Soil ecological restoration detection method and system based on artificial intelligence

Abstract

The invention provides a soil ecological restoration detection method and system based on artificial intelligence, and the method comprises the steps: optimizing the network parameters of a first soil restoration state detection network and a second soil restoration state detection network according to a comparison result, and obtaining the trained first soil restoration state detection network and second soil restoration state detection network; and receiving to-be-processed soil ecological restoration data, extracting feature information of the to-be-processed soil ecological restoration data through the trained first soil restoration state detection network, and carrying out soil ecological restoration data detection to obtain a detection result. Multi-label training is performed on the two networks through combination of contained labels between soil object labels and soil ecological restoration data with similar soil ecological restoration data characteristics and training of the soil ecological restoration data, network parameters of the two networks are optimized according to a comparison result, and the working efficiency of the trained networks is improved. Therefore, the information detection precision is improved, and the reliability of ecological restoration is guaranteed.

Description

Technical Field

[0001] This application relates to the technical field of soil data detection. Specifically, it relates to a method and system for detecting soil ecological restoration based on artificial intelligence. Background Art

[0002] Soil refers to a layer of loose material on the earth's surface, composed of various granular minerals, organic matter, moisture, air, microorganisms, etc., and can grow plants. Soil is composed of minerals weathered from rocks, organic matter produced by the decomposition of animal and plant residues and microorganisms, soil organisms (solid-phase substances), as well as moisture (liquid-phase substances), air (gas-phase substances), oxidized humus, etc.

[0003] Solid substances include soil minerals, organic matter, and nutrients obtained after irradiating bacteria and sterilizing microorganisms by light. Liquid substances mainly refer to soil moisture. Gas is the air existing in soil pores. These three types of substances in the soil form a contradictory unity. They are interconnected and restrict each other, providing essential living conditions for crops and being the material basis of soil fertility.

[0004] In the actual process, after soil remediation, the information of the remediated soil needs to be brought back to the laboratory for detection, and various information of the soil needs to be detected. Therefore, a large amount of time and labor costs are required, and the detection cycle of the soil is very long and may be inaccurate (the reason for inaccuracy is that during the process of bringing the soil back to the laboratory, the water content and microorganisms of the soil are difficult to accurately detect). Therefore, there is an urgent need for a technical solution to improve the above technical problems. Summary of the Invention

[0005] To improve the technical problems existing in the related art, this application provides a method and system for detecting soil ecological restoration based on artificial intelligence.

[0006] In a first aspect, a method for detecting soil ecological restoration based on artificial intelligence is provided. The method includes: loading training soil ecological restoration data, soil object labels, and inclusion labels into a first soil restoration state detection network and a second soil restoration state detection network for training, where the inclusion labels are used to calibrate training soil ecological restoration data with a soil ecological restoration data feature similarity greater than a preset specified value; obtaining a comparison result between the predictions of the first soil restoration state detection network and the second soil restoration state detection network; optimizing the network parameters of the first soil restoration state detection network and the second soil restoration state detection network in combination with the comparison result to obtain the trained first soil restoration state detection network and the second soil restoration state detection network; where the second soil restoration state detection network is used to load important soil restoration indicators for feature extraction by the second soil restoration state detection network into the first soil restoration state detection network to achieve mutual feature extraction learning between the first soil restoration state detection network and the second soil restoration state detection network, obtaining the trained first soil restoration state detection network and the second soil restoration state detection network, so that the trained first soil restoration state detection network has the important soil restoration indicators for feature extraction of the second soil restoration state detection network; receiving soil ecological restoration data to be processed, and extracting feature information of the soil ecological restoration data to be processed through the trained first soil restoration state detection network for soil ecological restoration data detection to obtain a detection result.

[0007] Further, the method for detecting soil ecological restoration based on artificial intelligence further includes: obtaining training soil ecological restoration data, where the training soil ecological restoration data has soil object labels; determining preset soil ecological restoration data corresponding to each soil object label type; classifying the preset soil ecological restoration data corresponding to different soil object label types, and classifying the preset soil ecological restoration data with a soil ecological restoration data feature similarity greater than a preset specified value; generating inclusion labels for the classified preset soil ecological restoration data.

[0008] Further, the step of classifying the preset soil ecological restoration data corresponding to different soil object label types and classifying the preset soil ecological restoration data with a soil ecological restoration data feature similarity greater than a preset specified value includes: extracting soil ecological restoration data features of the preset soil ecological restoration data corresponding to each soil object label type; classifying the soil ecological restoration data features of the preset soil ecological restoration data corresponding to each soil object label type through a classification method; and classifying the preset soil ecological restoration data with a soil ecological restoration data feature similarity greater than a preset specified value.

[0009] Further, the step of loading the training soil ecological restoration data, the soil object labels, and the inclusion labels into the first soil restoration status detection network and the second soil restoration status detection network for training includes: optimizing the training soil ecological restoration data; loading the optimized training soil ecological restoration data, the soil object labels, and the inclusion labels into the first soil restoration status detection network and the second soil restoration status detection network for training respectively.

[0010] Further, the step of obtaining the comparison result between the predictions of the first soil restoration status detection network and the second soil restoration status detection network includes: calculating the comparison result between the predictions of the first soil restoration status detection network and the second soil restoration status detection network through a comparison result calculation strategy, and constructing a quantization evaluation thread corresponding to the comparison result.

[0011] Further, the step of optimizing the network parameters of the first soil restoration status detection network and the second soil restoration status detection network by combining the comparison result to obtain the trained first soil restoration status detection network and the second soil restoration status detection network includes: splicing and training the quantization evaluation thread with the initial loss value thread of the first soil restoration status detection network to obtain the trained first soil restoration status detection network; splicing and training the quantization evaluation thread with the initial loss value thread of the second soil restoration status detection network to obtain the trained second soil restoration status detection network.

[0012] Further, the step of receiving the soil ecological restoration data to be processed, extracting the feature information of the soil ecological restoration data to be processed through the trained first soil restoration status detection network for soil ecological restoration data detection, and obtaining the detection result includes: receiving the predicted soil ecological restoration data, and identifying the data to be processed in the predicted soil ecological restoration data through the target parsing network; reading the data to be processed to generate the soil ecological restoration data to be processed; loading the soil ecological restoration data to be processed into the trained first soil restoration status detection network, extracting the feature information after global convolution for soil ecological restoration data detection, and obtaining the detection result.

[0013] Further, the step of extracting the feature information after global convolution for soil ecological restoration data detection to obtain a detection result includes: extracting the feature information of a preset layer after global convolution; determining multiple directories to which the feature information belongs; obtaining a preset number of target soil ecological restoration data under each directory sorted in ascending order of similarity to the feature information; respectively loading the preset number of target soil ecological restoration data under each directory into the trained first soil restoration status detection network, and extracting the feature information of the preset layer after global convolution of the preset number of target soil ecological restoration data under each directory; calculating the splicing value of the discrimination data between the feature information of each target soil ecological restoration data and the feature information of the soil ecological restoration data to be processed in the same directory; taking the directory with the smallest splicing value as the target directory, and obtaining the target soil ecological restoration data with the smallest discrimination data under the target directory as the soil ecological restoration data detection result.

[0014] Further, the step of calculating the splicing value of the discrimination data between the feature information of each target soil ecological restoration data and the feature information of the soil ecological restoration data to be processed in the same directory includes: calculating the discrimination data between the feature information of each target soil ecological restoration data and the feature information of the soil ecological restoration data to be processed in the same directory; sorting each target soil ecological restoration data in the same directory in descending order of discrimination data; performing weighted processing on the discrimination data of each target soil ecological restoration data in the same directory according to the sorting situation; calculating the splicing value of the discrimination data of each target soil ecological restoration data in the same directory after weighted processing.

[0015] In a second aspect, a soil ecological restoration detection system based on artificial intelligence is provided, including a processor and a memory that communicate with each other, and the processor is configured to read and execute a computer program from the memory to implement the above method.

[0016] The soil ecological restoration detection method and system based on artificial intelligence provided by the embodiments of the present application train by loading training soil ecological restoration data, soil object tags, and inclusion tags into a first soil restoration status detection network and a second soil restoration status detection network respectively; obtain the comparison result between the predictions of the first soil restoration status detection network and the second soil restoration status detection network; optimize the network parameters of the first soil restoration status detection network and the second soil restoration status detection network according to the comparison result to obtain the trained first soil restoration status detection network and the second soil restoration status detection network; receive the soil ecological restoration data to be processed, extract the feature information of the soil ecological restoration data to be processed through the trained first soil restoration status detection network for soil ecological restoration data detection, and obtain the detection result. Therefore, through the inclusion tags between the soil ecological restoration data with similar soil object tags and soil ecological restoration data features, the training soil ecological restoration data is used to perform multi-label training on the two networks, and the network parameters of the two networks are optimized according to the comparison result, improving the working efficiency of the trained networks, thereby improving the accuracy of information detection and ensuring the reliability of ecological restoration. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings required in the embodiments. It should be understood that the following drawings only show some embodiments of the present application and should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0018] Figure 1 It is a flowchart of a soil ecological restoration detection method based on artificial intelligence provided by the embodiments of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] In order to better understand the above technical solutions, the following will make a detailed description of the technical solutions of the present application through the drawings and specific embodiments. It should be understood that the embodiments of the present application and the specific features in the embodiments are detailed descriptions of the technical solutions of the present application, rather than limitations on the technical solutions of the present application. Without conflict, the technical features in the embodiments of the present application and the embodiments can be combined with each other.

[0020] Please refer to Figure 1 , which shows a soil ecological restoration detection method based on artificial intelligence. The method may include the technical solutions described in the following steps 101-104.

[0021] In step 101, the training soil ecological restoration data, the soil object labels, and the inclusion labels are respectively loaded into the first soil restoration status detection network and the second soil restoration status detection network for training. The inclusion labels are used to calibrate the training soil ecological restoration data whose similarity of soil ecological restoration data features is greater than a preset specified value.

[0022] Furthermore, the soil object label can be understood as the soil identifier of a certain area, including information such as soil mineral content, soil fertility, and water retention rate of the soil.

[0023] For example, the soil ecological restoration data can be obtained from detection devices.

[0024] The soil restoration status detection network can include Artificial Intelligence (AI). AI uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, and is a theory, method, technology, and application system that perceives the environment, acquires knowledge, and uses knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technology in computer science. It attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a way similar to human intelligence. Artificial intelligence also studies the design principles and implementation methods of various intelligent machines to enable machines to have the functions of perception, reasoning, and decision-making.

[0025] Artificial intelligence technology is an interdisciplinary subject with a wide range of fields involved, including both hardware-level technologies and software-level technologies. The basic technologies of artificial intelligence generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, sequential storage, big data processing technology, operation / interaction systems, and mechatronics. The software technologies of artificial intelligence mainly include several major directions such as computer vision technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0026] Among them, the first soil restoration status detection network is used to calculate soil material information, and the second soil restoration status detection network is used to calculate the information of microorganisms in the soil.

[0027] The embodiment of this application applies multi-label classification technology, that is, the training soil ecological restoration data can have one or two labels, and the first soil restoration status detection network can be a soil restoration status detection network with a deeper network layer and a narrower network channel.

[0028] Further, the training soil ecological restoration data, the soil object labels it has, and the inclusion labels are respectively loaded into the first soil restoration status detection network and the second soil restoration status detection network for training. Since inclusion labels are introduced on the premise of soil object labels, the learned first soil restoration status detection network and second soil restoration status detection network can not only learn the ability to identify the soil object types of the training soil ecological restoration data, but also strengthen the feature correlation between similar soil ecological restoration data in the same directory.

[0029] Regarding a possible embodiment, the step of respectively loading the training soil ecological restoration data, the soil object labels it has, and the inclusion labels into the first soil restoration status detection network and the second soil restoration status detection network for training may include the following.

[0030] (1) Optimize the training soil ecological restoration data.

[0031] (2) Respectively load the optimized training soil ecological restoration data, the soil object labels it has, and the inclusion labels into the first soil restoration status detection network and the second soil restoration status detection network for training.

[0032] Among them, the training soil ecological restoration data can be processed for optimization in terms of random degree of correlation, similarity, interaction degree, and transformation to enhance the richness of the training set. Loading the optimized training soil ecological restoration data, the soil object labels it has, and the inclusion labels into the first soil restoration status detection network and the second soil restoration status detection network for training can enhance the robustness of the soil restoration status detection network.

[0033] In step 102, obtain the comparison result between the predictions of the first soil restoration status detection network and the second soil restoration status detection network.

[0034] Among them, since the network structures of the first soil restoration status detection network and the second soil restoration status detection network are inconsistent and the important soil restoration indicators for feature extraction are different, the embodiments of the present application can draw on the advantages of the first soil restoration status detection network and the second soil restoration status detection network, that is, obtain the comparison result formed between the predictions of the first soil restoration status detection network and the second soil restoration status detection network for the same training soil ecological restoration data. The lower this comparison result is, the closer the predictions of the first soil restoration status detection network and the second soil restoration status detection network are. The higher this comparison result is, the network parameters of the first soil restoration status detection network and the second soil restoration status detection network can be tuned subsequently based on this comparison result.

[0035] For a possible embodiment, the step of obtaining a comparison result between the predictions of the first soil remediation status detection network and the second soil remediation status detection network may include: calculating the comparison result between the predictions of the first soil remediation status detection network and the second soil remediation status detection network through a comparison result calculation strategy, and constructing a quantization evaluation thread corresponding to the comparison result.

[0036] For example, the quantization evaluation thread can be understood as a divergence loss method.

[0037] In step 103, optimize the network parameters of the first soil remediation status detection network and the second soil remediation status detection network according to the comparison result to obtain the trained first soil remediation status detection network and the second soil remediation status detection network.

[0038] Among them, in order to enable the first soil remediation status detection network and the second soil remediation status detection network to learn each other's important indicators for soil remediation in feature extraction, the network parameters of the first soil remediation status detection network and the second soil remediation status detection network can be optimized according to the reverse conduction of the comparison result, so as to load the important indicators of the first soil remediation status detection network for soil remediation in feature extraction to the second soil remediation status detection network. Similarly, load the important indicators of the second soil remediation status detection network for soil remediation in feature extraction to the first soil remediation status detection network.

[0039] Furthermore, it is possible to continue to obtain the comparison result formed between the predictions of the first soil remediation status detection network and the second soil remediation status detection network with optimized network parameters for the same training soil ecological remediation data. Due to the previous network parameter optimization, the first soil remediation status detection network and the second soil remediation status detection network learn from each other in feature extraction, and this comparison result will become smaller. Therefore, this comparison result will become smaller and smaller until it converges, that is, the mutual learning of the first soil remediation status detection network and the second soil remediation status detection network is realized, and the trained first soil remediation status detection network and the second soil remediation status detection network are obtained. The trained first soil remediation status detection network not only processes the details of soil ecological remediation data more accurately, but also has better feature extraction ability because it has learned the important indicators of soil remediation of the second soil remediation status detection network. And because of multi-label training, on the premise that the first soil remediation status detection network can identify the soil objects in the soil ecological remediation data, it can better determine the feature correlation between similar soil ecological remediation data in the same directory.

[0040] For a possible embodiment, the step of optimizing the network parameters of the first soil remediation status detection network and the second soil remediation status detection network according to the comparison result to obtain the trained first soil remediation status detection network and the second soil remediation status detection network may include the following content.

[0041] (1) Combine the quantization evaluation thread with the initial loss value thread of the first soil remediation status detection network for splicing training to obtain the trained first soil remediation status detection network.

[0042] (2) Combine the quantization evaluation thread with the initial loss value thread of the second soil remediation status detection network for splicing training to obtain the trained second soil remediation status detection network.

[0043] Among them, combining the quantization evaluation thread with the initial loss value thread of the first soil remediation status detection network for splicing training enables the initial loss value thread of the first soil remediation status detection network to be combined with the quantization evaluation thread for splicing training. That is, the first soil remediation status detection network can continuously learn the important soil remediation indicators of the second soil remediation status detection network under normal training conditions until the initial loss value thread and the quantization evaluation thread converge simultaneously, obtaining the trained first soil remediation status detection network.

[0044] Furthermore, combining the quantization evaluation thread with the initial loss value thread of the second soil remediation status detection network for splicing training enables the initial loss value thread of the second soil remediation status detection network to be combined with the quantization evaluation thread for splicing training. That is, the second soil remediation status detection network can continuously learn the important soil remediation indicators of the first soil remediation status detection network under normal training conditions until the initial loss value thread and the quantization evaluation thread converge simultaneously, obtaining the trained second soil remediation status detection network.

[0045] In step 104, receive the soil ecological remediation data to be processed, extract the feature information of the soil ecological remediation data to be processed through the trained first soil remediation status detection network for soil ecological remediation data detection, and obtain the detection result.

[0046] Among them, after the first soil remediation status detection network and the second soil remediation status detection network are trained, the predicted soil ecological remediation data uploaded by the user can be received. The predicted soil ecological remediation data may include the data to be processed, and the data to be processed can be read to generate the soil ecological remediation data to be processed.

[0047] To solve the above technical problems, in the embodiments of the present application, the first soil remediation status detection network after multi-label training can extract the feature information of the soil ecological remediation data to be processed. Different from the prior art, due to the introduction of training including labels, it is possible to better distinguish the target soil ecological remediation data in the same directory, avoiding the situation of abnormal detection of soil ecological remediation data caused by incorrect label selection.

[0048] For a possible embodiment, the steps of receiving the soil ecological remediation data to be processed, extracting the feature information of the soil ecological remediation data to be processed through the trained first soil remediation status detection network for soil ecological remediation data detection, and obtaining the detection result may include the following content.

[0049] (1) Receive the predicted soil ecological remediation data, and identify the data to be processed in the predicted soil ecological remediation data through the target parsing network.

[0050] (2) Read the data to be processed to generate the soil ecological remediation data to be processed.

[0051] (3) Load the soil ecological remediation data to be processed into the trained first soil remediation status detection network, extract the feature information after global convolution for soil ecological remediation data detection, and obtain the detection result.

[0052] Among them, when receiving the predicted soil ecological remediation data, the predicted soil ecological remediation data can be the soil ecological remediation data containing the data to be processed uploaded by the user. The data to be processed in the predicted soil ecological remediation data is identified through the target parsing network, and operations are performed on the data to be processed. A part of the data to be processed is read to generate the soil ecological remediation data to be processed.

[0053] Further, load the soil ecological remediation data to be processed into the trained first soil remediation status detection network, extract the feature information after global convolution, and from the front multiple soil ecological remediation data that are the most similar under each directory in the background library according to the feature information. Calculate the feature information of the front multiple soil ecological remediation data in each directory through the trained first soil remediation status detection network, calculate the difference splicing value between the feature information of the front multiple soil ecological remediation data in each directory and the feature information of the soil ecological remediation data to be processed, use the label with the lowest difference splicing value as the detected label, and use the soil ecological remediation data with the smallest difference in the detected label as the detected soil ecological remediation data for display.

[0054] For a possible embodiment, the steps of extracting the feature information after global convolution for soil ecological remediation data detection and obtaining the detection result may include the following content.

[0055] Extract the feature information of the preset layer after global convolution.

[0056] Determine multiple directories to which the feature information belongs.

[0057] Obtain the preset number of target soil ecological restoration data under each directory sorted in ascending order of the acquaintance degree with the feature information.

[0058] Load the preset number of target soil ecological restoration data under each directory into the trained first soil restoration state detection network respectively, and extract the feature information of the preset layer after global convolution of the preset number of target soil ecological restoration data under each directory.

[0059] Calculate the splicing value of the discrimination data between the feature information of each target soil ecological restoration data and the feature information of the soil ecological restoration data to be processed in the same directory.

[0060] Take the directory with the smallest splicing value as the target directory, and obtain the target soil ecological restoration data with the smallest discrimination data under this target directory as the detection result of the soil ecological restoration data.

[0061] Among them, extract the feature information of the preset layer after global convolution through the trained first soil restoration state detection network, and determine multiple directories related to the feature information, that is, determine multiple directories to which the feature information most likely belongs.

[0062] Furthermore, in each directory, detect the preset number of target soil ecological restoration data sorted in ascending order of acquaintance degree by the feature information, load the preset number of target soil ecological restoration data under each directory into the trained first soil restoration state detection network respectively, and extract the feature information of the preset layer after global convolution of the preset number of target soil ecological restoration data under each directory. Therefore, calculate the splicing value of the discrimination data between the feature information of each target soil ecological restoration data and the feature information of the soil ecological restoration data to be processed in the same directory. This splicing value can reflect the relevance between the soil ecological restoration data to be processed and each directory. Take the directory with the smallest splicing value as the target directory with the largest correlation degree, and obtain the target soil ecological restoration data with the smallest discrimination data under this target directory as the detection result of the soil ecological restoration data. Therefore, when detecting the soil ecological restoration data, sorting and comparing the proximity degrees of multiple directories can more accurately detect the target directory, and then find the best detected soil ecological restoration data.

[0063] As can be seen from the above, in the embodiment of the present application, the training soil ecological restoration data, the soil object labels, and the inclusion labels are respectively loaded into the first soil restoration status detection network and the second soil restoration status detection network for training; the comparison result between the predictions of the first soil restoration status detection network and the second soil restoration status detection network is obtained; the network parameters of the first soil restoration status detection network and the second soil restoration status detection network are optimized according to the comparison result to obtain the trained first soil restoration status detection network and the second soil restoration status detection network; the soil ecological restoration data to be processed is received, and the feature information of the soil ecological restoration data to be processed is extracted through the trained first soil restoration status detection network for soil ecological restoration data detection to obtain the detection result. Therefore, through the inclusion labels between the soil ecological restoration data with similar soil object labels and soil ecological restoration data features, the two networks are trained with multiple labels for the training soil ecological restoration data, and the network parameters of the two networks are optimized according to the comparison result, improving the working efficiency of the trained networks, thereby improving the accuracy of information detection and ensuring the reliability of ecological restoration.

[0064] The soil ecological restoration detection method based on artificial intelligence provided by the embodiment of the present application. The method process may include the following contents.

[0065] In step 201, the training soil ecological restoration data is obtained, and the preset soil ecological restoration data corresponding to each soil object label type is determined.

[0066] Among them, a large amount of training soil ecological restoration data can be obtained, and the training soil ecological restoration data has soil object labels.

[0067] In step 202, the soil ecological restoration data features of the preset soil ecological restoration data corresponding to each soil object label type are extracted, and the soil ecological restoration data features of the preset soil ecological restoration data corresponding to each soil object label type are classified by classification, and the preset soil ecological restoration data with a soil ecological restoration data feature similarity greater than the preset specified value is classified to generate the inclusion labels of the classified preset soil ecological restoration data.

[0068] In step 203, the training soil ecological restoration data is optimized, and the optimized training soil ecological restoration data, the soil object labels, and the inclusion labels are respectively loaded into the first soil restoration status detection network and the second soil restoration status detection network for training.

[0069] In step 204, the comparison result between the predictions of the first soil restoration status detection network and the second soil restoration status detection network is calculated through the comparison result calculation strategy, and a quantization evaluation thread corresponding to the comparison result is constructed.

[0070] In step 205, the quantization evaluation thread is spliced and trained with the initial loss value thread of the first soil remediation status detection network to obtain the trained first soil remediation status detection network. The quantization evaluation thread is spliced and trained with the initial loss value thread of the second soil remediation status detection network to obtain the trained second soil remediation status detection network.

[0071] In step 206, the predicted soil ecological remediation data is received, the data to be processed in the predicted soil ecological remediation data is identified through the target parsing network, the data to be processed is read, and the data to be processed for soil ecological remediation is generated.

[0072] In step 207, the data to be processed for soil ecological remediation is loaded into the trained first soil remediation status detection network, the feature information of the preset layer after global convolution is extracted, and multiple directories to which the feature information belongs are determined.

[0073] In step 208, the preset number of target soil ecological remediation data under each directory sorted in ascending order of similarity to the feature information is obtained. The preset number of target soil ecological remediation data under each directory is respectively loaded into the trained first soil remediation status detection network, and the feature information of the preset layer after global convolution of the preset number of target soil ecological remediation data under each directory is extracted.

[0074] In step 209, the discrimination data between the feature information of each target soil ecological remediation data and the feature information of the data to be processed for soil ecological remediation in the same directory is calculated. The target soil ecological remediation data in the same directory is sorted in descending order of the discrimination data, and the discrimination data of the target soil ecological remediation data in the same directory is weighted according to the sorting situation, and the splicing value of the discrimination data of the target soil ecological remediation data in the same directory after weighted processing is calculated.

[0075] Among them, the discrimination data between the feature information of each target soil ecological remediation data and the feature information of the data to be processed for soil ecological remediation in the same directory is calculated. The discrimination data can be the Euclidean difference or the cosine difference. The smaller the discrimination data, the closer the two are. The larger the discrimination data, the greater the difference between the two. The target soil ecological remediation data in the same directory can be sorted in descending order of the discrimination data.

[0076] Further, the discrimination data of each target soil ecological restoration data in each directory can be weighted according to the sorting situation. Since the trained first soil restoration status detection network has the ability to distinguish similar soil ecological restoration data in the same directory, the difference value between the to-be-processed soil ecological restoration data and the similar soil ecological restoration data under the most detected label will be significantly smaller than the difference value between the to-be-processed soil ecological restoration data and the soil ecological restoration data under other labels. To increase the discrimination degree of the difference values of different labels, the discrimination data of each target soil ecological restoration data in the same directory can be weighted in sequence according to the ranking order.

[0077] In step 210, calculate the splicing value of the discrimination data of each target soil ecological restoration data in the same directory after weighted processing, take the directory with the smallest splicing value as the target directory, and obtain the target soil ecological restoration data with the smallest discrimination data in the target directory as the detection result of the soil ecological restoration data.

[0078] Among them, calculate the splicing value of the discrimination data of each target soil ecological restoration data in the same directory after weighted processing. This splicing value reflects the detection degree of the to-be-processed soil ecological restoration data and each directory. Take the directory with the smallest splicing value as the most detected target directory, and the shoe label with the smallest splicing value can be used as the target directory. Obtain the target soil ecological restoration data with the smallest discrimination data in this target directory as the detection result of the soil ecological restoration data. Therefore, through the above multi-label training method and label detection method, the target directory to which the to-be-processed soil ecological restoration data belongs can be accurately determined, and then a soil ecological restoration data detection result with high accuracy can be generated.

[0079] As can be seen from the above, in the embodiment of the present application, the training soil ecological restoration data, the soil object labels, and the inclusion labels are respectively loaded into the first soil restoration status detection network and the second soil restoration status detection network for training; obtain the comparison result between the predictions of the first soil restoration status detection network and the second soil restoration status detection network; optimize the network parameters of the first soil restoration status detection network and the second soil restoration status detection network according to the comparison result to obtain the trained first soil restoration status detection network and the second soil restoration status detection network; receive the to-be-processed soil ecological restoration data, and extract the feature information of the to-be-processed soil ecological restoration data through the trained first soil restoration status detection network for soil ecological restoration data detection to obtain the detection result. Therefore, through the inclusion labels between the soil ecological restoration data with similar soil object labels and soil ecological restoration data features, the two networks are trained with the training soil ecological restoration data for multi-label training, and the network parameters of the two networks are optimized according to the comparison result, improving the working efficiency of the trained networks, thereby improving the accuracy of information detection and ensuring the reliability of ecological restoration.

[0080] On this basis, a soil ecological restoration detection device based on artificial intelligence is provided. The device includes: A network training module for loading training soil ecological restoration data, the soil object labels it has, and the inclusion labels into the first soil restoration status detection network and the second soil restoration status detection network respectively for training. The inclusion labels are used to calibrate the training soil ecological restoration data with the similarity of soil ecological restoration data features greater than a preset specified value; A result comparison module for obtaining the comparison result between the predictions of the first soil restoration status detection network and the second soil restoration status detection network; A parameter optimization module for optimizing the network parameters of the first soil restoration status detection network and the second soil restoration status detection network in combination with the comparison result to obtain the trained first soil restoration status detection network and the second soil restoration status detection network. Among them, the second soil restoration status detection network is used to load the important soil restoration indicators for feature extraction of the second soil restoration status detection network into the first soil restoration status detection network to realize the mutual feature extraction learning between the first soil restoration status detection network and the second soil restoration status detection network, and obtain the trained first soil restoration status detection network and the second soil restoration status detection network, so that the trained first soil restoration status detection network has the important soil restoration indicators for feature extraction of the second soil restoration status detection network; An information detection module for receiving the soil ecological restoration data to be processed, and extracting the feature information of the soil ecological restoration data to be processed through the trained first soil restoration status detection network for soil ecological restoration data detection to obtain the detection result.

[0081] On this basis, a soil ecological restoration detection system based on artificial intelligence is shown, including a processor and a memory that communicate with each other. The processor is used to read and execute a computer program from the memory to implement the above method.

[0082] On this basis, a computer-readable storage medium is also provided, and the computer program stored thereon implements the above method when running.

[0083] In summary, based on the above solution, the training soil ecological restoration data, the soil object tags, and the inclusion tags are respectively loaded into the first soil restoration status detection network and the second soil restoration status detection network for training; the comparison result between the predictions of the first soil restoration status detection network and the second soil restoration status detection network is obtained; the network parameters of the first soil restoration status detection network and the second soil restoration status detection network are optimized according to the comparison result to obtain the trained first soil restoration status detection network and the second soil restoration status detection network; the soil ecological restoration data to be processed is received, and the feature information of the soil ecological restoration data to be processed is extracted through the trained first soil restoration status detection network for soil ecological restoration data detection to obtain the detection result. Therefore, through the inclusion tags between the soil ecological restoration data with similar soil object tags and soil ecological restoration data features, the two networks are trained with multiple labels for the training soil ecological restoration data, and the network parameters of the two networks are optimized according to the comparison result, improving the working efficiency of the trained networks, thereby improving the accuracy of information detection and ensuring the reliability of ecological restoration.

[0084] It should be understood that the above-described systems and their modules can be implemented in various ways. For example, in some embodiments, the systems and their modules can be implemented by hardware, software, or a combination of software and hardware. Among them, the hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by a suitable instruction execution system, such as a microprocessor or dedicated design hardware. Those skilled in the art can understand that the above methods and systems can be implemented using computer-executable instructions and / or included in processor control code, such as provided on a carrier medium such as a disk, CD, or DVD-ROM, a programmable memory such as read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The systems and their modules of the present application can be implemented not only by hardware circuits of programmable hardware devices such as very large scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable logic devices such as field programmable gate arrays, but also by software executed by various types of processors, or by a combination of the above hardware circuits and software (e.g., firmware).

[0085] It should be noted that the beneficial effects that may be produced by different embodiments are different. In different embodiments, the beneficial effects that may be produced can be any one or several of the above combinations, or any other beneficial effects that may be obtained.

Claims

1. An artificial intelligence-based soil ecological restoration detection method, characterized in that, The method includes: Loading the training soil ecological restoration data, the soil object labels it has, and the inclusion labels into the first soil restoration status detection network and the second soil restoration status detection network for training respectively. The inclusion labels are used to calibrate the training soil ecological restoration data with the similarity of soil ecological restoration data features greater than a preset specified value; Obtaining the comparison result between the predictions of the first soil restoration status detection network and the second soil restoration status detection network; Optimizing the network parameters of the first soil restoration status detection network and the second soil restoration status detection network in combination with the comparison result to obtain the trained first soil restoration status detection network and the second soil restoration status detection network. Among them, the second soil restoration status detection network is used to load the important soil restoration indicators for feature extraction of the second soil restoration status detection network into the first soil restoration status detection network, so as to realize the mutual feature extraction learning between the first soil restoration status detection network and the second soil restoration status detection network, and obtain the trained first soil restoration status detection network and the second soil restoration status detection network, so that the trained first soil restoration status detection network has the important soil restoration indicators for feature extraction of the second soil restoration status detection network; Receiving the soil ecological restoration data to be processed, and extracting the feature information of the soil ecological restoration data to be processed through the trained first soil restoration status detection network for soil ecological restoration data detection to obtain the detection result.

2. The method for detecting soil ecological restoration based on artificial intelligence according to claim 1, wherein The artificial intelligence-based soil ecological restoration detection method further includes: Obtaining the training soil ecological restoration data, and the training soil ecological restoration data has soil object labels; determining the preset soil ecological restoration data corresponding to each soil object label type; Classifying the preset soil ecological restoration data corresponding to different soil object label types, and classifying the preset soil ecological restoration data with the similarity of soil ecological restoration data features greater than a preset specified value; Generating the inclusion labels of the classified preset soil ecological restoration data.

3. The method for detecting soil ecological restoration based on artificial intelligence according to claim 2, wherein, The step of classifying the preset soil ecological restoration data corresponding to different soil object label types and classifying the preset soil ecological restoration data with the similarity of soil ecological restoration data features greater than a preset specified value includes: Extracting the soil ecological restoration data features of the preset soil ecological restoration data corresponding to each soil object label type; Classifying the soil ecological restoration data features of the preset soil ecological restoration data corresponding to each soil object label type through a classification method; Classifying the preset soil ecological restoration data with the similarity of soil ecological restoration data features greater than a preset specified value.

4. The method for detecting soil ecological restoration based on artificial intelligence according to claim 1, wherein, The step of loading the training soil ecological restoration data, the soil object labels it has, and the inclusion labels into the first soil restoration status detection network and the second soil restoration status detection network for training respectively includes: Performing optimization processing on the training soil ecological restoration data; Load the optimized training soil ecological restoration data, the soil object labels it has, and the inclusion labels into the first soil restoration status detection network and the second soil restoration status detection network for training respectively.

5. The method for detecting soil ecological restoration based on artificial intelligence according to claim 4, wherein The step of obtaining the comparison result between the predictions of the first soil restoration status detection network and the second soil restoration status detection network includes: calculating the comparison result between the predictions of the first soil restoration status detection network and the second soil restoration status detection network through a comparison result calculation strategy, and constructing a quantization evaluation thread corresponding to the comparison result.

6. The method for detecting soil ecological restoration based on artificial intelligence according to claim 5, wherein, The step of optimizing the network parameters of the first soil restoration status detection network and the second soil restoration status detection network by combining the comparison result to obtain the trained first soil restoration status detection network and the second soil restoration status detection network includes: Splice and train the quantization evaluation thread with the initial loss value thread of the first soil restoration status detection network to obtain the trained first soil restoration status detection network; Splice and train the quantization evaluation thread with the initial loss value thread of the second soil restoration status detection network to obtain the trained second soil restoration status detection network.

7. The method for detecting soil ecological restoration based on artificial intelligence according to claim 4, wherein, The step of receiving the soil ecological restoration data to be processed, extracting the feature information of the soil ecological restoration data to be processed through the trained first soil restoration status detection network for soil ecological restoration data detection, and obtaining the detection result includes: Receive the predicted soil ecological restoration data, and identify the data to be processed in the predicted soil ecological restoration data through the target parsing network; Read the data to be processed and generate the soil ecological restoration data to be processed; Load the soil ecological restoration data to be processed into the trained first soil restoration status detection network, extract the feature information after global convolution for soil ecological restoration data detection, and obtain the detection result.

8. The method for detecting soil ecological restoration based on artificial intelligence according to claim 7, characterized in that The step of extracting the feature information after global convolution for soil ecological restoration data detection and obtaining the detection result includes: Extract the feature information of a preset layer after global convolution; Determine multiple directories to which the feature information belongs; obtain a preset number of target soil ecological restoration data in each directory sorted from smallest to largest in terms of similarity to the feature information; load the preset number of target soil ecological restoration data in each directory into the trained first soil restoration status detection network respectively, and extract the feature information of the preset layer after global convolution of the preset number of target soil ecological restoration data in each directory; Calculate the splicing value of the discrimination data between the feature information of each target soil ecological restoration data in the same directory and the feature information of the soil ecological restoration data to be processed; Take the directory with the smallest splicing value as the target directory, and obtain the target soil ecological restoration data with the smallest discrimination data in the target directory as the soil ecological restoration data detection result.

9. The method for detecting soil ecological restoration based on artificial intelligence according to claim 8, characterized in that The step of calculating the splicing value of the discrimination data between the feature information of each target soil ecological restoration data in the same directory and the feature information of the soil ecological restoration data to be processed includes: Calculate the discrimination data between the characteristic information of each target soil ecological restoration data in the same directory and the characteristic information of the to-be-processed soil ecological restoration data; Sort each target soil ecological restoration data in the same directory according to the discrimination data from large to small; Perform weighted processing on the discrimination data of each target soil ecological restoration data in the same directory according to the sorting situation; Calculate the splicing value of the discrimination data of each target soil ecological restoration data in the same directory after weighted processing.

10. An artificial intelligence-based soil ecological restoration detection system, characterized in that, It includes a processor and a memory that communicate with each other. The processor is used to read and execute a computer program from the memory to implement the method according to any one of claims 1-9.

Images