Value-oriented land composition detection method and land composition detection system

Abstract

The invention discloses a value-oriented land component detection method and a land component detection system, and relates to the technical field of component detection, and the method comprises the steps: connecting a spectrometer, deploying a spectrum configuration region in a control center in an embedded manner, obtaining a land utilization scene, carrying out component conversion according to a scene value index, and setting a multispectral band according to component correlation; response driving and land detection of the spectrograph are carried out, and distributed spectral information is determined; and by taking distributed mean value fusion of the spectral features as a first preprocessing condition and distributed difference measurement of the spectral features as a second preprocessing condition, executing decision of spectral feature-land components in parallel, and determining a first global result and a second equilibrium result as land component detection results. The technical problem that in the prior art, due to the fact that detection parameters cannot be set adaptively according to target requirements, land composition detection lacks value guidance is solved, and the technical effect of improving the pertinence of land composition detection is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of component detection, and particularly to a value-oriented land component detection method and a land component detection system. Background Art

[0002] With the acceleration of the urbanization process and the continuous improvement of the land resource management requirements, land component detection has received extensive attention in many fields such as agricultural production and ecological restoration. Especially in application scenarios such as refined land use assessment and cultivated land protection, accurate identification based on soil components is of great significance for improving the efficiency of land resource allocation. However, with the continuous expansion of the detection area and the increasing diversification of land use types, traditional land component detection means have gradually revealed problems such as poor adaptability, rough assessment, and lack of application orientation. Existing land component detection methods mostly adopt a unified spectral detection standard, lacking the ability to conduct differential analysis according to different land use values, and unable to effectively distinguish the different requirements of high-value uses (such as high-yield farmland and planned construction land) and general areas in terms of detection accuracy and component priority. Summary of the Invention

[0003] This application provides a value-oriented land component detection method and a land component detection system, which are used to solve the technical problem that the prior art cannot adaptively set detection parameters according to target requirements, resulting in the lack of value orientation in land component detection.

[0004] In view of the above problems, this application provides a value-oriented land component detection method and a land component detection system.

[0005] In the first aspect of this application, a value-oriented land component detection method is provided. The method includes:

[0006] Connect a spectrometer, embed a spectral configuration area in the control center, convert components with scene value indicators by obtaining the land use scenario, set multi-spectral bands according to component correlation, including essential components and inferior components; set the distributed sampling rules for the target land area, fuse them with the multi-spectral bands, perform response drive of the spectrometer and land detection, and determine the distributed spectral information to be transmitted back; take the distributed mean fusion of spectral features as the first preprocessing condition and the distributed difference measure of spectral features as the second preprocessing condition, and perform the decision of spectral feature - land component in parallel to determine the first global result and the second balance result as the land component detection result.

[0007] In the second aspect of this application, a value-oriented land component detection system is provided. The system includes:

[0008] The component conversion module is used to connect to the spectrometer, and is embedded in the control center to deploy the spectral configuration area. By obtaining the land use scenario, it performs component conversion based on the scenario value index, and sets multi-spectral bands according to component correlation, including necessary components and inferior components. The spectral information determination module is used to set the distributed sampling rules for the target land area, fuse with the multi-spectral bands, perform response drive of the spectrometer and land detection, and determine the distributed spectral information. The land component detection result generation module is used to take the distributed mean fusion of spectral features as the first preprocessing condition and the distributed difference measure of spectral features as the second preprocessing condition, and parallelly execute the decision of spectral feature - land component to determine the first global result and the second balance result as the land component detection result.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] This application connects to the spectrometer, embeds and deploys the spectral configuration area in the control center. By obtaining the land use scenario, it performs component conversion based on the scenario value index, and sets multi-spectral bands according to component correlation, including necessary components and inferior components; sets the distributed sampling rules for the target land area, fuses with the multi-spectral bands, performs response drive of the spectrometer and land detection, and determines the distributed spectral information; takes the distributed mean fusion of spectral features as the first preprocessing condition and the distributed difference measure of spectral features as the second preprocessing condition, and parallelly executes the decision of spectral feature - land component to determine the first global result and the second balance result as the land component detection result. This invention solves the technical problem that the prior art cannot adaptively set detection parameters according to target requirements, resulting in the lack of value orientation in land component detection. By constructing a component conversion mechanism based on scenario value index, a distributed response sampling strategy for fusing multi-spectral bands, and a parallel spectral decision channel, it achieves the technical effect of improving the pertinence of land component detection. Description of the Drawings

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0012] Figure 1 It is a schematic flowchart of the value-oriented land component detection method provided by the embodiment of this application;

[0013] Figure 2 It is a schematic structural diagram of the value-oriented land component detection system provided by the embodiment of this application.

[0014] Explanation of the reference numerals: component conversion module 11 , spectral information determination module 12 , land component detection result generation module 13 . DETAILED DESCRIPTION

[0015] This application provides a value-oriented land composition detection method and a land composition detection system to solve the technical problem that the existing technology is unable to adaptively set detection parameters according to target requirements, resulting in a lack of value orientation in land composition detection. By constructing a component conversion mechanism based on scene value indicators, a distributed response sampling strategy integrating multi-spectral bands, and a parallel spectral decision channel, the technical effect of improving the targeted nature of land composition detection is achieved.

[0016] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0017] It should be noted that any variations of the terms "include" and "have" are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or inherent to these processes, methods, products or devices.

[0018] Embodiment 1, as Figure 1 As shown, the embodiment of the present application provides a value-oriented land composition detection method, the method comprising:

[0019] Step S100: Connect the spectrometer and deploy the spectral configuration area in the control center. Obtain the land use scene, convert the components according to the scene value index, and set the multi-spectral band according to the component correlation, which includes the necessary components and the inferior components.

[0020] In the embodiment of the present application, first, a spectrometer is connected, and an embedded deployment of a spectral configuration area is completed in the front area of the control response end of its control center to establish a stable data communication and response mechanism. On this basis, land use scenario information is obtained, and combined with the value indicators reflected in the scenario, the land components are classified and identified and prioritized. By retrieving and mining the spectral relationship library containing the maximum correlation response band between the surface reflectance and soil components, a spectral configuration area matching the scenario value is constructed. With the support of the spectral configuration area, a first type of spectral band is determined according to the necessary components defined in the scenario value indicator and is given a first identifier; then a second type of spectral band is determined according to the inferior components and is given a second identifier. Finally, through the aggregation of the above-mentioned first and second types of spectral bands, a multi-spectral band configuration is formed for a specific land use scenario, reflecting the component correlation.

[0021] Further, in the method provided by the application embodiment, for the embedded deployment of the spectral configuration area in the control center, it further includes:

[0022] Connect to the land database, retrieve and mine the spectral relationship library of land component - spectral response, wherein the maximum correlation response band under the surface reflectance and soil components is the mining target; according to the spectral relationship library, build a spectral configuration area; in the front area of the control response end of the spectrometer control center, perform an embedded deployment on the spectral configuration area and establish a connection with the control response end.

[0023] In the embodiment of the present application, first, connect to the land database, which includes soil type information, geographical location data, component measurement values, and surface reflectance data of various soil components in different bands. The surface reflectance represents the degree of reflection of radiation energy by the soil in different electromagnetic bands and is a key parameter characterizing the spectral characteristics of the soil. By processing the retrieved data and using the Pearson correlation coefficient analysis method, with the maximum correlation response band between the surface reflectance and soil components as the mining target, the linear correlation between various components in the soil and the reflectance of each spectral band is calculated one by one, and the spectral band with the highest correlation is selected to establish the corresponding relationship between the land components and the spectral bands, and build the spectral relationship library of land component - spectral response.

[0024] Next, a spectral configuration area is built according to the spectral relationship library. This spectral configuration area is a structured data area used to define the band combination scheme for the spectrometer response. During the building process, the parameter mapping configuration method is adopted to map and organize each item of land component in the spectral relationship library with its maximum correlation response band, and classify and mark them according to the component category to clarify the priority and call order of the necessary components and the inferior components in the configuration area. Each set of configuration parameters includes the band center value, bandwidth range, function identifier, call mark, etc. After all configuration items are classified and summarized according to the component attributes, the construction of the spectral configuration area is completed. This spectral configuration area supports component-oriented band rapid switching and dynamic call, providing precise adaptation ability for spectral detection requirements in different land use scenarios.

[0025] Finally, in the front area of the control response end of the control center of the spectrometer, the spectral configuration area is embedded and deployed and connected to the control response end. The control center is the core execution area of the spectrometer, and the front area of the control response end is the specific interface position for implementing parameter call and spectral scheduling logic embedding. Through the embedded parameter loading method, the completed spectral configuration area is loaded into the parameter reading channel of the control response end, so that each band combination parameter in the configuration area can directly participate in the real-time response instruction of the spectrometer. After the connection is completed, the spectrometer can call the corresponding configuration in different control scenarios to implement the component-driven spectral response control mechanism, ensuring the accuracy and differential target recognition ability of the subsequent spectral acquisition process.

[0026] Furthermore, in the method provided by the application embodiment, setting the multi-spectral band according to the component correlation further includes:

[0027] Based on the necessary components of the scenario value index, making a decision according to the spectral configuration area to determine a class of spectral bands, where the class of spectral bands has a first identifier; based on the inferior components of the scenario value index, making a decision according to the spectral configuration area to determine a second class of spectral bands, where the second class of spectral bands has a second identifier; aggregating the first class of spectral bands and the second class of spectral bands as the set multi-spectral band.

[0028] In the embodiment of the present application, first, the necessary components that should be present in the current land use scenario are identified, such as nitrogen, phosphorus, and organic matter in agricultural land, or water content and mineral elements in ecological land. This step is based on the spectral configuration area established in the early stage, and the relevant matching method is used to extract the set of spectral bands corresponding to the necessary components defined by the scenario value index in the configuration area. The corresponding surface reflectance characteristic curves are compared one by one, and according to the principle of maximum correlation response, the dominant band that can best reflect the characteristics of the necessary component is selected as a type of spectral band. To facilitate the distinction of the sources of these bands in detection and interpretation, a first identifier is uniformly assigned to all the first-type spectral bands, and this identifier indicates that these bands correspond to the "should-have" necessary components.

[0029] Secondly, the adverse components that need to be monitored or avoided in the current scenario are identified, such as phenomena like excessive salt content, heavy metal exceeding the standard, or acidification trend. This step also uses the spectral configuration area data and the characteristic indication method to determine the bands. Specifically, a list of adverse components is listed according to the scenario value index, the spectral bands corresponding to them in the configuration area are extracted, and by analyzing the change trend of the surface reflectance characteristics, the most indicative representative bands are selected and classified as the second-type spectral bands. To distinguish them from the first-type bands and indicate that they are the detection channels for adverse components in subsequent analysis, a second identifier is assigned to all the second-type spectral bands, and this identifier indicates that these bands are used to detect the "should-avoid" adverse components.

[0030] Finally, the first-type spectral bands and the second-type spectral bands are summarized and integrated to form a unified multi-spectral band configuration. This step uses the numbering collection and identifier retention method to merge the first-type spectral bands with the first identifier and the second-type spectral bands with the second identifier into a band list. During the merging process, the band entries with repeated values or overlapping bandwidths are removed, and the corresponding identifier information is retained.

[0031] Step S200: Set the distributed sampling rule for the target land area, fuse it with the multi-spectral bands, perform the response drive of the spectrometer and land detection, and determine the distributed spectral information.

[0032] In the embodiment of the present application, first, the distributed sampling rule for the target land area is set, and this rule is formulated based on the surface utilization value. The surface utilization value is an important indicator to measure the current or potential use of the land, such as agricultural production capacity, ecological protection value, or development suitability. By spatially dividing the target area, representative locations are preferentially selected, including high-value core areas, boundary transition areas, and typical geomorphic units, to form a sampling network with strong coverage and balanced structure.

[0033] After the sampling rules are set, they are fused with the set multi-spectral bands. The multi-spectral bands are a set of spectral response bands associated with the aforementioned necessary components and inferior components. According to the position characteristics of the sampling points and the component detection objectives, a spatial matching and band adaptation method is used to assign specific detection band combinations to each sampling point, and a complete spectral sampling strategy is constructed. Subsequently, according to the above sampling strategy, the response-driven and land detection processes are executed. Response-driven means automatically sending control instructions to the spectrometer based on the sampling strategy, including operations such as starting specific bands, adjusting the sampling angle and frequency, etc., to ensure that the spectrometer operates according to the set parameters at each sampling position. Land detection means that after the spectrometer completes the response, it conducts on-site radiation measurement of the land surface and collects spectral data such as surface reflectance or radiance at the corresponding bands. The sampling position and timestamp information are recorded synchronously to ensure the spatio-temporal consistency of the data.

[0034] Finally, through the distributed response sampling process, distributed spectral information is obtained and integrated. This information consists of N sets of spectral data, each corresponding to a distributed sampling position, and the data content corresponds one-to-one with the multi-spectral bands.

[0035] Furthermore, in the method provided by the application embodiment, when setting the distributed sampling rules for the target land area, fusing with the multi-spectral bands, and performing response-driven of the spectrometer, it further includes:

[0036] According to the surface utilization value, conduct distributed sampling planning to determine the distributed sampling rules; use the distributed sampling rules for detection position constraints, use the multi-spectral bands for detection condition constraints, and fuse to determine the spectral sampling strategy; based on the spectral sampling strategy, control and drive the spectrometer.

[0037] Furthermore, the method provided by the application embodiment further includes:

[0038] The distributed spectral information includes N sets of spectral information corresponding to each distributed position, and each set of spectral information corresponds one-to-one with the multi-spectral bands, where N is the value of the distributed positions.

[0039] In the embodiment of the present application, first, spatial sampling planning of the target area is carried out according to the surface utilization value, and a distributed sampling rule is output. The surface utilization value refers to the comprehensive potential evaluation result of the land under a specific application background, reflecting its value level in agricultural production, ecological protection or urban development. For example, areas with high fertility and sufficient water sources belong to high-utilization value areas in agriculture, while areas with steep slopes and wastelands belong to low-utilization value areas. By using the fixed grid overlay method, after dividing the target area into regular grids, the surface utilization value layer is overlaid, and the grid cells are classified and screened according to the value level. High-value and boundary transition areas are preferentially selected as representative sampling points. Through this process, the balance of spatial distribution is ensured, and it is also ensured that the selected sampling points can cover various land use types, and finally a distributed sampling rule including sampling point coordinates, grid numbers and value labels is formed.

[0040] Next, based on the above distributed sampling rule, detection position constraints are carried out, and combined with multi-spectral bands, detection condition constraints are carried out, and a spectral sampling strategy for actual detection tasks is fused and constructed. The multi-spectral bands are band combinations used for component identification, covering response bands for necessary components (such as positive components that improve land quality, such as nitrogen, phosphorus, organic matter, etc.) and adverse components (such as harmful components such as salts and heavy metals). For example, the near-infrared band is often used to identify vegetation cover, and the short-wave infrared can reflect mineral or salt conditions. This step uses the position-band matching method, that is, according to the land type or use of the sampling point (such as cultivated land, saline-alkali land, construction area), the most suitable band combination is selected from the multi-spectral bands. For example, red and near-infrared bands are selected in cultivated land, and blue and mid-infrared bands are selected in saline-alkali areas, and they are bound to the sampling points to form a spectral sampling strategy including sampling position, band configuration, and observation requirements.

[0041] Then, the spectrometer is controlled and driven according to the spectral sampling strategy to perform specific detection tasks. The spectrometer is a sensing device that can collect the reflected information of ground objects in specific bands and can output spectral characteristics such as surface reflectance or radiance at the set bands for each sampling point. This step uses the parameterized task scheduling method to package the spatial information of the sampling points, their corresponding band combinations, sampling duration, start commands and other parameters into unit tasks, and send them to the spectrometer in sequence to control the spectrometer to start sampling at the corresponding position and according to the set bands for land detection, that is, to obtain the spectral response value after the interaction between the surface material and the incident light in the field. During the whole process, accessory information such as band numbers, sampling times, and geographical locations is recorded simultaneously to ensure the traceability and comparability of the data in the later processing.

[0042] Finally, the task scheduling and data acquisition of N sampling points are completed to form complete distributed spectral information. Here, N is the number of distributed positions set in the sampling rule; each group of spectral information corresponds one by one to the preset multi-spectral bands, and the data content includes the reflectivity or radiation value of each sampling point under all response bands, along with the spatial coordinates and acquisition timestamp of this point.

[0043] Step S300: Taking the distributed mean fusion of spectral features as the first preprocessing condition and the distributed difference measure of spectral features as the second preprocessing condition, perform the decision of spectral feature - land component in parallel to determine the first global result and the second balance degree result as the land component detection result.

[0044] In the embodiment of this application, taking the distributed mean fusion of spectral features as the first preprocessing condition and the distributed difference measure of spectral features as the second preprocessing condition in sequence, a component detection mechanism with the ability of global trend recognition and local difference perception is constructed. By performing the decision of spectral feature - land component in parallel, the first global result and the second balance degree result are finally output, jointly constituting the complete land component detection result.

[0045] In the first preprocessing path, for the spectral information of N groups of sampling points, extract the spectral response values in the same band one by one according to the multi-spectral bands, and adopt the mean fusion processing method, that is, perform average calculation on the spectral reflectivity data in the same band, so as to obtain the mean spectral information of each band. Multiple mean spectral information corresponds one by one to the set multi-spectral bands, which is used to characterize the overall spectral performance of the land and serves as the basis for global trend judgment, constituting the first preprocessing condition.

[0046] In the second preprocessing path, based on the neighborhood position relationship between sampling points, perform differential calculation on the spectral values of adjacent sampling points in the same band to obtain the spectral difference information reflecting the degree of spatial change. Subsequently, all difference results are integrated to generate a difference fluctuation map, which is used to express the fluctuation intensity and change pattern in the spectral dimension of the local area. This process is the distributed difference measure of spectral features, and its role is to capture the possible component change boundaries, abnormal trends or spatial inhomogeneities within the region, constituting the second preprocessing condition.

[0047] After completing the dual-path preprocessing, using spectral information as the input and the characteristics of the components to be identified in the land as the target, a component detection channel based on feature mapping and component recognition is constructed. This channel is replicated through a mirror twin structure to form two parallel channels with the same structure but different input features. Among them, the first detection channel receives the mean spectral information and outputs the first global result for describing the comprehensive performance of the regional components; the second detection channel receives the differential fluctuation map and outputs the second equilibrium degree result for reflecting the spatial distribution balance. The two channels together form a parallel detection mechanism, running concurrently and making decisions separately.

[0048] The finally output first global result and the second equilibrium degree result together constitute a set of land component detection results with statistical representativeness and spatial perception ability.

[0049] Furthermore, in the method provided by the application embodiment, taking the distributed mean fusion of spectral features as the first preprocessing condition, it further includes:

[0050] For the N groups of spectral information, extract the spectral information of the same spectral band for each group and calculate the mean of the spectral features, and reorganize and determine multiple mean spectral information, where the multiple mean spectral information characterizes the regional comprehensive state, and the multiple mean spectral information corresponds one-to-one with the multi-spectral bands.

[0051] In the embodiment of the present application, for N groups of spectral information, extract the spectral information of the same spectral band for each group and calculate the mean of the spectral features. Specifically, first, from N distributed sampling points, extract the spectral response values of each sampling point under all set multi-spectral bands. The N groups of spectral information refer to the original spectral reflection data obtained from N different geographical locations, covering the entire multi-spectral band range. The multi-spectral band is a set of bands selected for having high spectral response correlation with essential components or adverse components, including the visible light region (such as red light, green light), the near-infrared region, and the short-wave infrared region, etc.

[0052] Then perform a band alignment processing method on these spectral data, that is, group the response values of all sampling points under the same spectral band into one group to form multiple same-band spectral data vectors. For each band vector, further use the arithmetic mean method to calculate its mean to obtain the representative spectral reflectance value of this band in the entire region.

[0053] By repeating the above operations, process all set multi-spectral bands in turn. The finally generated multiple mean spectral information corresponds one-to-one with the multi-spectral bands in structure, that is, each mean spectral information exactly matches a specific band. These results together constitute a complete spectral mean feature sequence for summarizing the response levels of the entire region under each key band.

[0054] Further, in the method provided by the application embodiment, with the distributed difference metric of spectral features as the second preprocessing condition, it further includes:

[0055] For the N groups of spectral information, perform the difference calculation of the same spectral features between groups based on the neighborhood positions to determine N - 1 groups of spectral difference information; according to the N - 1 groups of spectral difference information, construct a difference fluctuation map based on spectral features.

[0056] In the embodiment of the present application, first, for the N groups of spectral information, perform the difference calculation of the same spectral features between groups based on the neighborhood positions. Among them, the N groups of spectral information refer to the spectral response data obtained from N distributed sampling points set in the target land area, and each group of information contains the reflectance or radiance values under a plurality of preset multispectral bands; and the neighborhood position refers to the natural arrangement order or geometric adjacency relationship of the sampling points in space, such as adjacent points arranged according to geographical coordinates or sampling grids. Pair the sampling points in sequence according to the sampling point numbers, and perform a subtraction operation on the characteristic values of each pair of adjacent sampling points in the same spectral band. For example, for the reflectance of the 1st and 2nd sampling points in a certain near-infrared band being 0.43 and 0.37 respectively, the calculated difference in this band is -0.06. Through such operations, the difference processing of N - 1 groups of adjacent point pairs is completed in each band, and finally N - 1 groups of spectral difference information are output.

[0057] Subsequently, according to the N - 1 groups of spectral difference information, construct a difference fluctuation map based on spectral features. In this process, use the band mapping rearrangement method to reorganize the N - 1 groups of spectral difference information into a two-dimensional matrix structure, with the band number as the horizontal axis and the adjacent point pair number as the vertical axis, and each cell represents the difference value of a specific band between adjacent point pairs. After all the data is filled, a difference fluctuation map with two-dimensional spatial characteristics is formed.

[0058] Further, in the method provided by the application embodiment, when parallelly performing the decision of spectral features - land components to determine the first global result and the second balance degree result, it further includes:

[0059] Taking spectral information as the input and component characteristics as the output, construct a component detection channel; perform mirror twinning on the component detection channel to determine a parallel detection channel; input the multiple mean spectral information into the first detection channel of the parallel detection channel, and input the difference fluctuation map into the second detection channel of the parallel detection channel, and perform parallel decision to determine the first global result and the second balance degree result.

[0060] In the embodiments of the present application, first, a component detection channel is constructed with spectral information as the input and component characteristics as the output. In this process, the multi-layer perceptron method is adopted, and its input data is a structured spectral vector, including reflectance features in multiple bands, such as the feature sequence after unfolding from "multiple mean spectral information" or "difference fluctuation map"; the output data is the component label of each sample point, such as the nitrogen content, organic matter ratio or heavy metal grade in the soil. The training data is composed of a measured sampling data set, containing paired "spectral feature - component label" samples, and fitting is achieved by minimizing the error between the model prediction value and the true component value. The mean square error loss function is used to evaluate the prediction accuracy during the training process. Through this process, the construction of the component detection channel is completed.

[0061] Next, a structure replication operation is performed on the trained component detection channel to construct a parallel detection channel with the same structure for separately processing different types of spectral feature inputs. By adopting the structure mirror replication method, that is, all network settings such as the number of layers, activation function, and parameter initialization of the original channel are retained, and only the input data source is changed.

[0062] Subsequently, the two types of spectral feature data are input into the parallel detection channel for component judgment respectively. Among them, the channel alignment input method is adopted, and "multiple mean spectral information" reflecting the macroscopic trend is input into the first detection channel, and the first global result representing the overall component level of the region is output by this channel; the "difference fluctuation map" reflecting the spatial difference characteristics is input into the second detection channel, and the second equilibrium degree result describing the equilibrium degree of the regional component distribution is output by this channel.

[0063] Finally, through the parallel decision-making mechanism, the first global result and the second equilibrium degree result are synchronously output within the same recognition cycle as the land component detection result.

[0064] Furthermore, the method provided by the application embodiments further includes:

[0065] Determine the baseline land components based on the scenario value indicators of the land use scenario; according to the baseline land components, perform a pass judgment on the land component detection results to generate land quality data.

[0066] In the embodiment of the present application, first, according to the target use purpose of the land, the basic evaluation criteria are determined. The method adopted in this step is the preset scenario look-up method, that is, using the land use scenario set by the user as the query condition to find the corresponding baseline land components from the built-in scenario value index database. The so-called "land use scenario" refers to the planned use of the land, such as agricultural cultivation, landscaping, building development, or ecological restoration, etc., and the "scenario value index" is the minimum standard of land components required under these scenarios. For example, if the set scenario is "green vegetable base", the corresponding baseline values are obtained as organic matter content ≥ 2.5%, pH value 6.5 - 7.5, and heavy metal lead content ≤ 50 mg / kg. This process does not require manual input of standards, and a set of corresponding baseline land component indicators is automatically determined according to the scenario as the reference system for passing judgment.

[0067] After the baseline setting is completed, the obtained component identification results are compared with the baseline data to determine whether the plot meets the quality requirements under the target use. This step adopts the interval comparison method, that is, each test result is compared item by item with the baseline range to determine whether it falls within the allowable interval. For example, in a certain test plot, the detected organic matter content is 2.7%, pH is 7.1, and lead content is 53 mg / kg. Then, comparing with the above "green vegetable base" baseline standard, the first two items are qualified, and the third item, the lead content, is on the high side. So, this plot is determined as "not up to standard", and "heavy metal exceeding the limit" is marked as a component risk reminder. In addition, combining the first global result (judging the regional average level) and the second balance result (judging the volatility of component distribution), supplementary verification is carried out in the spatial dimension. For example, even if the component mean value meets the standard, if there is serious imbalance locally, it can be marked as "conditionally qualified" or "needs treatment".

[0068] Finally, structured land quality data is output through this processing flow, which includes fields such as detected values, baseline values, pass / fail status marks, and abnormal item descriptions.

[0069] In the embodiment of the present application, in summary, the embodiment of the present application has at least the following technical effects:

[0070] This application is connected to a spectrometer. A spectral configuration area is embedded and deployed in the control center. By obtaining the land use scenario, component conversion is performed using scenario value indicators, and multi-spectral bands are set based on component correlations, including necessary components and inferior components. A distributed sampling rule for the target land area is set and fused with the multi-spectral bands to perform response driving of the spectrometer and land detection to determine distributed spectral information. Using the distributed mean fusion of spectral features as the first preprocessing condition and the distributed difference measure of spectral features as the second preprocessing condition, the decision-making of spectral feature-land components is executed in parallel to determine the first global result and the second balance result as the land component detection result. The present invention solves the technical problem that the prior art cannot adaptively set detection parameters according to target requirements, resulting in a lack of value orientation in land component detection. By constructing a component conversion mechanism based on scenario value indicators, a distributed response sampling strategy for fusing multi-spectral bands, and a parallel spectral decision-making channel, the technical effect of improving the pertinence of land component detection is achieved.

[0071] Embodiment 2, based on the same inventive concept as the value-oriented land component detection method in the foregoing embodiment, as Figure 2 shown, this application provides a value-oriented land component detection system. The system in the embodiment of this application and the method embodiment are based on the same inventive concept. Among them, the system includes:

[0072] A component conversion module 11, configured to connect to a spectrometer, embed and deploy a spectral configuration area in the control center, perform component conversion using scenario value indicators by obtaining the land use scenario, and set multi-spectral bands based on component correlations, including necessary components and inferior components; a spectral information determination module 12, configured to set a distributed sampling rule for the target land area, fuse it with the multi-spectral bands, perform response driving of the spectrometer and land detection, and determine the returned distributed spectral information; a land component detection result generation module 13, configured to use the distributed mean fusion of spectral features as the first preprocessing condition and the distributed difference measure of spectral features as the second preprocessing condition, execute the decision-making of spectral feature-land components in parallel, and determine the first global result and the second balance result as the land component detection result.

[0073] Furthermore, the system is also used to implement the following functions:

[0074] Connect to the land database, retrieve and mine the spectral relationship library of land component-spectral response, where the maximum correlation response band under surface reflectance and soil components is the mining target; build a spectral configuration area according to the spectral relationship library; in the front area of the control response end of the spectrometer control center, embed and deploy the spectral configuration area and establish a connection with the control response end.

[0075] Further, the system is also used to implement the following functions:

[0076] Based on the necessary components of the scenario value index, make a decision according to the spectral configuration area to determine a type of spectral band, where the type of spectral band has a first identifier; based on the inferior components of the scenario value index, make a decision according to the spectral configuration area to determine a second type of spectral band, where the second type of spectral band has a second identifier; aggregate the first type of spectral band and the second type of spectral band as the set multi-spectral band.

[0077] Further, the system is also used to implement the following functions:

[0078] According to the surface utilization value, conduct a distributed sampling plan to determine the distributed sampling rule; use the distributed sampling rule to constrain the detection position, use the multi-spectral band to constrain the detection conditions, and fuse to determine the spectral sampling strategy; based on the spectral sampling strategy, control and drive the spectrometer.

[0079] Further, the system is also used to implement the following functions:

[0080] The distributed spectral information includes N groups of spectral information corresponding to each distributed position, and each group of spectral information corresponds one-to-one with the multi-spectral band, where N is the value of the distributed position.

[0081] Further, the system is also used to implement the following functions:

[0082] For the N groups of spectral information, extract the spectral information of the same spectral band for each group and calculate the mean value of the spectral features, and reorganize to determine multiple mean spectral information, where the multiple mean spectral information characterizes the regional comprehensive state, and the multiple mean spectral information corresponds one-to-one with the multi-spectral band.

[0083] Further, the system is also used to implement the following functions:

[0084] For the N groups of spectral information, perform a difference calculation of the same spectral features between groups based on the neighborhood position to determine N - 1 groups of spectral difference information; according to the N - 1 groups of spectral difference information, construct a difference fluctuation map based on the spectral features.

[0085] Further, the system is also used to implement the following functions:

[0086] Taking spectral information as input and component characteristics as output, a component detection channel is constructed; the component detection channel is mirror-twinned to determine a parallel detection channel; the multiple mean spectral information is input into the first detection channel of the parallel detection channel, and the differential fluctuation map is input into the second detection channel of the parallel detection channel, and parallel decision-making is performed to determine the first global result and the second balance result.

[0087] Further, the system is also used to implement the following functions:

[0088] Based on the scenario value index of the land use scenario, the baseline land components are determined; according to the baseline land components, the qualified determination of the land component detection results is performed to generate land quality data.

[0089] It should be noted that the above order of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above description of specific embodiments of this specification is made. The processes depicted in the drawings do not necessarily require the specific order and continuous order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0090] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application shall be included within the protection scope of the present application.

[0091] This specification and the drawings are only exemplary descriptions of the present application and are considered to have covered any and all modifications, variations, combinations, or equivalents within the scope of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is intended to include these changes and modifications.

Claims

1. A value-oriented land component detection method, characterized in that, The method includes: Connect the spectrometer, embed and deploy a spectral configuration area in the control center. By obtaining the land use scenario, perform component conversion based on scenario value indicators, and set multi-spectral bands according to component correlations, including necessary components and inferior components. Set the distributed sampling rules for the target land area, fuse them with the multi-spectral bands, perform response driving of the spectrometer and land detection, and determine the distributed spectral information. Take the distributed mean fusion of spectral features as the first preprocessing condition and the distributed difference measure of spectral features as the second preprocessing condition, and perform the decision-making of spectral feature - land components in parallel to determine the first global result and the second balance degree result as the land component detection result.

2. The value-oriented land component detection method according to claim 1, wherein Embed and deploy a spectral configuration area in the control center, including: Connect to the land database, retrieve and mine the spectral relationship library of land component - spectral response, with the maximum correlation response band under surface reflectivity and soil components as the mining target. Build a spectral configuration area according to the spectral relationship library. In the front area of the control response end of the spectrometer control center, perform embedded deployment on the spectral configuration area and establish a connection with the control response end.

3. The value-oriented land component detection method according to claim 2, wherein Set multi-spectral bands according to component correlations, including: Based on the necessary components of the scenario value indicator, make a decision according to the spectral configuration area to determine a type of spectral band, where the type of spectral band has a first identifier. Based on the inferior components of the scenario value indicator, make a decision according to the spectral configuration area to determine a second type of spectral band, where the second type of spectral band has a second identifier. Aggregate the first type of spectral band and the second type of spectral band as the set multi-spectral bands.

4. The value-oriented land component detection method according to claim 1, characterized in that Set the distributed sampling rules for the target land area, fuse them with the multi-spectral bands, and perform response driving of the spectrometer, including: According to the surface utilization value, perform distributed sampling planning to determine the distributed sampling rules. Use the distributed sampling rules for detection position constraints and the multi-spectral bands for detection condition constraints, and fuse them to determine the spectral sampling strategy. Based on the spectral sampling strategy, control and drive the spectrometer.

5. The value-oriented land component detection method according to claim 1, characterized in that The distributed spectral information includes N groups of spectral information corresponding to each distributed position, and each group of spectral information corresponds one-to-one with the multi-spectral bands, where N is the value of the distributed position.

6. The value-oriented land component detection method according to claim 5, wherein, Take the distributed mean fusion of spectral features as the first preprocessing condition, including: For the N groups of spectral information, extract the spectral information of the same spectral band group by group and calculate the mean of spectral features, and reorganize to determine multiple mean spectral information, where the multiple mean spectral information represents the regional comprehensive state, and the multiple mean spectral information corresponds one-to-one with the multi-spectral bands.

7. The value-oriented land component detection method according to claim 6, wherein, Take the distributed difference measure of spectral features as the second preprocessing condition, including: For the N groups of spectral information, perform the difference calculation of the same spectral features between groups based on neighborhood positions to determine N - 1 groups of spectral difference information. According to the N - 1 groups of spectral difference information, construct a difference fluctuation map based on spectral features.

8. The value-oriented land component detection method according to claim 7, wherein Parallelly execute the decision-making of spectral features - land components to determine the first global result and the second balance degree result, including: Construct a component detection channel with spectral information as the input and component characteristics as the output; Mirror and twin the component detection channel to determine a parallel detection channel; Input the multiple mean spectral information into the first detection channel of the parallel detection channel, input the differential fluctuation map into the second detection channel of the parallel detection channel, and make parallel decisions to determine the first global result and the second balance degree result.

9. The value-oriented land component detection method according to claim 1, characterized in that After determining the land component detection result, including: Determine the baseline land component based on the scene value index of the land use scenario; According to the baseline land component, conduct a pass / fail judgment on the land component detection result to generate land quality data.

10. A value-oriented land component detection system, characterized in that, The system includes: A component conversion module, which is used to connect to a spectrometer, embed and deploy a spectral configuration area in the control center, convert components based on the obtained land use scenario with scene value indicators, and set multi-spectral bands according to component correlations, including necessary components and inferior components; A spectral information determination module, which is used to set the distributed sampling rules for the target land area, fuse with the multi-spectral bands, perform response driving of the spectrometer and land detection, and determine the back-transmitted distributed spectral information; A land component detection result generation module, which is used to take the distributed mean fusion of spectral features as the first preprocessing condition and the distributed differential metric of spectral features as the second preprocessing condition, parallelly execute the decision-making of spectral features - land components, and determine the first global result and the second balance degree result as the land component detection result.

Images