CN111007013B - Crop rotation fallow remote sensing monitoring method and device for northeast cold region - Google Patents

Crop rotation fallow remote sensing monitoring method and device for northeast cold region Download PDF

Info

Publication number
CN111007013B
CN111007013B CN201911059350.4A CN201911059350A CN111007013B CN 111007013 B CN111007013 B CN 111007013B CN 201911059350 A CN201911059350 A CN 201911059350A CN 111007013 B CN111007013 B CN 111007013B
Authority
CN
China
Prior art keywords
remote sensing
ndvi
image
pixel
condition
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.)
Active
Application number
CN201911059350.4A
Other languages
Chinese (zh)
Other versions
CN111007013A (en
Inventor
徐飞飞
陆洲
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhongke Hexin Remote Sensing Technology Suzhou Co ltd
Institute of Geographic Sciences and Natural Resources of CAS
Original Assignee
Zhongke Hexin Remote Sensing Technology Suzhou Co ltd
Institute of Geographic Sciences and Natural Resources of CAS
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Zhongke Hexin Remote Sensing Technology Suzhou Co ltd, Institute of Geographic Sciences and Natural Resources of CAS filed Critical Zhongke Hexin Remote Sensing Technology Suzhou Co ltd
Priority to CN201911059350.4A priority Critical patent/CN111007013B/en
Publication of CN111007013A publication Critical patent/CN111007013A/en
Application granted granted Critical
Publication of CN111007013B publication Critical patent/CN111007013B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
    • G01N21/55—Specular reflectivity
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/10—Terrestrial scenes
    • G06V20/188—Vegetation
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
    • G01N2021/1793—Remote sensing

Abstract

The application relates to a crop rotation fallow remote sensing monitoring method and device for a northeast cold cool area. The crop rotation and fallow identification method has the advantages that the identification accuracy is high, the crops can be accurately identified, the crop types in two years can be compared, and whether crop rotation and fallow are conducted or not can be effectively distinguished.

Description

Crop rotation fallow remote sensing monitoring method and device for northeast cold region
Technical Field
The invention relates to the field of crop remote sensing monitoring, in particular to a crop rotation fallow remote sensing monitoring method and device for northeast cold areas.
Background
The adjustment of the crop planting structure is a national policy of sustainable development of agricultural production in China, and crop rotation fallow is a strategic layout for implementing grain storage in the ground and grain storage in the technology, so that the method has profound significance for guaranteeing grain safety and agricultural green transformation in China. Crop rotation (crop rotation) refers to a mode of planting different crops alternately or in a multiple mode in sequence between different years on the same field. Fallow (fallow) refers to a mode in which only no crop or no crop is cultivated in a cultivation area in a season in which crops can be planted. The crop rotation fallow area is divided into 4 areas such as a northeast cold and cool area, a groundwater funnel area, a heavy metal pollution area, an ecological serious degradation area and the like. At present, the crop planting area in the northeast cold region is acquired based on statistical data, the time efficiency and the authenticity of the crop are poor, and the spatial differentiation characteristic cannot be presented. The crop rotation monitoring is carried out by utilizing the satellite remote sensing technology, the method has the obvious advantages of low cost, convenience, high efficiency, systematicness and the like, the interference of human factors is reduced, the accurate monitoring of the crop area is realized, and the fairness and the justice are realized. Compared with the countries which have achieved better results on crop rotation fallow, china is still in the primary exploration stage, at present, researches are mainly conducted on the aspects of comparison of different crop rotation multiple cropping modes, influence of crop rotation on soil environment and the like, and the research on the remote sensing inspection aspect of crop rotation is less. In addition, the extraction of main crops in the northeast area in a large range mainly has medium and low resolution, cannot meet the requirement of refined crop extraction, and is difficult to realize accurate alternate fallow and fallow area estimation.
The crop rotation mode of the northeast cold region and the fallow mode coexist, the crop rotation is complex and various, the crop rotation mode mainly adopts 'one main crop and multiple auxiliary crops', namely, the corn and soybean wheel are used as main crops, the corn and potatoes, the kernel corn and silage corn, alfalfa, ryegrass, millet, peanut, sunflower, and other oil crop sorghum are used as auxiliary crops. The fallow mode is deep ploughing and planting green manure, planting alfalfa or rape and other field-fertilizing crops. The rotation and fallow test period is three years, rotation and fallow between the annual lines are implemented, the same plot can be dynamically adjusted, and rotation and fallow monitoring of the northeast cold area is more complex. Therefore, in order to meet the requirement of crop rotation remote sensing inspection business in the northeast cold area, a new remote sensing monitoring method is urgently needed.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: in order to solve the defects in the prior art, the remote sensing monitoring method and the remote sensing monitoring device for crop rotation fallow facing northeast cold regions are provided.
The technical scheme adopted by the invention for solving the technical problems is as follows:
a crop rotation fallow remote sensing monitoring method for northeast cold areas comprises the following steps:
s1: acquiring remote sensing images of the monitoring area of the Nth year and the Nth-1 st year; the remote sensing image comprises blue light, green light, red light, near infrared, 3 red side wave bands and short wave infrared wave bands, and the time of the remote sensing image comprises images of 5-9 months;
s2: respectively identifying the crop type of each pixel on the remote sensing images of the Nth year and the Nth-1 th year or whether the pixel is fallow land, and extracting land blocks only planting single crops according to the same crop type;
s3: if the identification results of the same plot on the remote sensing images of the Nth year and the Nth-1 th year are respectively two different crops, the rotation is performed;
and if the identification results of the same plot on the remote sensing images of the Nth year and the Nth-1 st year are respectively fallow land and crops, the fallow land is explained.
Preferably, in the remote sensing monitoring method for crop rotation fallow facing northeast cold areas, in the step S2,
the corn is identified when the pixel on the remote sensing image meets the condition 1, and the condition 1 is as follows: t is a unit of 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 ＞T4；
When the pixels on the remote sensing image meet the condition 2, rice is identified, and the condition 2: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 Less than or equal to T4, and NDRI 08 ≤T 5 ；
When the pixels on the remote sensing image meet the condition 3, the pixels are identified as beet, and the condition 3: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 ≤T 4 And NDRI 08 ＞T 5 And GI 08 ≤T 6 ；
When the pixel on the remote sensing image meets the condition 4, the image is identified as soybean, and the condition 4: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 Less than or equal to T4, and NDVI 08 ＞T 5 And GI 08 ＞T 6 ；
And when the pixel identification on the remote sensing image accords with the condition 5, identifying the image as a fallow land, wherein the condition 5: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＜T 7 And NDBI 08 ＞T 8 ；
Wherein NDVI 05 NDVI representing 5-month image;
NDVI 08 represents the 8-month image NDVI value;
REP 08 a red edge position index representing an 8-month image;
NDRI 08 normalized rice index NDRI representing 8-month images;
GI 08 green index representing 8-month image;
NDBI 08 a normalized architectural index representing an image in month 8;
threshold value T 1 -T 8 All are obtained by sample data statistics.
Preferably, the remote sensing monitoring method for rotation tillage and fallow facing northeast cold areas of the invention,
Figure GDA0003804873920000031
where ρ is red 、ρ nir 、ρ swir Respectively the reflectance values of red light wave band, near infrared wave band and short wave infrared wave band pixels;
Figure GDA0003804873920000032
where ρ is the inflection point reflectivity, ρ = (ρ) 670 +ρ 780 )/2，ρ 670 、ρ 700 、ρ 740 、ρ 780 Reflectance at wavelengths of 670nm,700nm,740nm and 780nm, respectively;
Figure GDA0003804873920000033
GI＝10000*(ρ blue -ρ green )；
where ρ is red 、ρ swir 、ρ blue 、ρ green Respectively the reflectance values of red light wave band, short wave infrared, blue light wave band and green light slope band.
Preferably, in the remote sensing monitoring method for crop rotation fallow facing northeast cold areas, in the step S2, the extraction of the plots comprises the following steps:
s201: identifying a land parcel on the remote sensing image;
s202: identifying crops in the plots, and marking the plots as plots to be corrected if one plot contains at least two types of crops;
s203: smoothing the remote sensing image containing the plot to be corrected by using a Gaussian filter;
s204: calculating the amplitude and direction of the image gradient after filtering by using the finite difference of a first-order differential operator;
s205: carrying out non-maximum value suppression on the gradient amplitude, traversing the image, if the gradient value of a certain pixel is smaller than the gradient values of two adjacent pixels along the gradient direction, judging that the pixel point is not an edge point, and setting the gray value of the pixel point as 0;
s206: detecting and connecting edges by using a dual-threshold algorithm, calculating two thresholds which are respectively a high threshold and a low threshold by using an accumulative histogram, and respectively recording T high And T low Where is greater than T high Is judged as an edge, if less than T low Is judged not to be an edge; if the detection result is greater than or equal to T low But less than or equal to T high If yes, the pixel point is an edge point, otherwise, the pixel point is not an edge point;
s207: if the pixel points serving as the edge points in the detection result are discontinuous to form breakpoints, searching end points by using an n multiplied by n template according to the invariance of the edge local direction; connecting any two searched end points, evaluating whether the connection line of the end points is linear, if the result is linear, using the connection line as an edge, and if the end points cannot be searched or are not linear, manually marking;
s208: and cutting the extracted linear boundary line into the contour boundary of the farmland block to finally form a farmland block unit, thereby obtaining the block unit with only one crop.
Preferably, in the remote sensing monitoring method for crop rotation fallow facing northeast cold areas, statistics is further carried out in the step S3 according to different change conditions of the identification results of the same plot on the remote sensing images of the Nth year and the Nth-1 th year, and crop rotation modes, crop rotation areas and fallow areas are analyzed.
The invention also provides a rotation tillage remote sensing monitoring device facing northeast cold and cool areas, which comprises:
an image acquisition module: the remote sensing image acquisition system is used for acquiring remote sensing images of monitoring areas in the Nth year and the Nth-1 st year; the remote sensing image comprises blue light, green light, red light, near infrared, 3 red side wave bands and short wave infrared wave bands, and the time of the remote sensing image comprises images of 5-9 months;
the land parcel type identification module: the remote sensing image processing system is used for respectively identifying the crop type of each pixel on the remote sensing images of the Nth year and the Nth-1 th year or whether the pixel is fallow land or not, and extracting land blocks only used for planting single crops according to the same crop type;
and a result judgment module: the remote sensing image is used for judging whether the plot is crop rotation or fallow, and if the identification results of the same plot on the remote sensing images of the Nth year and the Nth-1 th year are respectively two different crops, the crop rotation is performed; and if the identification results of the same plot on the remote sensing images of the Nth year and the Nth-1 st year are respectively fallow land and crops, the fallow land is explained.
Preferably, the remote sensing monitoring device for crop rotation fallow facing northeast cold areas of the invention is arranged in a plot type identification module,
the corn is identified when the pixel on the remote sensing image meets the condition 1, and the condition 1 is as follows: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 ＞T4；
When the pixels on the remote sensing image meet the condition 2, rice is identified, and the condition 2: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 ≤T 4 And NDRI 08 ≤T 5 ；
When the pixels on the remote sensing image meet the condition 3, the pixels are identified as beet, and the condition 3: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 T4 or less, and NDRI 08 ＞T 5 And GI 08 ≤T 6 ；
When the pixel on the remote sensing image meets the condition 4, the image is identified as soybean, and the condition 4: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 Less than or equal to T4, and NDVI 08 ＞T 5 And GI 08 ＞T 6 ；
And identifying the fallow land when the pixel identification on the remote sensing image meets the condition 5, wherein the condition 5: t is a unit of 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＜T 7 And NDBI 08 ＞T 8 ；
Wherein NDVI 05 NDVI representing 5-month image;
NDVI 08 represents the 8-month image NDVI value;
REP 08 a red edge position index representing an image of 8 months;
NDRI 08 normalized rice index NDRI representing 8-month images;
GI 08 green index representing 8-month image;
NDBI 08 a normalized architectural index representing an image for 8 months;
threshold value T 1 -T 8 All are obtained by sample data statistics.
Preferably, the remote sensing monitoring device for crop rotation fallow facing northeast cold areas of the invention,
Figure GDA0003804873920000061
where ρ is red 、ρ nir 、ρ swir Respectively the reflectance values of the red wave band pixel, the near infrared pixel and the short wave infrared pixel;
Figure GDA0003804873920000062
where ρ is the inflection point reflectivity, ρ = (ρ) 670 +ρ 780 )/2，ρ 670 、ρ 700 、ρ 740 、ρ 780 Reflectance at wavelengths of 670nm,700nm,740nm and 780nm, respectively;
Figure GDA0003804873920000063
GI＝10000*(ρ blue -ρ green )；
where ρ is red 、ρ swir 、ρ blue 、ρ green Respectively the reflectance values of red light wave band, short wave infrared, blue light wave band and green light slope band.
Preferably, the remote sensing monitoring device for crop rotation fallow facing northeast cold areas of the invention comprises the following steps of extracting plots in the plot type identification module:
s201: identifying a land parcel on the remote sensing image;
s202: identifying crops in the land parcels, and marking the land parcels to be corrected if one land parcel contains at least two types of crops;
s203: smoothing the remote sensing image containing the plot to be corrected by using a Gaussian filter;
s204: calculating the amplitude and direction of the image gradient after filtering by using the finite difference of a first-order differential operator;
s205: carrying out non-maximum value suppression on the gradient amplitude, traversing the image, if the gradient value of a certain pixel is smaller than the gradient values of two adjacent pixels along the gradient direction, judging that the pixel point is not an edge point, and setting the gray value of the pixel point as 0;
s206: detecting and connecting edges with a dual threshold algorithm, calculating two thresholds, high and low, respectively, using cumulative histograms, and separately noting T high And T low Where is greater than T high Is judged as an edge, if less than T low Is judged not to be an edge; if the detection result is greater than or equal to T low But less than or equal to T high If yes, the pixel point is the edge point, otherwise, the pixel point is not the edge pointOr not an edge;
s207: if the pixel point as the edge point in the detection result is discontinuous to form a breakpoint, searching an end point by an n multiplied by n template according to the invariance of the edge local direction; connecting any two searched end points, evaluating whether the connection line of the end points is linear, if the result is linear, using the connection line as an edge, and if the end points cannot be searched or are not linear, manually marking;
s208: and cutting the extracted linear boundary line into a cultivated land block contour boundary to finally form a cultivated land block unit, thereby obtaining the block unit with only one crop.
Preferably, the remote sensing monitoring device for crop rotation fallow facing the northeast cold region is characterized in that the result judgment module is used for carrying out statistics according to different change conditions of the identification results of the same plot on the remote sensing images of the Nth year and the Nth-1 th year, and analyzing a crop rotation mode, a crop rotation area and a fallow area.
The invention has the beneficial effects that:
the application provides a crop rotation fallow remote sensing monitoring method and device for a northeast cold area. The crop rotation recognition method has the advantages that the recognition accuracy is high, the crops can be recognized accurately, and finally, the comparison is carried out according to the crop types of two years, so that whether rotation fallow is carried out or not can be effectively distinguished.
Drawings
The technical solution of the present application is further explained below with reference to the drawings and the embodiments.
FIG. 1 is a general technical flow diagram of crop rotation fallow;
FIG. 2 is a flow chart of a cultivated land plot extraction and rotation fallow mode monitoring technique;
FIG. 3 is a diagram of the recognition result of cultivated land parcel;
FIG. 4 is a multi-temporal, multi-exponential based crop extraction technique
FIG. 5 shows the results of the crop extraction in this and the previous year;
FIG. 6 is a schematic view of a cultivated land parcel correction;
FIG. 7 is a view showing the results of crop rotation cultivation monitoring;
Detailed Description
It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict.
The technical solutions of the present application will be described in detail below with reference to the accompanying drawings in conjunction with embodiments.
Example 1
The embodiment provides a crop rotation fallow remote sensing monitoring method facing northeast cold areas, as shown in fig. 1, comprising the following steps:
s1: acquiring remote sensing images of the monitoring area of the Nth year and the Nth-1 th year;
the remote sensing image comprises blue light, green light, red light, near infrared, 3 red-side wave bands and short-wave infrared wave bands, such as a sentinel No. 2 image, and the remote sensing image is subjected to wave band synthesis, atmospheric correction, geometric correction, cutting, inlaying and other processing, comprises images of 5-9 months and covers the whole growth period of crops such as soybean, rice, corn and the like in a research area;
s2: respectively identifying the crop type of each pixel on the remote sensing images of the Nth year and the Nth-1 st year or whether the pixel is fallow land, extracting land blocks only planting single crops according to the same crop type, and extracting the land blocks only planting the single crops according to the same crop type;
the corn is identified when the pixel on the remote sensing image meets the condition 1, and the condition 1 is as follows: t is a unit of 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 ＞T 4 ；
And when the pixels on the remote sensing images meet the condition 2, rice is identified, and the condition 2: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 ≤T 4 And NDRI 08 ≤T 5 ；
When the pixels on the remote sensing image meet the condition 3, the pixels are identified as beet, and the condition 3: t is a unit of 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 ≤T 4 And NDRI 08 ＞T 5 And GI 08 ≤T 6 ；
When the pixel on the remote sensing image meets the condition 4, the image is identified as soybean, and the condition 4: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 ≤T 4 And NDVI 08 ＞T 5 And GI 08 ＞T 6 ；
And identifying the fallow land when the pixel identification on the remote sensing image meets the condition 5, wherein the condition 5: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＜T 7 And NDBI 08 ＞T 8 ；
Wherein NDVI 05 NDVI, threshold T, representing 5-month image 1 And T 2 Statistically derived from sample data, e.g. T 1 ＝0，T 2 ＝0.5；
NDVI 08 Indicating the NDVI value of the image in 8 months, threshold T 3 And T 7 Statistically derived from sample data, e.g. T 3 ＝0.6，T 7 ＝0.2；
REP 08 Red edge position index, threshold T, representing 8-month image 4 Derived from sample data statistics, e.g. T 4 ＝728；
NDRI 08 Normalized rice index NDRI representing 8-month image, threshold T5 is obtained from statistics of sample point data, such as T 5 ＝0.5；
GI 08 Greenness index, threshold T, representing 8-month image 6 From the statistics of the sample point data, T 6 ＝52；
NDBI 08 Normalized building index, threshold T, representing 8-month image 8 Derived from statistics of sample data, e.g. T 8 ＝-0.1，
The sample data statistical threshold is a threshold for creating random points on the image, visually distinguishing the types of the random points, respectively extracting the reflectivity of each wave band and calculating the classification of crops.
The specific identification process is as follows:
s21: removing water and forest land by using NDVI of the 5-month image of the N-1 year or the N year, and judging the NDVI 05 Whether or not the value range of (1) meets T 1 ＜NDVI 05 ＜T 2 If the condition is not, the pixel is marked as a water body or a forest land, and if the condition is true, S22 judgment is carried out;
s22: then, 8-month image NDVI is used for eliminating buildings or other artificial ground objects. Judging pixel NDVI 08 Whether or not greater than T 3 If not, the picture element is marked as a building or other artificial ground object, and if so, S23 judgment is carried out;
can further judge the NDVI 08 Whether or not less than T 7 Wherein NDVI 08 Indicating the NDVI value of the image in 8 months, threshold T 7 Statistically derived from sample data, e.g. T 7 =0.2 if NDVI 08 ＞T 7 The picture element is identified as dense vegetation if NDVI 08 ＜T 7 Then setting the condition NDBI 08 ＞T 8 Wherein NDBI 08 Normalized building index, threshold T, representing 8-month image 8 Derived from statistics of sample data, e.g. T 8 And = 0.1, if the condition is that the pixel is not the building, the pixel is identified as the fallow land, and if the condition is true, the pixel is identified as the fallow land.
S23: the red edge position index of the 8-month corn is larger than that of crops such as soybean, rice and the like, and the REP is favorable for extracting the corn. Judging REP of picture element 08 Whether or not T is less than or equal to 4 If the condition is not, the pixel is marked as corn, and if the condition is true, S4 judgment is carried out;
s24: the reflectivity of the short-wave infrared spectrum rises along with the reduction of the water content of the soil, and the short-wave infrared spectrum is sensitive to the change of vegetation water. And constructing a normalized rice index NDRI based on the short-wave infrared band and the red band, wherein the normalized rice index NDRI is defined as the ratio of the difference between the reflection value of the short-wave infrared band and the reflection value of the red band to the sum of the reflection values. Judging the NDRI of the pixel after step S3 08 Whether or not greater than T 5 If the condition is not, the pixel is marked as rice, and if the condition is true, S25 judgment is carried out;
s25: the color difference is used for distinguishing soybean beet, and the soybean beet in 8 months shows twoThe color of the soybean is dark green, and the color of the beet is bright green. The greenness index GI is constructed based on the blue and green wavelength bands and is defined as the difference between the reflectance values of the blue and green wavelength bands. Determining GI of Pixel after step S24 08 ＞T 6 If the condition is not, the pixel is identified as beet, and if the condition is true, the pixel is identified as soybean;
in this step, the various indices are defined as follows:
normalized vegetation index (NDVI) to distinguish between vegetation and non-vegetation; normalized building index (NDBI) to differentiate between buildings and fallow land, red edge position index (REP) to extract corn; the normalized rice index NDRI is used for extracting rice; greenness index GI is used to distinguish between soybean and sugar beet.
NDVI is defined as the ratio of the difference of the reflectivity of the near infrared band and the reflectivity of the red band to the sum of the reflectivities, NDBI is defined as the ratio of the reflectivity of the short wave infrared band to the difference of the reflectivity of the near infrared band to the sum of the reflectivities, and the calculation formula is defined as follows:
Figure GDA0003804873920000111
Figure GDA0003804873920000112
where ρ is red 、ρ nir 、ρ swir The reflectivity values of the red wave band pixel, the near infrared pixel and the short wave infrared pixel are respectively.
Red edge position index (REP) with the formula
ρ＝(ρ 670 +ρ 780 )/2
Figure GDA0003804873920000121
Where ρ is the inflection point reflectivity, ρ 670 、ρ 700 、ρ 740 、ρ 780 Respectively, the reflectance at wavelengths of 670nm,700nm,740nm and 780nm. 700 and 740 are 70Constant generated by interpolation in the interval of 0 nm-740 nm.
In addition to the above commonly used vegetation indexes, the present invention defines two new vegetation indexes as normalized rice index NDRI and greenness index GI, respectively. The NDRI is defined as the ratio of the difference between the reflection value of the short wave infrared band and the reflection value of the red wave band to the sum of the two, and is calculated as follows:
Figure GDA0003804873920000122
greenness index, GI, calculated as follows:
GI＝10000*(ρ blue -ρ green )
where ρ is red 、ρ swir 、ρ blue 、ρ green Respectively the reflectance values of red light wave band, short wave infrared, blue light wave band and green light slope band.
Crop rotation and fallow are monitored on the basis of a plot, and the plot is a plot edge vector result obtained based on deep learning. The cultivated land plot extraction result does not consider the ground planting condition, various crop plants exist on the obtained plot, and the abnormal cultivated land plot needs to be corrected in order to meet the requirement of accurate monitoring of crop rotation and fallow and further improve the accuracy of the plot extraction result. Abnormal fields typically represent fields containing at least two types of crops, with the crops having distinct boundaries due to color differences, the boundaries being of small width and generally being visible as straight lines. The correction of the abnormal farmland plots is to detect the straight line and cut the farmland plots based on the straight line. The cut field section contains only one crop.
In the embodiment, soybeans, rice and corns are mainly planted in the northeast cold area, the red edge position of the corn in 8 months is the largest, and the threshold value of the red edge position is set at the moment, so that the corn is convenient to identify;
most of rice extraction in the prior art is based on NDVI or NDWI and is realized by utilizing the difference of NDVI change between the transplanting period and the heading period of rice. The method needs multi-scene image participation and increases the calculation amount. In addition, this method often misclassifies soybeans to rice. The short wave infrared has obvious reflection spectrum characteristics on vegetation moisture and mineral type identification. The false color image composed of 8-month images in near-infrared, short-wave infrared and red wave bands has obvious characteristics, and is represented as that rice is rose red, and soybeans and corns are yellow. The NDRI is the ratio of the sum of the reflectivity of the short-wave infrared and the red wave band to the difference of the reflectivity, the characteristic that the short-wave infrared is sensitive to the soil water content is fully utilized, and the index is more beneficial to extracting rice. The representation forms of the soybeans and the beet are different on remote sensing images, and the green degree index is constructed based on the difference between the reflection values of blue light and green light wave bands and is used for distinguishing the soybeans from the beet.
The decision tree model constructed by combining the NDVI, the REP and the NDRI in the embodiment makes full use of NDVI characteristics, red edge characteristics and short wave infrared rice characteristics of crops by using less image data, reduces data volume, and is convenient to extract and accurate in precision.
S3: if the identification results of the same plot on the remote sensing images of the Nth year and the Nth-1 th year are respectively two different crops, the rotation is performed;
and if the identification results of the same plot on the remote sensing images of the Nth year and the Nth-1 st year are respectively fallow land and crops, the fallow land is explained.
Furthermore, the conditions that corns are changed into soybeans, soybeans are changed into corns, corns are changed into beets, corns are changed into rice, rice is changed into fallow land, fallow land is changed into rice and the like can be counted, crop rotation and fallow modes are analyzed according to the change conditions, crop rotation plots, crop rotation types and fallow plots are marked, and crop rotation modes, crop rotation areas and fallow areas are counted and analyzed.
In step S2, the following parcel extraction method may be adopted:
the method comprises the following steps: firstly, fusing a true-color three-channel and a panchromatic waveband of a remote sensing image to obtain an image which is stored in an R-G-B three-channel mode, wherein the resolution is 0.8 m, and the data format is tiff. Manually drawing a contour boundary of a farmland plot on an original image by utilizing ArcGIS software to manufacture label data, creating a labeled image which has the same projection information as the original data and is stored in a single wave band based on vector data of the contour boundary, wherein the image content is 0 and 1, and the labeled image respectively represents a non-farmland (background type) and the farmland plot. And randomly cutting the original image and the marked image at the same time, wherein the cut size is 256 multiplied by 256 pixels, performing data enhancement processing on the cut image, and obtaining 10000 pictures with 256 multiplied by 256 pixels and png format after random cutting and data enhancement, wherein 25% of the pictures are 2500 pictures as verification samples, and 7500 pictures are training samples.
Step two: and carrying out model training by utilizing a U-net network model built by a Tensorflow and Keras deep learning framework. Before training, the learning rate parameter is set to be 0.001, the training batch size is 32, the verification batch size is 16, the training times are 2000, and the activation function is a Sigmod function.
Step three: the trained facility agriculture extraction model is used for predicting the spatial distribution and type of the facility agriculture, images to be classified are cut into image blocks with fixed size (256 multiplied by 256) pixels to be predicted respectively, image blocks with certain overlapped areas are obtained in a sliding window pixel 32 mode, then classification results of a certain middle area are reserved for each predicted image block, inaccurate edge results are abandoned, and the image blocks are spliced in sequence.
And (3) cutting the image to be classified into blocks according to the size of the image and the memory of a computer, cutting the blocked image into small images with the size of 256 multiplied by 256 pixels, inputting the cut data into a trained model for classification, and splicing the classified small images. And performing classification post-processing operations such as mode filtering, corrosion, expansion, opening and closing operation and the like on the classification result. And (3) removing the pixels with less number by mode filtering, and keeping the target with more pixels. Erosion operations erode the target area "smaller" in extent, which essentially causes the boundaries of the field patch in the image to shrink, which can be used to eliminate small and meaningless objects.
Step four: and converting the result of the cultivated land block into a vector. And obtaining raster data as a deep learning result, converting the raster data into vectors through binarization, thinning and the like by utilizing ArcGIS software raster vector conversion operation, and then obtaining the initial cultivated land block boundary through smooth surface operation. The cultivated land plot extraction results are shown in fig. 3. Cultivated land and land mass in research area6196 blocks are counted, and the area is 42.52km in total 2 The minimum land area is 0.45 mu, and the maximum cultivated land area is 455 mu.
In step S2, the following parcel extraction method may be adopted:
and selecting the gray level image or the texture image of the key period image. The gray level image is a single-band image, and the texture image is a mean value, a variance, an entropy value and the like calculated by utilizing a gray level co-occurrence matrix.
S201: identifying a land parcel on the remote sensing image;
s202: identifying crops in the plots, and marking the plots as plots to be corrected if one plot contains at least two types of crops;
s203: smoothing the remote sensing image containing the plot to be corrected by using a Gaussian filter;
s204: calculating the amplitude and the direction of the gradient of the filtered image by using the finite difference of a first-order differential operator;
s205: carrying out non-maximum value suppression on the gradient amplitude, traversing the image, if the gradient value of a certain pixel is smaller than the gradient values of two adjacent pixels along the gradient direction, judging that the pixel point is not an edge point, and setting the gray value of the pixel point as 0;
s206: detecting and connecting edges by using a double-threshold algorithm, calculating two thresholds which are respectively a high threshold and a low threshold by using an accumulative histogram, respectively recording T _ high and T _ low, judging the edges if the thresholds are more than T _ high, and judging the edges if the thresholds are less than T _ low; if the detection result is more than or equal to T _ low but less than or equal to T _ high, continuously judging whether a pixel point which exceeds a high threshold value exists in the adjacent pixel of the pixel, if so, determining that the pixel point is an edge point, otherwise, determining that the pixel point is not an edge;
although Canny can extract partial edge information, by analyzing the result of Canny edge detection, a single-pixel discontinuity phenomenon is found in the middle of the extracted edge. The broken line edge can not reach the boundary of the farmland plots, namely the farmland plots can not be divided.
S207: if the pixel points serving as the edge points in the detection result are discontinuous to form breakpoints, searching end points by using an n multiplied by n (3 multiplied by 3 pixel) template according to the invariance of the edge local direction; connecting any two searched end points, evaluating whether the connection line of the end points is linear, if the result is linear, using the connection line as an edge, and if the end points cannot be searched or are not linear, manually marking; one commonly used linear evaluation method is the aspect ratio index, i.e., the ratio of the length to the width of the smallest circumscribed rectangle of the linear region of interest;
s208: and cutting the extracted linear boundary line into the contour boundary of the farmland block to finally form a farmland block unit, thereby obtaining the block unit with only one crop.
The sentinel No. 2 is only one data containing three wave bands in the red edge range by the application date, the resolution is 10 meters, the monitoring of the crop type is facilitated, and a new remote sensing monitoring data is provided. The high-resolution remote sensing image is rich in texture features, the boundary of the plot can be determined by combining the high-resolution remote sensing image with a deep learning technology, and an accurate plot vector is provided for crop rotation checking. The invention provides a crop rotation fallow monitoring technology in a northeast cold region based on a high-resolution remote sensing image and a sentinel image. Firstly, obtaining a plot farmland boundary by utilizing a deep learning technology based on a high-resolution remote sensing image, secondly, extracting crops such as soybean, rice, corn, wheat, beet and the like based on a sentinel image in combination with crop phenological characteristics, red edge parameter positions, normalized vegetation indexes and normalized rice indexes, correcting the crop plot result to a farmland plot, analyzing the variation condition of crops between the annual rings, and obtaining crop rotation and fallow and area.
Example 2
The embodiment also provides a rotation tillage remote sensing monitoring device facing northeast cold and cool areas, which comprises:
an image acquisition module: the remote sensing image acquisition system is used for acquiring remote sensing images of monitoring areas in the Nth year and the Nth-1 st year; the remote sensing image comprises blue light, green light, red light, near infrared, 3 red side wave bands and short wave infrared wave bands, and the time of the remote sensing image comprises images of 5-9 months;
the land parcel type identification module: the remote sensing image processing system is used for respectively identifying the crop type of each pixel on the remote sensing images of the Nth year and the Nth-1 th year or whether the pixel is fallow land or not, and extracting land blocks only used for planting single crops according to the same crop type;
and a result judgment module: the remote sensing image is used for judging whether the plot is crop rotation or fallow, and if the identification results of the same plot on the remote sensing images of the Nth year and the Nth-1 th year are respectively two different crops, the crop rotation is performed; and if the identification results of the same plot on the remote sensing images of the Nth year and the Nth-1 st year are respectively fallow land and crops, the fallow land is explained.
Further, in the parcel type identification module,
the corn is identified when the pixel on the remote sensing image meets the condition 1, and the condition 1 is as follows: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 ＞T 4 ；
And when the pixels on the remote sensing images meet the condition 2, rice is identified, and the condition 2: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 ≤T 4 And NDRI 08 ≤T 5 ；
When the pixels on the remote sensing image meet the condition 3, the pixels are identified as beet, and the condition 3: t is a unit of 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 Less than or equal to T4, and NDRI 08 ＞T 5 And GI 08 ≤T 6 ；
When the pixels on the remote sensing image meet the condition 4, the pixels are identified as soybeans, and the condition 4 is as follows: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 Less than or equal to T4, and NDVI 08 ＞T 5 And GI 08 ＞T 6 ；
And identifying the fallow land when the pixel identification on the remote sensing image meets the condition 5, wherein the condition 5: t is 1 ＜NDVI 05 ＜T 2 And NDVI 0s ＜T 7 And NDBI 08 ＞T 8 ；
Wherein NDVI 05 NDVI representing 5-month image;
NDVI 08 represents the 8-month image NDVI value;
REP 08 a red edge position index representing an 8-month image;
NDRI 08 normalized rice index NDRI representing 8-month images;
GI 08 green index representing 8-month image;
NDBI 08 a normalized architectural index representing an image for 8 months;
threshold value T 1 -T 8 All are obtained by sample data statistics.
Preferably, the remote sensing monitoring device for crop rotation fallow facing northeast cold areas of the invention,
Figure GDA0003804873920000181
where ρ is red 、ρ nir 、ρ swir Respectively the reflectance values of red light wave band, near infrared wave band and short wave infrared wave band pixels;
Figure GDA0003804873920000182
where ρ is the inflection point reflectivity, ρ = (ρ) 670 +ρ 780 )/2，ρ 670 、ρ 700 、ρ 740 、ρ 780 Reflectance at wavelengths of 670nm,700nm,740nm and 780nm respectively;
Figure GDA0003804873920000183
GI＝10000*(ρ blue -ρ green )；
where ρ is red 、ρ swir 、ρ blue 、ρ green Respectively the reflectance values of red light wave band, short wave infrared, blue light wave band and green light slope band.
Further, in the land parcel type identification module, the extraction of the land parcel comprises the following steps:
s201: identifying a land parcel on the remote sensing image;
s202: identifying crops in the plots, and marking the plots as plots to be corrected if one plot contains at least two types of crops;
s203: smoothing the remote sensing image containing the plot to be corrected by using a Gaussian filter;
s204: calculating the amplitude and direction of the image gradient after filtering by using the finite difference of a first-order differential operator;
s205: carrying out non-maximum value suppression on the gradient amplitude, traversing the image, if the gradient value of a certain pixel is smaller than the gradient values of two adjacent pixels along the gradient direction, judging that the pixel point is not an edge point, and setting the gray value of the pixel point as 0;
s206: detecting and connecting edges by using a dual-threshold algorithm, calculating two thresholds which are respectively a high threshold and a low threshold by using an accumulative histogram, and respectively recording T high And T low Where is greater than T high Is judged as an edge, if less than T low Is judged not to be an edge; if the detection result is greater than or equal to T low But less than or equal to T high If yes, the pixel point is an edge point, otherwise, the pixel point is not an edge point;
s207: if the pixel point as the edge point in the detection result is discontinuous to form a breakpoint, searching an end point by an n multiplied by n template according to the invariance of the edge local direction; connecting any two searched end points, evaluating whether the connection line of the end points is linear, if the result is linear, using the connection line as an edge, and if the end points cannot be searched or are not linear, manually marking; one commonly used linear evaluation method is the aspect ratio index, i.e., the ratio of the length to the width of the smallest circumscribed rectangle of the linear region of interest;
s208: and cutting the extracted linear boundary line into a cultivated land block contour boundary to finally form a cultivated land block unit, thereby obtaining the block unit with only one crop.
Furthermore, the result judgment module is used for carrying out statistics according to different change conditions of the identification results of the same plot on the remote sensing images of the Nth year and the Nth-1 st year, and analyzing the crop rotation mode, the crop rotation area and the fallow area.
In light of the foregoing description of the preferred embodiments according to the present application, it is to be understood that various changes and modifications may be made without departing from the spirit and scope of the invention. The technical scope of the present application is not limited to the contents of the specification, and must be determined according to the scope of the claims.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

Claims (6)

1. A rotation tillage remote sensing monitoring method facing northeast cold areas is characterized by comprising the following steps:
s1: acquiring remote sensing images of the monitoring area of the Nth year and the Nth-1 st year; the remote sensing image comprises blue light, green light, red light, near infrared, 3 red side wave bands and short wave infrared wave bands, and the time of the remote sensing image comprises images of 5-9 months;
s2: respectively identifying the crop type of each pixel on the remote sensing images of the Nth year and the Nth-1 st year or whether the pixel is a fallow land, and extracting land blocks only planting single crops according to the same crop type;
s3: if the identification results of the same plot on the remote sensing images of the Nth year and the Nth-1 th year are respectively two different crops, the crop rotation is performed;
if the identification results of the same plot on the remote sensing images of the Nth year and the Nth-1 th year are fallow land and crops respectively, the fallow land is indicated;
in the step of S2, the step of,
the corn is identified when the pixel on the remote sensing image meets the condition 1, and the condition 1 is as follows: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 ＞T 4 ；
And when the pixels on the remote sensing images meet the condition 2, rice is identified, and the condition 2: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 ≤T 4 And NDRI 08 ≤T 5 ；
Pixel on remote sensing image meets conditionsSugar beet at 3, condition 3: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 ≤T 4 And NDRI 08 ＞T 5 And GI 08 ≤T 6 ；
When the pixel on the remote sensing image meets the condition 4, the image is identified as soybean, and the condition 4: t is a unit of 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 ≤T 4 And NDVI 08 ＞T 5 And GI 08 ＞T 6 ；
And when the pixel identification on the remote sensing image accords with the condition 5, identifying the image as a fallow land, wherein the condition 5: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＜T 7 And NDBI 08 ＞T 8 ；
Wherein NDVI is a normalized vegetation index, REP is a red edge position index, NDBI is a normalized construction index, NDRI is a normalized rice index, GI is a greenness index;
NDVI 05 NDVI representing 5-month image;
NDVI 08 represents the 8-month image NDVI value;
REP 08 a red edge position index representing an 8-month image;
NDRI 08 normalized rice index NDRI representing 8-month images;
GI 08 green index representing 8-month image;
NDBI 08 a normalized architectural index representing an image in month 8;
threshold value T 1 -T 8 All are obtained by sample data statistics; creating random points on an image, visually judging the type of the random points, respectively extracting the reflectivity of each wave band, and calculating the threshold value of crop classification;
Figure FDA0003804873910000021
where ρ is red 、ρ nir 、ρ swir Are respectively red lightWave band, near infrared and short wave infrared band pixel reflectivity values;
Figure FDA0003804873910000022
where ρ is the inflection point reflectivity, ρ = (ρ) 670 +ρ 780 )/2，ρ 670 、ρ 700 、ρ 740 、ρ 780 Reflectance at wavelengths of 670nm,700nm,740nm and 780nm, respectively;
Figure FDA0003804873910000023
GI＝10000*(ρ blue -ρ green )；
where ρ is red 、ρ swir 、ρ blue 、ρ green Reflectance values of a red light wave band, a short wave infrared band, a blue light wave band and a green light wave band respectively;
the specific identification process is as follows:
s21: removing water and forest land by using NDVI of the 5-month image of the N-1 year or the N year, and judging the NDVI 05 Whether or not the value range of (1) meets T 1 ＜NDVI 05 ＜T 2 If the condition is not, the pixel is marked as a water body or a forest land, and if the condition is true, S22 judgment is carried out;
s22: then, removing buildings or other artificial ground objects by using the 8-month image NDVI; judging pixel NDVI 08 Whether or not greater than T 3 If not, the pixel is marked as a building or other artificial ground object, and if so, S23 judgment is carried out;
s23: judging REP of picture element 08 Whether or not T is less than or equal to 4 If the condition is not, the pixel is marked as corn, and if the condition is true, S24 judgment is carried out;
s24: constructing a normalized rice index NDRI based on short wave infrared and red wave bands, wherein the normalized rice index NDRI is defined as the ratio of the difference between the reflection value of the short wave infrared wave band and the reflection value of the red wave band to the sum of the reflection values; judging the NDRI of the pixel after step S23 08 Whether or not it is greater than T 5 If a strip ofIf the condition is not true, the pixel is marked as rice, and if the condition is true, S25 judgment is carried out;
s25: constructing a green index GI based on blue light and green light wave bands, and defining the green index GI as the difference of reflection values of the blue light and the green light wave bands; determining GI of Pixel after step S24 08 ＞T 6 And if the condition is not, the pixel is identified as beet, and if the condition is true, the pixel is identified as soybean.
2. The remote sensing monitoring method for rotation tillage in northeast cold areas according to claim 1, wherein in the step S2, the extraction of the land parcel comprises the following steps:
s201: identifying a land parcel on the remote sensing image;
s202: identifying crops in the plots, and marking the plots as plots to be corrected if one plot contains at least two types of crops;
s203: smoothing the remote sensing image containing the plot to be corrected by using a Gaussian filter;
s204: calculating the amplitude and direction of the image gradient after filtering by using the finite difference of a first-order differential operator;
s205: carrying out non-maximum value suppression on the gradient amplitude, traversing the image, if the gradient value of a certain pixel is smaller than the gradient values of two adjacent pixels along the gradient direction, judging that the pixel point is not an edge point, and setting the gray value of the pixel point as 0;
s206: detecting and connecting edges with a dual threshold algorithm, calculating two thresholds, high and low, respectively, using cumulative histograms, and separately noting T high And T low Where is greater than T high Is judged as an edge, if less than T low Is judged not to be an edge; if the detection result is greater than or equal to T low But less than or equal to T high If yes, the pixel point is an edge point, otherwise, the pixel point is not an edge;
s207: if the pixel point as the edge point in the detection result is discontinuous to form a breakpoint, searching an end point by an n multiplied by n template according to the invariance of the edge local direction; connecting any two searched end points, evaluating whether the connection line of the end points is linear, if the result is linear, using the connection line as an edge, and if the end points cannot be searched or are not linear, manually marking;
s208: and cutting the extracted linear boundary line into the contour boundary of the farmland block to finally form a farmland block unit, thereby obtaining the block unit with only one crop.
3. The remote sensing monitoring method for crop rotation fallow facing to the northeast cold region as claimed in claim 1, wherein in the step S3, statistics is further carried out according to the different change conditions of the recognition results of the same plot on the remote sensing images of the nth year and the nth-1 year, and crop rotation mode, crop rotation area and fallow area are analyzed.
4. The utility model provides a rotation fallow remote sensing monitoring devices towards northeast cold cool district which characterized in that includes:
an image acquisition module: the remote sensing image acquisition system is used for acquiring remote sensing images of monitoring areas in the Nth year and the Nth-1 st year; the remote sensing image comprises blue light, green light, red light, near infrared, 3 red side wave bands and short wave infrared wave bands, and the time of the remote sensing image comprises images of 5-9 months;
the land parcel type identification module: the remote sensing image processing system is used for respectively identifying the crop type of each pixel on the remote sensing images of the Nth year and the Nth-1 th year or whether the pixel is fallow land or not, and extracting land blocks only used for planting single crops according to the same crop type;
and a result judgment module: the remote sensing image is used for judging whether the plot is crop rotation or fallow, and if the identification results of the same plot on the remote sensing images of the Nth year and the Nth-1 th year are respectively two different crops, the crop rotation is performed; if the identification results of the same plot on the remote sensing images of the Nth year and the Nth-1 th year are fallow land and crops respectively, the fallow land is indicated;
in the land parcel type identification module,
the corn is identified when the pixels on the remote sensing image accord with the condition 1, and the condition 1 is as follows: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 ＞T 4 ；
And when the pixels on the remote sensing images meet the condition 2, rice is identified, and the condition 2: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 ≤T 4 And NDRI 08 ≤T 5 ；
When the pixels on the remote sensing image meet the condition 3, the pixels are identified as beet, and the condition 3: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 ≤T 4 And NDRI 08 ＞T 5 And GI 08 ≤T 6 ；
When the pixel on the remote sensing image meets the condition 4, the image is identified as soybean, and the condition 4: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＞T 3 And REP 08 ≤T 4 And NDVI 08 ＞T 5 And GI 08 ＞T 6 ；
And identifying the fallow land when the pixel identification on the remote sensing image meets the condition 5, wherein the condition 5: t is 1 ＜NDVI 05 ＜T 2 And NDVI 08 ＜T 7 And NDBI 08 ＞T 8 ；
Wherein NDVI is a normalized vegetation index, REP is a red edge position index, NDBI is a normalized construction index, NDRI is a normalized rice index, GI is a greenness index;
NDVI 05 NDVI representing 5-month image;
NDVI 08 represents the 8-month image NDVI value;
REP 08 a red edge position index representing an 8-month image;
NDRI 08 normalized rice index NDRI representing 8-month images;
GI 08 a greenness index representing an image of 8 months;
NDBI 08 a normalized architectural index representing an image for 8 months;
threshold value T 1 -T 8 All are obtained by sample data statistics; means to create random points on the image and visually judgeThe types of the plants are distinguished, the reflectivity of each wave band is respectively extracted, and the threshold value of the crop classification is calculated;
Figure FDA0003804873910000061
where ρ is red 、ρ nir 、ρ swir Respectively the reflectance values of red light wave band, near infrared wave band and short wave infrared wave band pixels;
Figure FDA0003804873910000062
where ρ is the inflection point reflectivity, ρ = (ρ) 670 +ρ 780 )/2，ρ 670 、ρ 700 、ρ 740 、ρ 780 Reflectance at wavelengths of 670nm,700nm,740nm and 780nm respectively;
Figure FDA0003804873910000063
GI＝10000*(ρ blue -ρ green )；
where ρ is red 、ρ swir 、ρ blue 、ρ green Respectively the reflectance values of red light wave band, short wave infrared, blue light wave band and green light wave band.
5. The remote sensing monitoring device for crop rotation fallow facing northeast cold areas as claimed in claim 4, wherein the land parcel type identification module comprises the following steps:
s201: identifying a land parcel on the remote sensing image;
s202: identifying crops in the plots, and marking the plots as plots to be corrected if one plot contains at least two types of crops;
s203: smoothing the remote sensing image containing the plot to be corrected by using a Gaussian filter;
s204: calculating the amplitude and the direction of the gradient of the filtered image by using the finite difference of a first-order differential operator;
s205: carrying out non-maximum value suppression on the gradient amplitude, traversing the image, if the gradient value of a certain pixel is smaller than the gradient values of two adjacent pixels along the gradient direction, judging that the pixel point is not an edge point, and setting the gray value of the pixel point as 0;
s206: detecting and connecting edges with a dual threshold algorithm, calculating two thresholds, high and low, respectively, using cumulative histograms, and separately noting T high And T low Where is greater than T high Is judged as an edge, if less than T low Is judged not to be an edge; if the detection result is greater than or equal to T low But less than or equal to T high If yes, the pixel point is an edge point, otherwise, the pixel point is not an edge point;
s207: if the pixel points serving as the edge points in the detection result are discontinuous to form breakpoints, searching end points by using an n multiplied by n template according to the invariance of the edge local direction; connecting any two searched end points, evaluating whether the connection line of the end points is linear, if the result is linear, using the connection line as an edge, and if the end points cannot be searched or are not linear, manually marking;
s208: and cutting the extracted linear boundary line into the contour boundary of the farmland block to finally form a farmland block unit, thereby obtaining the block unit with only one crop.
6. The remote sensing monitoring device for crop rotation fallow facing northeast cold regions as claimed in claim 4, wherein the result judging module is further used for carrying out statistics according to different change conditions of the identification results of the same plot on the remote sensing images of the Nth year and the Nth-1 year, and analyzing a crop rotation mode, a crop rotation area and a fallow area.
CN201911059350.4A 2019-11-01 2019-11-01 Crop rotation fallow remote sensing monitoring method and device for northeast cold region Active CN111007013B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911059350.4A CN111007013B (en) 2019-11-01 2019-11-01 Crop rotation fallow remote sensing monitoring method and device for northeast cold region

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911059350.4A CN111007013B (en) 2019-11-01 2019-11-01 Crop rotation fallow remote sensing monitoring method and device for northeast cold region

Publications (2)

Publication Number Publication Date
CN111007013A CN111007013A (en) 2020-04-14
CN111007013B true CN111007013B (en) 2022-10-14

Family

ID=70111321

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911059350.4A Active CN111007013B (en) 2019-11-01 2019-11-01 Crop rotation fallow remote sensing monitoring method and device for northeast cold region

Country Status (1)

Country Link
CN (1) CN111007013B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111860150B (en) * 2020-06-11 2024-01-26 中科禾信遥感科技（苏州）有限公司 Lodging rice identification method and device based on remote sensing image
CN112435292B (en) * 2020-12-18 2023-10-31 太原理工大学 Remote sensing image-based rice planting area extraction method
CN113221790A (en) * 2021-05-21 2021-08-06 广东省科学院广州地理研究所 Method and device for generating field crop rotation mode based on radar data
CN115082582B (en) * 2022-06-09 2023-03-10 珠江水利委员会珠江水利科学研究院 True color simulation method, system, equipment and medium for satellite remote sensing data

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7058197B1 (en) * 1999-11-04 2006-06-06 Board Of Trustees Of The University Of Illinois Multi-variable model for identifying crop response zones in a field
CN107064932A (en) * 2017-02-28 2017-08-18 华南农业大学 A kind of construction land change detection method based on time series SAR remote sensing images
WO2018081043A1 (en) * 2016-10-24 2018-05-03 Board Of Trustees Of Michigan State University Methods for mapping temporal and spatial stability and sustainability of a cropping system
CN109471131A (en) * 2018-11-15 2019-03-15 中科禾信遥感科技（苏州）有限公司 It is lain fallow the method and apparatus of situation by remote sensing satellite photo statistical monitoring crop rotation
CN109493247A (en) * 2017-09-11 2019-03-19 苏州市农业科学院 The method that confirmation field crop rotation is lain fallow
CN109508633A (en) * 2018-09-30 2019-03-22 广州地理研究所 A kind of sugarcane distribution recognition methods based on optical remote sensing data
CN109948596A (en) * 2019-04-26 2019-06-28 电子科技大学 A method of rice identification and crop coverage measurement are carried out based on vegetation index model

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10564316B2 (en) * 2014-09-12 2020-02-18 The Climate Corporation Forecasting national crop yield during the growing season
CN106295576B (en) * 2016-08-12 2017-12-12 中国水利水电科学研究院 A kind of water source type analytic method based on nature geography characteristic
CN109635731B (en) * 2018-12-12 2021-04-20 中国科学院深圳先进技术研究院 Method and device for identifying valid farmland, storage medium and processor

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7058197B1 (en) * 1999-11-04 2006-06-06 Board Of Trustees Of The University Of Illinois Multi-variable model for identifying crop response zones in a field
WO2018081043A1 (en) * 2016-10-24 2018-05-03 Board Of Trustees Of Michigan State University Methods for mapping temporal and spatial stability and sustainability of a cropping system
CN107064932A (en) * 2017-02-28 2017-08-18 华南农业大学 A kind of construction land change detection method based on time series SAR remote sensing images
CN109493247A (en) * 2017-09-11 2019-03-19 苏州市农业科学院 The method that confirmation field crop rotation is lain fallow
CN109508633A (en) * 2018-09-30 2019-03-22 广州地理研究所 A kind of sugarcane distribution recognition methods based on optical remote sensing data
CN109471131A (en) * 2018-11-15 2019-03-15 中科禾信遥感科技（苏州）有限公司 It is lain fallow the method and apparatus of situation by remote sensing satellite photo statistical monitoring crop rotation
CN109948596A (en) * 2019-04-26 2019-06-28 电子科技大学 A method of rice identification and crop coverage measurement are carried out based on vegetation index model

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
Analysis of cropping patterns in Sudan"s Gash Spate Irrigation System using Landsat 8 images;YoichiFujihara et al;《Journal of Arid Environments》;20191031;第173卷;全文 *
耕地轮作模式遥感监测;顾晓鹏等;《中国土地科学》;20111231;第25卷(第12期);全文 *

Also Published As

Publication number Publication date
CN111007013A (en) 2020-04-14

Similar Documents

Publication Publication Date Title
CN111007013B (en) Crop rotation fallow remote sensing monitoring method and device for northeast cold region
Chen et al. Mapping croplands, cropping patterns, and crop types using MODIS time-series data
Ludwig et al. A highly automated algorithm for wetland detection using multi-temporal optical satellite data
Zhu et al. Continuous monitoring of forest disturbance using all available Landsat imagery
Johnson et al. Change vector analysis: A technique for the multispectral monitoring of land cover and condition
Zhang et al. A simple automated dynamic threshold extraction method for the classification of large water bodies from landsat-8 OLI water index images
Shahi et al. A novel spectral index to automatically extract road networks from WorldView-2 satellite imagery
Song et al. Cloud detection and analysis of MODIS image
Rahmati et al. Quantitative remote sensing of soil electrical conductivity using ETM+ and ground measured data
CN110648347A (en) Coastline extraction method and device based on remote sensing image
Razak et al. Mapping rubber trees based on phenological analysis of Landsat time series data-sets
Watkins et al. Automating field boundary delineation with multi-temporal Sentinel-2 imagery
CN112183209A (en) Regional crop classification method and system based on multi-dimensional feature fusion
Lu et al. Detection of urban expansion in an urban-rural landscape with multitemporal QuickBird images
CN111060455B (en) Northeast cold-cool area oriented remote sensing image crop marking method and device
S Bhagat Use of remote sensing techniques for robust digital change detection of land: A review
Chust et al. Land cover mapping with patch-derived landscape indices
Wang et al. Unsupervised discrimination between lodged and non-lodged winter wheat: A case study using a low-cost unmanned aerial vehicle
Taghadosi et al. Trend analysis of soil salinity in different land cover types using Landsat time series data (case study Bakhtegan Salt Lake)
Izzuddin et al. Analysis of multispectral imagery from unmanned aerial vehicle (UAV) using object-based image analysis for detection of ganoderma disease in oil palm
Sabzchi-Dehkharghani et al. Recognition of different yield potentials among rain-fed wheat fields before harvest using remote sensing
Bucha et al. Analysis of MODIS imagery for detection of clear cuts in the boreal forest in north-west Russia
Zhang et al. Chlorophyll content detection of field maize using RGB-NIR camera
Qiu et al. Automated cropping intensity extraction from isolines of wavelet spectra
Yadav et al. An automated image processing method for segmentation and quantification of rust disease in maize leaves

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
GR01 Patent grant
GR01 Patent grant