CN115861844A - Rice early-stage remote sensing identification method based on planting probability - Google Patents
Rice early-stage remote sensing identification method based on planting probability Download PDFInfo
- Publication number
- CN115861844A CN115861844A CN202211690242.9A CN202211690242A CN115861844A CN 115861844 A CN115861844 A CN 115861844A CN 202211690242 A CN202211690242 A CN 202211690242A CN 115861844 A CN115861844 A CN 115861844A
- Authority
- CN
- China
- Prior art keywords
- rice
- time sequence
- probability
- planting
- vegetation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Image Processing (AREA)
Abstract
The invention discloses a rice early remote sensing identification method based on planting probability, which comprises the following steps: step 1: selecting an image time window; step 2: dividing the image to obtain a land object; and step 3: extracting object-level time sequence characteristic parameters; and 4, step 4: distinguishing and marking vegetation and non-vegetation areas; and 5: calculating rice planting probabilities under different time sequence characteristics; step 6: calculating a rice probability index TRPI based on time sequence characteristics; and 7: and (4) threshold classification, and extracting rice planting distribution. According to the method, the time sequence characteristics of the rice transplanted to the tillering stage are extracted by utilizing SAR image data, the probability of planting the rice under each time sequence characteristic index is calculated by a formula, meanwhile, the vegetation index calculated by using multispectral data of the rice previous crop in the vigorous stage is used for marking an obvious non-rice planting area, so that the rice probability index TRPI based on the time sequence characteristics is integrated and constructed, and finally, threshold binary classification is utilized for realizing early rice identification aiming at the TRPI calculation result.
Description
Technical Field
The invention belongs to the technical field of remote sensing monitoring and identification, and particularly relates to a rice early-stage remote sensing identification method based on planting probability.
Background
Timely mastering the planting distribution of the rice has important significance for relevant agricultural applications such as crop declaration and subsidy check, crop growth condition monitoring, crop insurable area investigation of agricultural insurance institutions, insurance coverage improvement, insurance authenticity check and the like. In recent years, with the rapid development of remote sensing technology, application research of remote sensing in the agricultural field is deepened continuously, and at present, researches for identifying and extracting crops by utilizing remote sensing are more and more, such as identification of grain crops such as wheat, rice, corn, soybean and the like. Most of related researches utilize the remote sensing data characteristics of crops in the whole growth period or the characteristics of crops in the vigorous growth period to identify the crops, and the researches and inventions for identifying the crops in the early planting and growth stages of the crops are relatively few, particularly the early identification of rice, and related researches are not discovered temporarily. In the early stage of rice planting, the influence of weather is caused, the obtained optical image quality is low, the data is less, and the biomass content is less, so that the difficulty in identifying the rice in the early stage of rice planting is high.
The Synthetic Airborne Radar (SAR) data has certain sensitivity to water and is not influenced by cloud and rain weather, the image time resolution is high, and the synthetic SAR data has certain advantages in agricultural application. The rice planting process generally includes the steps of soil preparation, field soaking, transplanting, rice growth (tillering, node pulling, ear bearing, ear sprouting and maturing) and the like, the water content in the field and the exposure condition of the surface water body of the rice field change in the early stage of rice planting and growth, and a basis is provided for capturing the rice planting condition through multi-time-phase SAR data. Therefore, the time sequence SAR image can be used as an effective data source for extracting early rice.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims to provide a rice early remote sensing identification method based on planting probability. The invention provides a rice probability index (TRPI) based on time sequence characteristics, and forms an early identification method of rice. The method extracts the time sequence characteristics of the rice transplanted to the tillering stage by utilizing SAR image data, calculates the probability of planting the rice under each time sequence characteristic index, and simultaneously marks an obvious non-rice planting area by using the vegetation index calculated by multispectral image data of the vigorous stage (history) of previous crops of the rice, thereby integrating and constructing TRPI, and finally realizing early rice identification by utilizing threshold value binary classification. The problem that the rice is difficult to extract by remote sensing identification in the early stage of rice planting and growth is solved.
In order to achieve the purpose of the invention, the invention adopts the technical scheme that:
the method for early remote sensing identification of the rice based on the planting probability comprises the following steps:
step 1: selecting a video time window
Selecting a rice planting and early growth stage (rice is transplanted to a tillering stage, generally from 6 months to 7 months) as a time window for identifying early rice, selecting time sequence SAR image data (Sentinel-1) in the time window as a data source for early rice identification, and spacing adjacent image time for 5-7 days; preprocessing the time sequence SAR image to obtain a backscattering coefficient of the target area in a VH polarization mode of each phase SAR image;
step 2: image segmentation to obtain land object
Performing multi-scale segmentation by utilizing a Sentinel-2 multispectral image of rice previous crop in the vigorous growth stage to obtain a land parcel object, and setting segmentation scales and factors according to the image resolution and the area size;
and 3, step 3: extracting object-level temporal feature parameters
Calculating the average value of backscattering coefficients of all pixels in each block object of each phase SAR VH polarization mode image to be used as the backscattering coefficient of the block object, and then extracting the maximum Value (VH) of the same block object among time sequence images Max ) Minimum Value (VH) Min ) Mean Value (VH) Mean ) Finally, generating a time sequence maximum value graph, a time sequence minimum value graph and a time sequence average value graph;
and 4, step 4: differentiating and marking vegetation and non-vegetation areas
Calculating a normalized vegetation index NDVI by using the multispectral image, setting a threshold value for distinguishing and marking vegetation and non-vegetation areas;
and 5: calculating the rice planting probability under different time sequence characteristics
Respectively calculating the rice planting probability of each plot object of the time sequence SAR data under the characteristics of the time sequence maximum value, the time sequence minimum value and the time sequence mean value;
step 6: calculating rice probability index TRPI based on time sequence characteristics
Constructing a rice probability index TRPI based on time sequence characteristics, and calculating the early rice planting probability by using the index;
and 7: threshold classification, extraction of rice planting distribution
Setting condition TRPI ≥ T 1 And if the condition is true, determining the rice is the rice, otherwise, determining the rice is non-rice, and realizing early rice identification.
Further, in step 1, performing orbit correction, thermal noise removal, radiometric calibration, filtering, terrain correction, decibel processing, registration and clipping preprocessing on the time-series SAR image to obtain a backscattering coefficient of the target area in a VH polarization mode of each phase of the SAR image.
Further, the method of step 4 specifically includes the following substeps:
step 4-1, according to a formula:
obtaining NDVI; where ρ is Red 、ρ NIR Reflectance for red and near infrared bands, respectively;
step 4-2, according to the function:
setting conditions NDVI to be more than or equal to T 0 When the condition is true, indicating a vegetation area, namely an area where rice can be planted, and marking as 1; when the condition is false, the non-vegetation area is shown, namely the obvious non-planting rice area, the rice planting probability is 0, and the mark is 0.
Further, in step 5, according to the formula:
obtaining the probability f (VH) that the object to be calculated is rice under different time sequence characteristic indexes i ) (ii) a Wherein the content of the first and second substances,is the mean value, x, of the rice sample under the current time sequence characteristics i For the value of the backscattering coefficient of the object to be calculated in the current time series characteristic index map, (VH) i ) max 、(VH i ) min Respectively obtaining the maximum value and the minimum value after the non-vegetation areas in the current time sequence characteristic index map are removed; />For reflecting the difference between the object to be recognized and the rice, the smaller the value, the smaller the difference, i.e. the greater the probability that it is rice, and the greater the value of the difference>Normalize it and pick it up>Indicates the probability that the object to be calculated is rice, the larger the value, the larger the probability that the object is rice, and f (VH) i ) The value range of (a) is 0 to 1. When Max, min and Mean are taken as i, according to the formula:
obtaining rice planting probability f (VH) under characteristic indexes of maximum value, minimum value and mean value of time sequence Max )、f(VH Min )、f(VH Mean )。
Further, in step 6, according to the formula:
TRPI=f()×f( Max )×f(VH Min )×f(VH Mean )
acquiring a rice probability index TRPI; wherein f (NDVI) is the vegetation and non-vegetation area marked in step 4, and the values are 1 and 0,f (VH) Max )、f(VH Min )、f(VH Mean ) And 5, respectively calculating the rice planting probability under the time sequence maximum value, minimum value and mean value characteristics calculated in the step 5.
The invention has the beneficial effects that:
1. in the early stage of rice planting (tillering stage), the characteristics of rice growth change and paddy field surface water body appearance condition change in the early stage of rice planting and growth are taken as breakthrough, and the synthetic aperture radar data which is not influenced by weather is utilized to identify the rice, so that the method has certain advantages in the timeliness of rice identification.
2. A rice probability index TRPI based on time sequence characteristics is established, and three characteristic indexes f (VH) Max )、f(VH Min )、f(VH Mean ) The rice planting possibility under different time sequence characteristics is respectively reflected, and finally the integration can be used for early identification of rice.
Drawings
FIG. 1 is a flow chart of the method of the present invention;
FIG. 2 is a timing maximum distribution diagram according to an embodiment of the present invention;
FIG. 3 is a time series minimum distribution graph according to an embodiment of the present invention;
FIG. 4 is a time-series mean distribution diagram according to an embodiment of the present invention;
fig. 5 is a graph showing a result of TRPI calculation provided in an embodiment of the present invention;
FIG. 6 is a distribution diagram of rice extraction results provided in the embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention is provided to facilitate the understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and it will be apparent to those skilled in the art that various changes may be made without departing from the spirit and scope of the invention as defined and defined by the appended claims, and all changes that can be made by the invention using the inventive concept are intended to be protected.
Examples
According to the embodiment of the invention, a partial region of a Dafeng region of a Yangtze city in Jiangsu province is selected as a research region, a Sentinel No. 1 (Sentinel-1) image is a data source for early rice identification, and a Sentinel No. 2 (Sentinel-2) image is a data source for object-oriented segmentation and NDVI calculation. Referring to fig. 1, a rice early remote sensing identification method based on planting probability is provided, which comprises the following steps:
step 1: video time window selection
In the Yancheng Dafeng area, rice transplanting or direct seeding is generally carried out from the first ten days of 6 months, most of the rice transplanting is completed before the end of 6 months, and the majority of the rice transplanting is completed by later farmers at the beginning of 7 months. Selecting 6-stage Sentinel-1GRD image data (SAR data) from 11 days 6 months to 12 days 7 months in 2021 for early rice identification, wherein the specific image time is 11 days 6 months, 18 days 6 months, 23 days 6 months, 29 days 6 months, 5 days 7 months and 12 months. And preprocessing the images such as orbit correction, thermal noise removal, radiation calibration, filtering, terrain correction, decibel processing, registration, cutting and the like to obtain backscattering coefficients of the SAR images in each period of the target area in a VH polarization mode.
Step 2: image segmentation to obtain land object
And performing multi-scale segmentation by using the No. 2 image of the sentinel in 5 and 8 days of 2021 to obtain a parcel object, wherein the segmentation scale is set to be 80, and the shape factor and the complexity are respectively set to be 0.6 and 0.4.
And 3, step 3: object-level temporal feature parameter extraction
Calculating object-level backscattering coefficients of each period image, and respectively extracting the maximum Value (VH) of the backscattering coefficients of the block objects among the time sequence images Max ) Minimum Value (VH) Min ) Mean Value (VH) Mean ) And obtaining a maximum value graph, a minimum value graph and a mean value graph, as shown in figures 2-4.
And 4, step 4: differentiation and marking of vegetation and non-vegetation areas
NDVI is calculated by using the ratio of the difference and the sum between the near infrared band and the red light band of the sentinel 2 image, and the condition NDVI is set to be more than or equal to 0.55, and is marked as 1 when the condition is true and is marked as 0 when the condition is false. Values were taken under different conditions using f (NDVI). Namely:
and 5: calculating the rice planting probability under each time sequence characteristic
F (VH) of the research region is respectively calculated according to the extraction result of the step 3 Max )、f(VH Max )、f(VH Mean ) The calculation formula is as follows:
in the above formula, the first and second carbon atoms are,respectively-18.1448, -23.8121, -21.0849; (VH) Max ) max 、(VH Max ) min 8.8171, -26.3483; (VH) Min ) max 、(VH Min ) min Respectively-10.8363, -55.0572; (VH) Mean ) max 、(VH Mean ) mix Are-6.2737, -32.3642 respectively. />
Step 6: calculating rice probability finger TRPI based on time sequence characteristics
Calculating TRPI according to the calculation results of the step 4 and the step 5. The calculation formula is as follows, and the TRPI calculation result is shown in fig. 5.
TRPI=f()×f( Max )×f(VH Min )×f(VH Mean )
And 7: threshold classification, extracting rice planting distribution
The condition TRPI is set to be more than or equal to 0.8, rice is set when the condition is true, and non-rice is set when the condition is false. The early rice extraction results are shown in FIG. 6.
The method takes the characteristics of rice growth change and paddy field surface water body appearance condition change in the early stage (tillering stage) of rice planting and the early stage of rice planting as breakthrough, utilizes Synthetic Aperture Radar (SAR) data which is not influenced by weather to identify the rice, and has certain advantages in the timeliness of rice identification. A rice probability index TRPI based on time sequence characteristics is established, and three characteristic indexes f (VH) Max )、f(VH Min )、f(VH Mean ) The rice planting possibility under different time sequence characteristics is respectively reflected, and finally the integration can be used for early identification of rice.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.
Furthermore, it should be understood that although the present specification describes embodiments, not every embodiment includes only a single embodiment, and such description is for clarity purposes only, and it is to be understood that all embodiments may be combined as appropriate by one of ordinary skill in the art to form other embodiments as will be apparent to those of skill in the art from the description herein.
Claims (5)
1. A rice early remote sensing identification method based on planting probability is characterized by comprising the following steps:
step 1: selecting a video time window
Selecting rice to be transplanted to a tillering stage as a time window for identifying early rice, selecting time sequence SARSentinel-1 image data in the time window as a data source for early rice identification, preprocessing a time sequence SAR image, and acquiring a backscattering coefficient of the target area in a VH polarization mode of the SAR image at each stage;
step 2: image segmentation to obtain land object
Carrying out multi-scale segmentation by utilizing a Sentinel-2 multispectral image of rice previous crop in the vigorous growth period to obtain a land parcel object, and setting segmentation scales and various factors according to the image resolution and the area size;
and step 3: extracting object-level temporal feature parameters
Calculating the average value of backscattering coefficients of all pixels in each block object of each phase SAR VH polarization mode image to be used as the backscattering coefficient of the block object, and then extracting the maximum value VH of the same block object among time sequence images Max Minimum value VH Min Mean value VH Mean Finally, generating a time sequence maximum value graph, a time sequence minimum value graph and a time sequence average value graph;
and 4, step 4: differentiating and marking vegetation and non-vegetation areas
Calculating a normalized vegetation index NDVI by using the multispectral image, setting a threshold value for distinguishing and marking vegetation and non-vegetation areas;
and 5: calculating the rice planting probability under different time sequence characteristics
Respectively calculating the rice planting probability of each block object of the time sequence SAR data under the characteristics of the time sequence maximum value, the time sequence minimum value and the time sequence mean value;
step 6: calculating rice probability index TRPI based on time sequence characteristics
Constructing a rice probability index TRPI based on time sequence characteristics, and calculating the early rice planting probability by using the index;
and 7: threshold classification, extracting rice planting distribution
Setting condition TRPI ≥ T 1 And when the condition is true, the rice is the rice, otherwise, the rice is non-rice, and early rice identification is realized.
2. The rice early remote sensing identification method based on planting probability as claimed in claim 1, characterized in that in step 1, the sequential SAR image is subjected to orbit correction, thermal noise removal, radiometric calibration, filtering, terrain correction and decibel processing, registration and clipping preprocessing, and the backscattering coefficient of the target area in the VH polarization mode of the SAR image at each stage is obtained.
3. The early rice remote sensing identification method based on planting probability as claimed in claim 1, wherein the method of step 4 specifically comprises the following substeps:
step 4-1, according to a formula:
obtaining NDVI; where ρ is Red 、ρ NIR Reflectance for red and near infrared bands, respectively;
step 4-2, according to the function:
setting condition NDVI is not less than T 0 When the condition is true, indicating a vegetation area, namely an area where rice can be planted, and marking as 1; when the condition is false, the non-vegetation area is shown, namely the obvious non-planting rice area, the rice planting probability is 0, and the mark is 0.
4. The planting probability-based early remote sensing identification method for rice as claimed in claim 1, wherein in step 5, according to a formula:
obtaining the probability f (VH) that the object to be calculated is rice under different time sequence characteristic indexes i ) (ii) a Wherein the content of the first and second substances,is the mean, x, of the rice sample under the current time series characteristics i Is the backscattering coefficient value of the object to be calculated in the current time series characteristic index diagram, (VH) i ) max 、(VH i ) min Respectively obtaining the maximum value and the minimum value after the non-vegetation areas in the current time sequence characteristic index map are removed; when Max, min and Mean are taken as i, according to the formula:
acquiring the rice planting probability f (VH) under the characteristic indexes of the maximum value, the minimum value and the mean value of the time sequence Max )、f(VH Min )、f(VH Mean )。
5. The planting probability-based early remote sensing identification method for rice as claimed in claim 1, wherein in step 6, according to the formula:
TRPI=f(NDVI)×f(VH Max )×f(VH Min )×f(VH Mean )
calculating a rice probability index TRPI; wherein f (NDVI) is the vegetation and non-vegetation area marked in step 4, and the values are 1 and 0,f (VH) Max )、f(VH Min )、f(VH Mean ) And 5, respectively calculating the rice planting probability under the time sequence maximum value, minimum value and mean value characteristics calculated in the step 5.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211690242.9A CN115861844A (en) | 2022-12-27 | 2022-12-27 | Rice early-stage remote sensing identification method based on planting probability |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211690242.9A CN115861844A (en) | 2022-12-27 | 2022-12-27 | Rice early-stage remote sensing identification method based on planting probability |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115861844A true CN115861844A (en) | 2023-03-28 |
Family
ID=85653549
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211690242.9A Pending CN115861844A (en) | 2022-12-27 | 2022-12-27 | Rice early-stage remote sensing identification method based on planting probability |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115861844A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117690024A (en) * | 2023-12-18 | 2024-03-12 | 宁波大学 | Integrated remote sensing identification method for paddy fields with multiple planting modes |
-
2022
- 2022-12-27 CN CN202211690242.9A patent/CN115861844A/en active Pending
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117690024A (en) * | 2023-12-18 | 2024-03-12 | 宁波大学 | Integrated remote sensing identification method for paddy fields with multiple planting modes |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Kontgis et al. | Mapping rice paddy extent and intensification in the Vietnamese Mekong River Delta with dense time stacks of Landsat data | |
CN111666815B (en) | Automatic garlic planting information extraction method based on Sentinel-2 remote sensing image | |
CN113009485A (en) | Remote sensing tobacco field identification method based on improved vegetation index | |
CN106372592B (en) | A kind of winter wheat planting area calculation method based on winter wheat area index | |
CN111345214A (en) | Xinjiang cotton region identification method and system based on satellite image data | |
CN112711989B (en) | Corn straw coverage estimation method based on radar remote sensing and optical remote sensing | |
CN114821362B (en) | Multi-source data-based rice planting area extraction method | |
CN110110595A (en) | A kind of farmland portrait and medicine hypertrophy data analysing method based on satellite remote-sensing image | |
CN109543552B (en) | Rice growth area mapping method and system based on multi-feature extraction in growth period | |
CN108519339B (en) | WT-L SSVR-based leaf cadmium content Vis-NIR spectral feature modeling method | |
CN114387516B (en) | Single-season rice SAR (synthetic aperture radar) identification method for small and medium-sized fields in complex terrain environment | |
CN112861810B (en) | Artificial forest planting time automatic detection method based on time sequence remote sensing observation data | |
Pang et al. | Pixel-level rice planting information monitoring in Fujin City based on time-series SAR imagery | |
CN116091938A (en) | Multisource remote sensing monitoring method for single-cropping rice planting area | |
CN115861844A (en) | Rice early-stage remote sensing identification method based on planting probability | |
CN115861629A (en) | High-resolution farmland image extraction method | |
CN116518935A (en) | Rice planting distribution and planting intensity recognition method, device and equipment | |
CN114266962A (en) | Crop identification method based on multi-source remote sensing data time sequence and textural features | |
CN117274798B (en) | Remote sensing rice identification method based on regularized time sequence variation model | |
CN111950361A (en) | Beet identification method based on single-time-sequence NDVI | |
CN116524225A (en) | Crop classification method and system based on multi-source remote sensing data | |
Chakraborty et al. | Early almond yield forecasting by bloom mapping using aerial imagery and deep learning | |
CN114972993A (en) | Tectorial membrane farmland identification and extraction method based on remote sensing image | |
CN115165773A (en) | Rice area extraction method based on Google Earth Engine | |
WO2023018387A1 (en) | A crop classification method using deep neural networks |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |