robustness. The method of the present invention can obtain image data of sand samples with different mud contents. The target mud content of the present invention is between 0% and 30% with a gradient of 0.01%, and the composed image data set samples are more uniform. After inputting it into the machine learning calculation, the obtained calculation model has good robustness, can efficiently identify the mud content, and has a high recognition accuracy. It can also achieve 0.01% mud content recognition with high recognition accuracy.

[0079] 2. The present invention first mixes the sand sample with the target mud content and then collects the image. Finally, the target mud content can be directly marked as a label, avoiding the time-consuming and laborious problem of the prior art of detecting the mud content of the sand sample after collecting the image, and effectively improving the efficiency of establishing the image data set.

[0080] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0081] Based on the same inventive concept, an embodiment of the present invention also provides a device for intelligently making a sand surface image-mud content relationship data set, which can be used to implement the method for intelligently making a sand surface image-mud content relationship data set described in the above embodiment, as described in the following embodiment. Since the principle of solving the problem by the device for intelligently making a sand surface image-mud content relationship data set is similar to the method for intelligently making a sand surface image-mud content relationship data set, the embodiment of the device for intelligently making a sand surface image-mud content relationship data set can refer to the embodiment of the method for intelligently making a sand surface image-mud content relationship data set, and the repeated parts will not be repeated. As used below, the term "unit" or "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceived.

[0082] Figure 4 is a structural block diagram of a device for intelligently preparing a sand surface image-mud content relationship data set according to an embodiment of the present invention, such as Figure 4 As shown, the device for intelligently preparing a sand surface image-mud content relationship data set according to an embodiment of the present invention includes:

[0083] A sand sample obtaining unit 1 is used to obtain the prepared sand sample;

[0084] A doping and image acquisition unit 2 is used to add stone powder to the prepared sand sample so that the mud content of the prepared sand sample reaches a preset target mud content, and then perform image acquisition on the prepared sand sample that reaches the target mud content to obtain a plurality of sand surface image data;

[0085] The data annotation unit 3 is used to annotate the sand surface image data according to the target mud content to obtain a sand surface image data set.

[0086] In one embodiment of the present invention, the target mud content is multiple;

[0087] The doping and image acquisition unit 2 is specifically used to perform a doping and image acquisition process on each target mud content among multiple target mud contents in turn according to the numerical value, and obtain sand surface image data corresponding to each target mud content, wherein, each time the doping and image acquisition process is performed, the mud content of the manufactured sand sample reaches the target mud content by doping the manufactured sand sample with stone powder, and then image acquisition is performed on the manufactured sand sample that reaches the target mud content.

[0088] In one embodiment of the present invention, the doping and image acquisition unit 2 dopes the manufactured sand sample when the doping and image acquisition process is performed for the first time, and dopes the manufactured sand sample obtained by the previous doping and image acquisition process each time the doping and image acquisition process is performed thereafter.

[0089] In one embodiment of the present invention, the doping and image acquisition unit 2 is specifically used to change the surface characteristics of the sand sample produced to achieve the target mud content multiple times, and perform image acquisition after each change of the surface characteristics to obtain multiple sand surface image data.

[0090] In one embodiment of the present invention, the doping and image acquisition unit 2 includes:

[0091] An image acquisition module, used for acquiring image data of the sand sample produced to reach the target mud content in a sampler, wherein the sampler vibrates according to a preset frequency to change the surface characteristics of the sand sample produced to reach the target mud content;

[0092] The image generation module is used to generate a plurality of sand surface image data according to the collected image data.

[0093] In one embodiment of the present invention, the image acquisition module is specifically used to collect the video image data of the sand sample that has reached the target mud content in the sampler through a video acquisition device; the image acquisition module is specifically used to generate a plurality of sand surface image data by extracting frames from the video image data, wherein the frequency of the extraction is the same as the vibration frequency of the sampler.

[0094] In one embodiment of the present invention, the image acquisition module is specifically used to collect image data of the sand sample produced to achieve the target mud content in the sampler by continuous photography, wherein the frequency of continuous photography is the same as the vibration frequency of the sampler.

[0095] In one embodiment of the present invention, the target mud content is a plurality of target mud contents generated according to a preset mud content range and a preset gradient.

[0096] In one embodiment of the present invention, the device for intelligently preparing a sand surface image-mud content relationship data set further includes:

[0097] The sand sample processing unit is used to screen, rinse and dry the initial sand sample in sequence to obtain the prepared sand sample.

[0098] In one embodiment of the present invention, the mud content of the prepared sand sample is 0.

[0099] In one embodiment of the present invention, the manufactured sand samples are multiple types, and the manufactured sand samples include: natural sand, machine-made sand, and a mixture of natural sand and machine-made sand.

[0100] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for intelligently preparing a sand surface image-mud content relationship data set, characterized in that: include: Get samples of the produced sand; The sand samples are doped with stone powder so that the mud content of the sand samples reaches a preset target mud content, and then the sand samples reaching the target mud content are imaged to obtain a plurality of sand surface image data; Annotating the sand surface image data according to the target mud content to obtain a sand surface image data set; The target mud content is multiple; the mud content of the sand sample is made to reach a preset target mud content by adding stone powder to the sand sample, and then the sand sample reaching the target mud content is imaged to obtain multiple sand surface image data, including: Performing a doping and image acquisition process on each target mud content among a plurality of target mud contents in turn according to the numerical value, and obtaining sand surface image data corresponding to each target mud content, wherein, each time the doping and image acquisition process is performed, the mud content of the manufactured sand sample reaches the target mud content by doping the manufactured sand sample with stone powder, and then performing image acquisition on the manufactured sand sample that reaches the target mud content, doping is performed on the manufactured sand sample when the doping and image acquisition process is performed for the first time, and doping is performed on the manufactured sand sample obtained by the previous doping and image acquisition process each time the doping and image acquisition process is performed thereafter; The step of collecting images of the sand sample produced to reach the target mud content to obtain a plurality of sand surface image data includes: The surface characteristics of the sand sample prepared to achieve the target mud content are changed multiple times, and image acquisition is performed after each change of the surface characteristics to obtain multiple sand surface image data.

2. The method for intelligently preparing a sand surface image-mud content relationship data set according to claim 1, characterized in that: The prepared sand sample having the target mud content is imaged to obtain a plurality of sand surface image data, including: Collecting image data of the sand sample that has reached the target mud content in a sampler, wherein the sampler vibrates according to a preset frequency to change the surface characteristics of the sand sample that has reached the target mud content; A plurality of sand surface image data are generated according to the collected image data.

3. The method for intelligently preparing a sand surface image-mud content relationship data set according to claim 2, characterized in that: The collecting of the image data of the sand sample produced to reach the target mud content in the sampler comprises: Collecting video image data of the sand sample that reaches the target mud content in the sampler by means of a video acquisition device; The method of generating a plurality of sand surface image data according to the collected image data comprises: A plurality of sand surface image data are generated by extracting frames from the video image data, wherein the frequency of the extraction of frames is the same as the vibration frequency of the sampler.

4. The method for intelligently preparing a sand surface image-mud content relationship data set according to claim 2, characterized in that: The collecting of the image data of the sand sample produced to reach the target mud content in the sampler comprises: The image data of the sand sample that reaches the target mud content in the sampler is collected by continuous photography, wherein the frequency of continuous photography is the same as the vibration frequency of the sampler.

5. The method for intelligently preparing a sand surface image-mud content relationship data set according to claim 1, characterized in that: The target mud content is a plurality of target mud contents generated according to a preset mud content range and a preset gradient.

6. The method for intelligently preparing a sand surface image-mud content relationship data set according to claim 1, characterized in that: Also includes: The initial sand sample is screened, washed and dried in sequence to obtain the prepared sand sample.

7. The method for intelligently preparing a sand surface image-mud content relationship data set according to claim 3, characterized in that: The mud content of the prepared sand sample is 0.

8. The method for intelligently preparing a sand surface image-mud content relationship data set according to claim 1, characterized in that: There are multiple types of sand samples, including natural sand, machine-made sand and a mixture of natural sand and machine-made sand.

9. The method for intelligently preparing a sand surface image-mud content relationship data set according to claim 1, characterized in that: Before collecting images of the prepared sand sample that reaches the target mud content, the method further includes: The surface of the prepared sand sample that reaches the target mud content is optically magnified at a magnification of 2 to 20 times, and the magnification is consistent for the same sand surface image data set.

10. The method for intelligently preparing a sand surface image-mud content relationship data set according to claim 1, characterized in that: Also includes: A mud content recognition model is trained based on the sand surface image data set using a preset machine learning model, so as to identify the mud content of the sand sample based on the mud content recognition model.

11. A device for intelligently preparing a sand surface image-mud content relationship data set, characterized in that: include: A sand sample obtaining unit, used for obtaining the prepared sand sample; A doping and image acquisition unit, configured to dope the prepared sand sample with stone powder so that the mud content of the prepared sand sample reaches a preset target mud content, and then perform image acquisition on the prepared sand sample that reaches the target mud content to obtain a plurality of sand surface image data; A data annotation unit, used for annotating the sand surface image data according to a target mud content to obtain a sand surface image data set; There are multiple target mud contents; the doping and image acquisition unit is specifically used to perform doping and image acquisition processes on each target mud content in the multiple target mud contents in sequence according to the numerical values, and obtain sand surface image data corresponding to each target mud content, wherein, each time the doping and image acquisition process is performed, the mud content of the manufactured sand sample reaches the target mud content by doping the manufactured sand sample with stone powder, and then the image of the manufactured sand sample that reaches the target mud content is acquired; The doping and image acquisition unit dopes the prepared sand sample when the doping and image acquisition process is performed for the first time, and dopes the prepared sand sample obtained by the previous doping and image acquisition process each time the doping and image acquisition process is performed thereafter; The doping and image acquisition unit is specifically used to change the surface characteristics of the sand sample produced to achieve the target mud content for multiple times, and to perform image acquisition after each change of the surface characteristics to obtain multiple sand surface image data.

Images