A corn straw information extraction method and system integrating prior knowledge in farmland

Abstract

The application provides a corn stalk information extraction method and system integrating prior knowledge in a farmland, and belongs to the technical field of agricultural remote sensing, and comprises the following steps: acquiring multi-temporal corn field remote sensing images of a target area; performing denoising and feature extraction on the multi-temporal corn field remote sensing images to obtain a plurality of remote sensing vegetation indexes; determining a first time period remote sensing vegetation index and a second time period remote sensing stalk index in the plurality of corn remote sensing vegetation indexes respectively; and obtaining target time period corn stalk distribution range information according to the first time period remote sensing vegetation index as prior knowledge and in combination with the second time period remote sensing stalk index. Through a remote sensing big data platform, the application comprehensively uses multi-time sequence and a plurality of corn field remote sensing vegetation indexes to extract corn stalks in a farmland in autumn and winter, and combines summer corn growing season remote sensing information as prior knowledge, so that the corn stalk satellite remote sensing extraction precision and extraction efficiency are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of agricultural remote sensing technology, and in particular to a corn stalk information extraction method and system integrated with prior knowledge. BACKGROUND

[0002] Corn stalks in farmland are harvested in autumn every year along with the ripening of corn. In the process of large-scale agricultural mechanization, corn and stalks are recycled at the same time, but because of the idle season in winter or the planting area of some scattered households after corn harvesting, the stalks are not timely processed and harvested, and there is a hidden danger of burning stalks in autumn or winter, which is a high-risk area for supervision by the agricultural and environmental protection departments. The autumn and winter seasons every year are the key period for the prevention and control of stalk burning, and the distribution survey of farmland stalks is the basic information for the prevention and control of stalk burning.

[0003] At present, the main method for the prevention and control of corn stalks is to adopt a large number of personnel to conduct field investigation, and to report the statistical survey information, which not only requires a large number of manpower, but also has low efficiency. With the introduction of auxiliary technologies such as remote sensing, remote sensing technology has been used for information extraction of corn stalks in farmland, but it is limited to direct extraction in autumn and winter, and does not consider that the stalks come from corn farmland in the summer growing season. Full consideration is given to the fact that in the images in autumn and winter, corn stalks are yellow, yellow-brown, and other foreign objects such as dry grass and tree leaves have the same spectrum as stalks, which will bring some errors to the remote sensing extraction of corn stalks.

[0004] In order to improve the application of the above-mentioned remote sensing technology in the extraction of corn stalks in farmland, improve the precision and efficiency, a new corn stalk information extraction method for farmland is needed. SUMMARY

[0005] The present application provides a corn stalk information extraction method and system integrated with prior knowledge for farmland, which solves the defects of low precision and efficiency in the extraction of corn stalk information in farmland in the prior art.

[0006] In a first aspect, the present application provides a corn stalk information extraction method integrated with prior knowledge for farmland, comprising:

[0007] Obtaining multi-temporal corn stalk remote sensing images of a target area;

[0008] Performing denoising and feature extraction on the multi-temporal corn stalk remote sensing images to obtain a plurality of remote sensing vegetation indices;

[0009] Respectively determining a first time period remote sensing vegetation index and a second time period remote sensing stalk index in the plurality of remote sensing vegetation indices;

[0010] According to the first time period remote sensing vegetation index as prior knowledge, combining the second time period remote sensing straw index, the target time period corn straw distribution range information is obtained through a classification algorithm.

[0011] According to the first time period remote sensing vegetation index as prior knowledge, combining the second time period remote sensing straw index, the target time period corn straw distribution range information is obtained through a classification algorithm.

[0012] The target time period corn straw distribution range information is converted to obtain a corn straw distribution map.

[0013] According to the first time period remote sensing vegetation index as prior knowledge, combining the second time period remote sensing straw index, the target time period corn straw distribution range information is obtained through a classification algorithm.

[0014] The target area of the second remote sensing image set of the sentinel is obtained.

[0015] Through the Google map engine GEE, the corn straw data in the first time period and the second time period in the second remote sensing image set of the sentinel is determined, and the multi-time phase corn straw remote sensing image is constituted.

[0016] According to the first time period remote sensing vegetation index as prior knowledge, combining the second time period remote sensing straw index, the target time period corn straw distribution range information is obtained through a classification algorithm.

[0017] The target area of the second remote sensing image set of the sentinel is obtained.

[0018] The corn phenological characteristics and the annual ground object spectral characteristics of the target area cloud-free image data are extracted, and the best classification time node and the plurality of remote sensing vegetation indexes are obtained.

[0019] According to the first time period remote sensing vegetation index as prior knowledge, combining the second time period remote sensing straw index, the target time period corn straw distribution range information is obtained through a classification algorithm.

[0020] The target area vector data is obtained, and the GEE area screening function is used to screen out the second image data of the sentinel which overlaps with the target area vector data.

[0021] The image set in the preset time period in the second image data of the sentinel is extracted through the time screening function.

[0022] Cloud amount screening is performed on the image set by using preset wave band data of the Sentinel-2 as cloud mask data, to obtain a screened image set;

[0023] The screened image set is spliced, and the screened image set is cropped according to a vector file, to obtain the target area cloud-free image data.

[0024] According to the corn crop growth data set, a target area crop annual phenology information map is obtained;

[0025] According to the corn crop growth data set, a target area crop annual phenology information map is obtained;

[0026] According to the corn crop growth data set, a target area crop annual phenology information map is obtained;

[0027] According to the corn crop growth data set, a target area crop annual phenology information map is obtained;

[0028] According to the corn crop growth data set, a target area crop annual phenology information map is obtained;

[0029] According to the corn crop growth data set, a target area crop annual phenology information map is obtained;

[0030] According to the corn crop growth data set, a target area crop annual phenology information map is obtained;

[0031] According to the corn crop growth data set, a target area crop annual phenology information map is obtained;

[0032] According to the corn crop growth data set, a target area crop annual phenology information map is obtained;

[0033] According to the corn crop growth data set, a target area crop annual phenology information map is obtained;

[0034] Based on the sentinel No. 2 fourth wave band reflectivity value, sentinel No. 2 fifth wave band reflectivity value and the sentinel No. 2 twelfth wave band reflectivity value, and the fourth wave band adjustment coefficient, the fifth wave band adjustment coefficient and the twelfth wave band adjustment coefficient, the second time period is obtained. Accumulation near red straw index.

[0035] According to the first time period remote sensing vegetation index as prior knowledge, combining the second time period remote sensing straw index, the target time period corn straw distribution range information is obtained through classification algorithm, including:

[0036] The first time period remote sensing vegetation index and the second time period remote sensing straw index are fused to obtain feature fusion data;

[0037] The feature fusion data is input into the trained maximum likelihood classification model for classification to obtain the target time period corn straw distribution range information.

[0038] According to the first time period remote sensing vegetation index as prior knowledge, combining the second time period remote sensing straw index, the target time period corn straw distribution range information is obtained through classification algorithm, including:

[0039] The target time period corn straw distribution range information is converted into a vector file;

[0040] The vector file is superimposed with a set of sentinel No. 2 remote sensing images to obtain the corn straw distribution map.

[0041] Secondly, the present application also provides a corn straw information extraction system for farmland integrating prior knowledge, comprising:

[0042] The acquisition module is used for acquiring multi-temporal corn straw remote sensing images of a target area;

[0043] The extraction module is used for denoising and feature extraction of the multi-temporal corn straw remote sensing images to obtain a plurality of remote sensing vegetation indexes;

[0044] The determination module is used for determining a first time period vegetation index and a second time period remote sensing straw index in the plurality of remote sensing vegetation indexes, respectively;

[0045] The calculation module is used for obtaining target time period corn straw distribution range information according to the first time period remote sensing vegetation index and the second time period remote sensing straw index.

[0046] In a third aspect, the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method for extracting corn stalk information in farmland by integrating prior knowledge according to any one of the above aspects.

[0047] In a fourth aspect, the present application provides a non-transitory computer-readable storage medium, which stores a computer program, wherein the computer program is executable on a processor to implement the method for extracting corn stalk information in farmland by integrating prior knowledge according to any one of the above aspects.

[0048] In a fifth aspect, the present application provides a computer program product, comprising a computer program, wherein the computer program is executable on a processor to implement the method for extracting corn stalk information in farmland by integrating prior knowledge according to any one of the above aspects.

[0049] The method and system for extracting corn stalk information in farmland by integrating prior knowledge provided by the present application can improve the extraction accuracy and efficiency of corn stalk satellite remote sensing by comprehensively extracting corn stalks in farmland in autumn and winter through remote sensing big data platform, multiple time sequences and multiple corn stalk vegetation indexes, and combining summer corn remote sensing information as prior knowledge. BRIEF DESCRIPTION OF DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the present application or prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0051] Figure 1 is one of the flowcharts of the method for extracting corn stalk information in farmland by integrating prior knowledge provided by the present application;

[0052] Figure 2 is another flowchart of the method for extracting corn stalk information in farmland by integrating prior knowledge provided by the present application;

[0053] Figure 3 is a structural schematic diagram of the system for extracting corn stalk information in farmland by integrating prior knowledge provided by the present application;

[0054] Figure 4 is a structural schematic diagram of the electronic device provided by the present application. DETAILED DESCRIPTION

[0055] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below in conjunction with the drawings in the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the protection scope of the present application.

[0056] In view of the deficiencies in the existing corn straw information extraction technology, the present application provides a new farmland corn straw information extraction method, which combines prior knowledge of corn straw growth and multi-temporal remote sensing data to effectively improve the extraction accuracy and efficiency of farmland corn straw.

[0057] Figure 1 is one of the flowcharts of the farmland corn straw information extraction method provided by the present application, as shown in Figure 1 , comprising:

[0058] Step 100: acquiring multi-temporal corn straw remote sensing images of a target region;

[0059] Step 200: denoising and feature extraction are performed on the multi-temporal corn straw remote sensing images to obtain a plurality of remote sensing vegetation indexes;

[0060] Step 300: a first time period remote sensing vegetation index and a second time period remote sensing straw index in the plurality of remote sensing vegetation indexes are determined respectively;

[0061] Step 400: according to the first time period remote sensing vegetation index as prior knowledge, combining the second time period remote sensing straw index, the target time period corn straw distribution range information is obtained through a classification algorithm.

[0062] Firstly, the remote sensing big data platform of Google Earth Engine (GEE) is adopted to acquire the Sentinel-2 remote sensing image data of the typical time period (usually July and August) of corn growth season and the mature time period (usually November) of corn in the selected target region, and a series of pretreatments are performed; according to the specific band in the Sentinel-2 data in the GEE platform, such as the QA60 band, the cloud removal operation is performed on the Sentinel-2 remote sensing image data, and the image with large cloud cover is removed; then the specified vegetation index in the typical time period of corn growth season is adopted as prior knowledge, the spectral characteristics of the straight straw and the flat straw in the farmland area of the target region are further calculated to obtain the straw index in the mature time period of corn, and the random forest algorithm is adopted to extract the corn straw distribution range information.

[0063] It should be noted that GEE has a global satellite database, provides a cloud platform that can be online visualized, calculated and analyzed, can access satellite images and other earth observation data databases, and provides sufficient computing power to process these data.

[0064] The Sentinel-2A satellite is the second satellite of the Global Environmental and Security Monitoring Plan, and the Sentinel-2 satellite carries a multi-spectral imager, which can cover 13 spectral bands, has a width of 290 kilometers, has a spatial resolution of 10 meters, and has a revisit period of 10 days. From visible light and near infrared to short wave infrared, different spatial resolutions, in optical data, Sentinel-2 data is the only data containing three bands in the red edge range, which is usually used to monitor vegetation health information.

[0065] The present application extracts corn stalks in the autumn and winter seasons by comprehensively using multi-time sequence and multiple corn stalk vegetation indexes through a remote sensing big data platform, and combines summer corn growth season remote sensing information as prior knowledge, thereby improving the satellite remote sensing extraction accuracy and extraction efficiency of corn stalks.

[0066] Based on the above embodiment, the multi-time phase corn stalk remote sensing image of the target area is obtained, comprising:

[0067] Obtain a set of Sentinel-2 remote sensing images of the target area;

[0068] Determine the corn stalk data in the first time period and the second time period in the set of Sentinel-2 remote sensing images through the Google Earth Engine (GEE), and constitute the multi-time phase corn stalk remote sensing image.

[0069] The present application first obtains a multi-time phase corn stalk remote sensing image through a GEE platform, extracts Sentinel-2 L2A level remote sensing images of the target area in July and August of the summer corn growth season and in November of the winter after the corn matures, respectively. Since the GEE platform contains all Sentinel-2 images since 2019, the data is processed by orthorectification and sub-pixel level geometric correction, and contains atmospheric bottom reflectivity data after atmospheric correction, so no additional preprocessing work is required.

[0070] Specifically, the data of the second generation of sentinel obtained through the GEE platform mainly contains 16 bands, including: B1 (443.9 nm), B2 (496.6 nm), B3 (560.0 nm), B4 (664.5 nm), B5 (703.9 nm), B6 (740.2 nm), B7 (782.5 nm), B8 (835.1 nm), B8b (864.8 nm), B9 (945.0 nm), B10 (1373.5 nm), B11 (1613.7 nm), B12 (2202.4 nm), QA10, QA20 and QA60, wherein B1-B12 are original data of the second generation of sentinel, and the application adopts QA60 as cloud mask data for subsequent processing.

[0071] The application adopts the second generation of sentinel data with processed vegetation index information, and has the advantages of comprehensive information and no need for further processing.

[0072] Based on any of the above embodiments, the multi-temporal corn stalk remote sensing image is denoised and feature extracted to obtain a plurality of remote sensing vegetation indexes, including:

[0073] The multi-temporal corn stalk remote sensing image is spliced and cloud layer removed to obtain target area cloud-free image data;

[0074] The corn phenological features and annual ground object spectral features of the target area cloud-free image data are extracted to obtain the best classification time node and the plurality of remote sensing vegetation indexes.

[0075] The multi-temporal corn stalk remote sensing image is spliced and cloud layer removed to obtain target area cloud-free image data, including:

[0076] The target area vector data is obtained, and the GEE area screening function is used to screen out the second generation of sentinel image data overlapping with the target area vector data;

[0077] The image set in the preset time period in the second generation of sentinel image data is extracted through the time screening function;

[0078] The image set is cloud amount screened using the preset band data of the second generation of sentinel as cloud mask data to obtain a screened image set;

[0079] The screened image set is spliced, and the screened image set is cropped according to the vector file to obtain the target area cloud-free image data.

[0080] The corn phenological features and annual ground object spectral features of the target area cloud-free image data are extracted to obtain the best classification time node and the plurality of remote sensing vegetation indexes, including:

[0081] Obtaining a crop growth and development dataset;

[0082] Based on the crop growth and development dataset, a target area crop annual phenology information map is obtained;

[0083] Based on the target area crop annual phenology information map, the corn phenology feature is extracted;

[0084] The corn phenology feature is analyzed by annual ground object spectral feature analysis, and the optimal classification time node and the plurality of remote sensing vegetation indexes are obtained.

[0085] Specifically, the application adopts a GEE big data platform to perform satellite remote sensing image splicing and cloud removal operation.

[0086] In view of the fact that satellite images are greatly affected by clouds, the shielding of clouds leads to the failure to obtain image band information, especially in summer, when there is more rain and thicker clouds, so that it is difficult to obtain high-quality images. The influence of cloud shielding on images can be effectively reduced through the GEE platform. With the help of the vector data uploading, image screening, time screening and cutting functions of the GEE, first, the vector data of the target area is uploaded, the GEE area screening function is used to screen out the Sentinel-2 image data overlapping with the vector data, then the time screening function is used to set the image set in a specific time period, and then the cloud amount is screened according to the value of the QA60 band in the Sentinel-2, a small amount of clouds in the image is programmed to remove the clouds, the clouds are masked to maximize the retention of complete images, finally the processed images are spliced, and the vector file is cut to obtain the cloud-free image data of the target area, which is used for subsequent feature extraction processing.

[0087] Further, the crop growth and development dataset provided by the meteorological data network and the local statistical agency yearbook is used to summarize the annual phenology information map of the main crops in the target area range by means of spectral analysis and on-site investigation inquiry.

[0088] Since the corn stalks in autumn and winter fields must belong to the summer corn farmland planting range, the summer corn remote sensing information is taken as prior knowledge, and the normalized difference vegetation index (NDVI) and the enhanced vegetation index-2 (EVI2) of the corn growth season at the peak or larger value are taken as features for extracting the corn stalks in autumn and winter fields. Through annual ground object spectral feature analysis, it can be found that there is a significant spectral difference between the spectral of the straw in November and the winter wheat area, which can be used as the best time to distinguish the corn stalks in autumn and winter fields.

[0089] The application adopts the GEE platform to perform cloud removal operation on data, reduces interference factors in the data, and adopts corn phenological characteristics and annual ground object spectral characteristics analysis, so that prior knowledge of corn straw distribution and the best time for distinguishing corn straw in autumn and winter are more accurately obtained.

[0090] Based on any one of the above embodiments, the first time period remote sensing vegetation index and the second time period remote sensing straw index in the plurality of corn straw vegetation indexes are determined respectively, including:

[0091] Based on the sentinel-2 fourth band reflectance value and the sentinel-2 eighth band reflectance value in the cloud-free image data of the target area, a first time period normalized vegetation index is obtained;

[0092] Based on the gain factor, the first aerosol impedance coefficient, the second aerosol impedance coefficient and the canopy background adjustment factor, and the sentinel-2 second band reflectance value, the sentinel-2 fourth band reflectance value and the sentinel-2 eighth band reflectance value, a first time period enhanced vegetation index is obtained;

[0093] Based on the sentinel-2 fourth band reflectance value and the sentinel-2 eighth band reflectance value, a second time period normalized short-wave infrared straw index is obtained;

[0094] Based on the sentinel-2 ninth band reflectance value and the sentinel-2 twelfth band reflectance value in the cloud-free image data of the target area, a second time period superimposed infrared straw index is obtained;

[0095] Based on the sentinel-2 fourth band reflectance value, the sentinel-2 fifth band reflectance value and the sentinel-2 twelfth band reflectance value, and the fourth band adjustment coefficient, the fifth band adjustment coefficient and the twelfth band adjustment coefficient, a second time period superimposed near infrared straw index is obtained.

[0096] Specifically, the application adopts NDVI and EVI two vegetation indexes of the cloud-removed remote sensing image of corn growth season in July and August in summer as important features of corn straw in autumn and winter, and according to the research on corn, it is not difficult to find that whether it is spring corn or summer corn, the corn growth in July and August is at the optimal time point, which can effectively highlight the corn planting situation, and the corresponding vegetation index calculation formula is as follows:

[0097]

[0098]

[0099] Wherein, B4 is the B4 band reflectivity value in the satellite image of Sentinel-2, B8 is the B8 band reflectivity value in the satellite image of Sentinel-2, B4 is the red light band of Sentinel-2, and B8 is the near-infrared band of Sentinel-2; G is a gain factor, and the value is 2.5; C1 and C2 are aerosol impedance coefficients, and the values are 6 and 7.5; and L is a canopy background adjustment factor, and the value is 1.

[0100] In addition, the application also uses three kinds of straw vegetation indexes in November after corn matures as straw extraction features in autumn and winter, including normalized difference short wave infrared straw index (NDSSI), accumulation infrared straw index (AIRSI) and product near-infrared straw index (PNISI), and the specific straw index calculation formula is as follows:

[0101]

[0102]

[0103]

[0104] Wherein, B4 is the B4 band reflectivity value in the satellite image of Sentinel-2, B5 is the B5 band reflectivity value in the satellite image of Sentinel-2, B8 is the B8 band reflectivity value in the satellite image of Sentinel-2, B9 is the B9 band reflectivity value in the satellite image of Sentinel-2, and B12 is the B12 band reflectivity value in the satellite image of Sentinel-2. a is a B4 band adjustment coefficient, and the value is 1.8, b is a B5 band adjustment coefficient, and the value is 1.5, and c is a B12 band adjustment coefficient, and the value is 10000.

[0105] The application extracts corn straw in autumn and winter by comprehensively using three time sequences and five kinds of straw vegetation indexes, and compared with existing satellite remote sensing extraction, the precision and extraction efficiency of corn straw extraction are improved.

[0106] According to any one of the above embodiments, the first time period remote sensing vegetation index is used as prior knowledge, the second time period remote sensing straw index is combined, and the classification algorithm is used to obtain the target time period corn straw distribution range information, including:

[0107] The first time period remote sensing vegetation index and the second time period remote sensing straw index are fused to obtain feature fusion data;

[0108] The feature fusion data is input into a trained maximum likelihood classification model for classification to obtain the corn stalk distribution range information in the target time period.

[0109] Specifically, according to the analysis of the foregoing embodiments, the time period when the summer corn grows best is from July to August, and the corn stalks in the farmland have obvious differences from other ground objects in the middle of November.

[0110] The application uses the NDVI data and EVI data in July and August as prior knowledge, and combines the three stalk index data in November to form the feature data of the corn stalks in the autumn and winter for extraction.

[0111] The calculation and download of the above-mentioned five vegetation indexes are completed through programming on the GEE platform, the band values are fused by using the ENVI software, and the corn stalks in the autumn and winter farmland are extracted. Based on the measured data, a training set of maximum likelihood classification is constructed, the combined feature data after fusion is put into the maximum likelihood classifier for classification, and the distribution range information of the corn stalks in the autumn and winter farmland is obtained.

[0112] The application has the characteristics of objectivity, accuracy and high training efficiency by fusing a plurality of vegetation indexes and using the maximum likelihood classification model for training and classification.

[0113] Based on any one of the foregoing embodiments, after obtaining the corn stalk distribution range information in the target time period according to the first time period remote sensing vegetation index as prior knowledge, combining the second time period remote sensing stalk index, and through a classification algorithm, the following is included:

[0114] The corn stalk distribution range information in the target time period is converted to obtain a corn stalk distribution map.

[0115] The conversion of the corn stalk distribution range information in the target time period to obtain a corn stalk distribution map includes:

[0116] The corn stalk distribution range information in the target time period is converted into a vector file;

[0117] The vector file is superimposed with a set of Sentinel-2 remote sensing images to obtain the corn stalk distribution map.

[0118] Specifically, since the generated corn stalk distribution range information in the target time period is a raster image, in order to facilitate identification in actual application, the raster-to-vector tool is used to convert the corn stalk distribution range information in the target time period into a farmland corn stalk distribution map in shp format, i.e. a vector file, which is completed by using the ArcGIS software.

[0119] The shp file is a shape file, which is developed by ESRI, and an ESRI (Environmental Systems Research Institute) shape file includes a main file, an index file, and a dBASE table, wherein the suffix of the main file is.shp. The shp file is composed of a fixed-length file header and then a variable-length record, and each variable-length record is composed of a fixed-length record header and then a variable-length record content.

[0120] Finally, the generated vector file, i.e., the farmland corn straw distribution map in the shp format, is superimposed on the second remote sensing image of the satellite, to generate the final corn straw distribution map.

[0121] As shown in the flowchart of the farmland corn straw information extraction method, the complete steps of the present application are as follows: Figure 2

[0122] (1) Obtain multi-temporal satellite remote sensing images through the GEE platform;

[0123] (2) Satellite remote sensing image splicing, cutting and cloud removal;

[0124] (3) Analysis of corn phenological characteristics and annual ground object spectral characteristics;

[0125] (4) Calculation of summer corn growing season remote sensing vegetation index;

[0126] (5) Calculation of straw index in autumn and winter;

[0127] (6) Extraction of farmland corn straw from multi-temporal images;

[0128] (7) Conversion of the extraction result of farmland corn straw in autumn and winter into a raster vector;

[0129] (8) Production of a farmland corn straw distribution map in autumn and winter.

[0130] The present application uses satellite remote sensing images of corn in the growing season in July and August as prior knowledge for straw extraction, and uses three straw indexes in November after the corn matures to extract farmland corn straw in autumn and winter, thereby improving the extraction accuracy and efficiency of farmland corn straw, reducing the time and cost of manual investigation, and providing data support for straw burning supervision.

[0131] The integrated prior knowledge farmland corn straw information extraction system provided by the present application is described below, and the integrated prior knowledge farmland corn straw information extraction system described below can be correspondingly referred to the integrated prior knowledge farmland corn straw information extraction method described above.

[0132] Figure 3 ​is a structural schematic diagram of the corn straw information extraction system integrating prior knowledge provided by the application, as shown in Figure 3 The application provides a corn straw information extraction system integrating prior knowledge.

[0133] The acquisition module 31 is configured to acquire multi-temporal corn straw remote sensing images of a target area. The extraction module 32 is configured to perform denoising and feature extraction on the multi-temporal corn straw remote sensing images to obtain a plurality of remote sensing vegetation indexes. The determination module 33 is configured to determine a first time period remote sensing vegetation index and a second time period remote sensing straw index in the plurality of remote sensing vegetation indexes, respectively. The calculation module 34 is configured to obtain target time period corn straw distribution range information by a classification algorithm according to the first time period remote sensing vegetation index as prior knowledge and in combination with the second time period remote sensing straw index.

[0134] The application improves the corn straw satellite remote sensing extraction precision and efficiency by comprehensively extracting corn straw in the autumn and winter seasons through a remote sensing big data platform, multi-time sequences and a plurality of corn straw vegetation indexes, and combining summer corn growth season remote sensing information as prior knowledge.

[0135] Figure 4 An example of an entity structure schematic diagram of an electronic device is shown in Figure 4 The electronic device can include a processor 410, a communications interface 420, a memory 430 and a communications bus 440, wherein the processor 410, the communications interface 420 and the memory 430 complete mutual communication through the communications bus 440. The processor 410 can call logical instructions in the memory 430 to execute a corn straw information extraction method integrating prior knowledge, which includes: acquiring multi-temporal corn straw remote sensing images of a target area; performing denoising and feature extraction on the multi-temporal corn straw remote sensing images to obtain a plurality of remote sensing vegetation indexes; determining a first time period remote sensing vegetation index and a second time period remote sensing straw index in the plurality of remote sensing vegetation indexes, respectively; and obtaining target time period corn straw distribution range information by a classification algorithm according to the first time period remote sensing vegetation index as prior knowledge and in combination with the second time period remote sensing straw index.

[0136] Further, the logic instructions in the memory 430 described above can be implemented in the form of software functional units and sold or used as independent products, and can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the parts that make contributions to the prior art or parts of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0137] In another aspect, the present application also provides a computer program product, which comprises a computer program, the computer program can be stored on a non-transitory computer readable storage medium, and the computer program can be executed by a processor to enable a computer to execute the integrated prior knowledge farmland corn straw information extraction method provided by the above-mentioned methods. The method comprises: acquiring multi-temporal corn straw remote sensing images of a target area; performing denoising and feature extraction on the multi-temporal corn straw remote sensing images to obtain a plurality of remote sensing vegetation indexes; determining a first time period remote sensing vegetation index and a second time period remote sensing straw index in the plurality of remote sensing vegetation indexes respectively; according to the first time period remote sensing vegetation index as prior knowledge, combining the second time period remote sensing straw index, and obtaining target time period corn straw distribution range information through a classification algorithm.

[0138] In yet another aspect, the present application also provides a non-transitory computer readable storage medium, which stores a computer program. The computer program is executed by a processor to implement the integrated prior knowledge farmland corn straw information extraction method provided by the above-mentioned methods. The method comprises: acquiring multi-temporal corn straw remote sensing images of a target area; performing denoising and feature extraction on the multi-temporal corn straw remote sensing images to obtain a plurality of remote sensing vegetation indexes; determining a first time period remote sensing vegetation index and a second time period remote sensing straw index in the plurality of remote sensing vegetation indexes respectively; according to the first time period remote sensing vegetation index as prior knowledge, combining the second time period remote sensing straw index, and obtaining target time period corn straw distribution range information through a classification algorithm.

[0139] The device embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed to multiple network units. Part or all of the modules can be selected to achieve the purposes of the embodiments according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0140] Through the description of the above embodiments, those skilled in the art can clearly understand that the embodiments can be realized by means of software and the necessary general hardware platform, and of course can also be realized by hardware. Based on such understanding, the above technical solutions can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in each embodiment or some parts of the embodiments.

[0141] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An integrated prior knowledge corn field straw information extraction method, characterized in that, The method comprises the following steps: acquiring multi-temporal corn stalk remote sensing images of a target area; performing denoising and feature extraction on the multi-temporal corn stalk remote sensing images to obtain a plurality of remote sensing vegetation indexes; determining a first time period remote sensing vegetation index and a second time period remote sensing stalk index in the plurality of remote sensing vegetation indexes respectively; obtaining target time period corn stalk distribution range information by a classification algorithm according to the first time period remote sensing vegetation index as prior knowledge and in combination with the second time period remote sensing stalk index; wherein the first time period refers to a typical time period in the corn growing season, and the second time period refers to a corn maturation time period; performing denoising and feature extraction on the multi-temporal corn stalk remote sensing images to obtain a plurality of remote sensing vegetation indexes, which comprises the following steps: performing splicing and cloud removal processing on the multi-temporal corn stalk remote sensing images to obtain target area cloud-free image data; extracting corn phenological features and annual ground object spectral features of the target area cloud-free image data to obtain a best classification time node and the plurality of remote sensing vegetation indexes; the step of extracting corn phenological features and annual ground object spectral features of the target area cloud-free image data to obtain a best classification time node and the plurality of remote sensing vegetation indexes comprises the following steps: acquiring a crop growth and development dataset; obtaining a target area crop annual phenological information map based on the crop growth and development dataset; extracting the corn phenological features based on the target area crop annual phenological information map; performing annual ground object spectral feature analysis on the corn phenological features to obtain the best classification time node and the plurality of remote sensing vegetation indexes; the step of obtaining target time period corn stalk distribution range information by a classification algorithm according to the first time period remote sensing vegetation index as prior knowledge and in combination with the second time period remote sensing stalk index comprises the following steps: performing feature fusion on the first time period remote sensing vegetation index and the second time period remote sensing stalk index to obtain feature fusion data; inputting the feature fusion data into a trained maximum likelihood classification model for classification to obtain the target time period corn stalk distribution range information.

2. The method for extracting information from farmland corn stalks by integrating prior knowledge according to claim 1, characterized in that, after the step of obtaining target time period corn stalk distribution range information by a classification algorithm according to the first time period remote sensing vegetation index as prior knowledge and in combination with the second time period remote sensing stalk index, the method further comprises the following step: converting the target time period corn stalk distribution range information to obtain a corn stalk distribution map. 3.The method of claim 1, wherein the step of acquiring multi-temporal corn stalk remote sensing images of a target area comprises the following steps: acquiring a set of Sentinel-2 remote sensing images of the target area; determining corn stalk data in the set of Sentinel-2 remote sensing images that are in a first time period and a second time period by using a Google Earth Engine (GEE) to form the multi-temporal corn stalk remote sensing images. 4.The method of claim 1, wherein the step of performing splicing and cloud removal processing on the multi-temporal corn stalk remote sensing images to obtain target area cloud-free image data comprises the following steps: acquiring target area vector data, and screening out Sentinel-2 image data that overlaps with the target area vector data by using a GEE area screening function; extracting a set of images in a preset time period from the Sentinel-2 image data by using a time screening function; Cloud amount screening is performed on the image set by using preset wave band data of the Sentinel-2 as cloud mask data, to obtain a screened image set; The screened image set is spliced, and the screened image set is cropped according to a vector file, to obtain the target area cloud-free image data. 5.The method of claim 1, wherein, The first time period remote sensing vegetation index and the second time period remote sensing straw index in the plurality of remote sensing vegetation indexes are determined respectively, including: A first time period normalized vegetation index is obtained based on a fourth wave band reflectivity value and an eighth wave band reflectivity value of the Sentinel-2 in the target area cloud-free image data; A first time period enhanced vegetation index is obtained based on a gain factor, a first aerosol impedance coefficient, a second aerosol impedance coefficient and a canopy background adjustment factor, and a second wave band reflectivity value, the fourth wave band reflectivity value and the eighth wave band reflectivity value of the Sentinel-2; A second time period normalized short wave infrared straw index is obtained based on the fourth wave band reflectivity value and the eighth wave band reflectivity value of the Sentinel-2; A second time period superimposed infrared straw index is obtained based on a ninth wave band reflectivity value and a twelfth wave band reflectivity value of the Sentinel-2 in the target area cloud-free image data; A second time period superimposed near infrared straw index is obtained based on the fourth wave band reflectivity value, a fifth wave band reflectivity value and the twelfth wave band reflectivity value of the Sentinel-2, and a fourth wave band adjustment coefficient, a fifth wave band adjustment coefficient and a twelfth wave band adjustment coefficient. 6.The method of claim 2, wherein, The target time period corn straw distribution range information is converted to obtain a corn straw distribution map, including: The target time period corn straw distribution range information is converted into a vector file; The vector file is superimposed on the Sentinel-2 remote sensing image set to obtain the corn straw distribution map.

7. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the program to realize the integrated prior knowledge farmland corn straw information extraction method according to any one of claims 1 to 6.

Images