CN118916831B - A method for detecting irrigation of farmland plots by integrating multi-source remote sensing data - Google Patents

Abstract

The invention discloses a farmland land parcel irrigation detection method integrating multisource remote sensing data, which comprises the steps of obtaining time series remote sensing data and auxiliary data, preprocessing, obtaining normalized difference moisture index, surface temperature and precipitation amount under the land parcel scale based on the data, constructing a NDMI-LST characteristic space scatter diagram with the normalized difference moisture index being an abscissa and the surface temperature being an ordinate under the land parcel scale, obtaining upper and lower boundary lines of NDMI-LST characteristic space, calculating temperature vegetation drought indexes, constructing a time series curve of the temperature vegetation drought indexes under the land parcel scale, setting a land parcel irrigation detection rule set, establishing a land parcel irrigation detection algorithm with a self-adaptive dynamic threshold, obtaining land parcel irrigation period, and obtaining analysis and statistics area irrigation information. Therefore, by adopting the method, farmland land parcel irrigation event and period information can be extracted rapidly and accurately, and references are provided for decisions such as water resource management, agricultural production activities and the like.

Description

A method for detecting irrigation of farmland plots by integrating multi-source remote sensing data

Technical Field

The present invention relates to the field of agricultural remote sensing monitoring, and in particular to a method for detecting irrigation of farmland plots by integrating multi-source remote sensing data.

Background Art

With the rapid development of spatial information technology, remote sensing and geospatial mapping technology, which has the advantages of wide coverage, high information timeliness and low operating cost, has become the main means of obtaining regional agricultural irrigation information.

Since irrigation can cause changes in soil moisture or crop moisture, it can also cause changes in crop growth. Therefore, by using multi-temporal remote sensing observations to capture the differences in spectral characteristics of soils with different moisture contents and crops in different growth states before and after irrigation, it is possible to quantify soil moisture changes or crop growth changes and extract farmland irrigation information. At present, the change detection model based on vegetation index has become the main method for extracting farmland irrigation information. For example, Tian Xin et al. constructed an irrigation area extraction model based on the difference between VSWI and TVDI, and realized the extraction of irrigation area in the Shenwu irrigation area in Inner Mongolia. Du Enyu et al. calculated the TVDI and MPDI indexes using Landsat images, and constructed an irrigation area extraction model based on the difference between the two indexes before and after irrigation. Yang Yongmin et al. quantified the changes in Sentinel-1 microwave backscatter coefficients caused by irrigation events and proposed an irrigation area extraction method based on time series difference and local threshold method. However, most of these studies focus on single irrigation information with known irrigation time (such as irrigation distribution and area, water consumption, etc.), but lack in-depth exploration of multiple irrigation information during the crop growing season (such as irrigation times, irrigation time, irrigation cycle, etc.).

Compared with the simple extraction of agricultural irrigation range and area mentioned above, using remote sensing technology to accurately identify agricultural irrigation events and cycle information has higher application value and is more challenging; and with the development of remote sensing technology, the integration of multi-source remote sensing satellites can obtain more dimensional and longer time series observations of crop-soil systems, which provides the possibility of extracting multi-dimensional irrigation information of farmland during the crop growing season. Zhai Yongguang et al. jointly used Sentinel-1\2\3 satellite data to construct the Vegetation Temperature Condition Index (VTCI) time series curve, and combined with peak peak valley detection and matching algorithms, successfully extracted the irrigation time, number and cycle of the irrigation area. Chen et al. integrated MODIS time series, Landsat images and auxiliary data, and based on the analysis of the crop greenness index's response to irrigation, constructed an irrigation event detection algorithm based on threshold segmentation, and extracted irrigation frequency and time information. These studies have greatly improved the precision of remote sensing monitoring of farmland irrigation.

Although existing studies have achieved good results in remote sensing monitoring of farmland irrigation, they still face new challenges: (1) There are many studies on irrigation spatial distribution and probability mapping using multi-source remote sensing data, but there is a relative lack of research on irrigation detection (i.e., irrigation events and cycles). (2) In existing irrigation event detection algorithms based on threshold segmentation, the same threshold is set for the entire study area, and there is a lack of dynamic threshold algorithms that can adapt to changes in the plot space and crop growth period. (3) Existing studies either take the entire study area as an irrigation unit, which makes it difficult to express the differences in irrigation between different farmland plots; or use a single pixel as an irrigation unit, which is easily interfered by salt and pepper noise, limits the accuracy of remote sensing irrigation detection, and also makes it difficult to accurately correspond irrigation remote sensing monitoring to actual agricultural farmland management units.

In summary, the current irrigation detection methods have many problems in remote sensing irrigation detection unit determination, multi-source data fusion strategy, computational model construction, algorithm application, etc., and lack a dynamic threshold method that can adapt to changes in plot space and crop growth period. Therefore, there is an urgent need for a plot-scale farmland irrigation event identification and cycle extraction method that integrates multi-source remote sensing data to support agricultural remote sensing monitoring and agricultural water resources management.

Summary of the invention

The purpose of the present invention is to provide a method for detecting irrigation of farmland plots by fusing multi-source remote sensing data, which can be realized.

To achieve the above object, the present invention provides a method for detecting irrigation of farmland plots by integrating multi-source remote sensing data, comprising the following steps:

S1. Obtain time series remote sensing data and auxiliary data, and perform preprocessing to obtain a raster dataset with the same geographic spatial reference and spatiotemporal resolution and with pixel spatiotemporal alignment;

S2. Based on the acquired time series remote sensing data and auxiliary data, the normalized difference moisture index, surface temperature and precipitation at the plot scale are obtained to construct a plot-scale irrigation remote sensing monitoring dataset;

S3, construct a scatter plot of the NDMI-LST feature space with the normalized difference moisture index as the horizontal coordinate and the surface temperature as the vertical coordinate at the plot scale, extract and linearly fit the upper and lower boundary lines of the NDMI-LST feature space, and calculate the temperature vegetation drought index of the plot;

S4, construct the time series curve of TVDI at the plot scale, set the plot irrigation detection rule set, establish the plot irrigation detection algorithm with adaptive dynamic threshold, and obtain the plot irrigation cycle;

S5. Analyze and count regional irrigation information based on the irrigation cycle of the plot.

Preferably, in step S3, the upper and lower boundary lines of the NDMI-LST feature space specifically include:

The normalized difference moisture index NDMI on the horizontal axis is divided into intervals, and the mean, standard deviation and confidence interval of the surface temperature LST in the divided interval are statistically calculated. The maximum and minimum values of LST in the confidence interval are found, and the corresponding scattered points are recorded as the upper boundary point and lower boundary point of the moisture of the crops in the divided interval; the upper boundary point and lower boundary point of the moisture are fitted to obtain the upper boundary line and lower boundary line of the NDMI-LST space.

Preferably, the upper boundary line and the lower boundary line of the NDMI-LST space are specifically expressed as:

f _u = a _u × NDMI + b _u

f _b = a _b × NDMI + b _b

Wherein, _fu is the upper and lower boundary lines of the NDMI-LST space, _au and _bu are the slope and intercept of the upper boundary line obtained by fitting, _fb is the upper and lower boundary lines of the NDMI-LST space, and _ab and _bb are the slope and intercept of the lower boundary line obtained by fitting.

Preferably, in step S3, the temperature vegetation drought index TVDI is calculated as follows:

Ts＝ _fu (NDMI)-LST

MI＝ _fu (NDMI) _-fb (NDMI)

TVDI＝Ts/MI

Where Ts is the distance between the plot coordinates (NDMI, LST) and the upper boundary line in the vertical direction, and MI is the distance between the upper and lower boundaries in the vertical direction.

Preferably, in step S4, setting the plot irrigation detection rule set includes:

Maximum number of irrigations during the entire growing season;

Minimum threshold for increase in TVDI before and after irrigation;

The minimum time interval between two adjacent irrigations;

The minimum precipitation that causes the TVDI to increase beyond the minimum threshold.

Therefore, the present invention adopts the above-mentioned farmland irrigation detection method integrating multi-source remote sensing data, which has the following technical effects:

(1) Using NDMI, we can achieve a stronger negative correlation between NDMI and LST in the crop-soil system and accurately express the degree of water stress in the crop-soil system;

(2) Using farmland plots as irrigation analysis units, the interference of pixel salt and pepper noise is reduced and the accuracy of irrigation detection is improved;

(3) Based on the TVDI time series curve of the plot, an adaptive dynamic threshold plot irrigation detection algorithm is established to more accurately extract the irrigation events and irrigation cycle information of the plot.

The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a flow chart of an embodiment of a method for detecting irrigation of farmland plots by integrating multi-source remote sensing data;

FIG2 is a flowchart of multi-source remote sensing data preprocessing in an embodiment of a method for detecting irrigation of farmland plots by fusing multi-source remote sensing data;

FIG3 is a TVDI calculation flow chart of an embodiment of a method for detecting irrigation of farmland plots by integrating multi-source remote sensing data;

FIG4 is a flow chart of irrigation cycle detection based on TVDI timing curve in an embodiment of a method for detecting irrigation of farmland plots by integrating multi-source remote sensing data;

FIG5 is an irrigation information map of farmland plots in Zhongning County in an embodiment of a method for detecting farmland plots by integrating multi-source remote sensing data. FIG5 (a) is a spatial distribution map of irrigation times, and FIG5 (b) is a monthly statistical map of irrigation areas.

DETAILED DESCRIPTION

The present invention can be explained in more detail by the following examples. The purpose of disclosing the present invention is to protect all changes and improvements within the scope of the present invention. The present invention is not limited to the following examples.

As shown in FIG1 , the present invention provides a method for detecting irrigation of farmland plots by integrating multi-source remote sensing data, comprising the following steps:

S1. Obtain time series remote sensing data (including Sentinel-2 images, MOD11A1 surface temperature data, precipitation data, etc.) and auxiliary data (including farmland plot vector data, crop planting and irrigation statistics, etc.), and preprocess them. Based on Sentinel-2 images, generate a raster dataset with the same geographic spatial reference and spatiotemporal resolution and pixel spatiotemporal alignment, as shown in Figure 2.

S2. Based on Sentinel-2 images, the Normalized Difference Moisture Index (NDMI) is calculated; based on MOD11A1 surface temperature data, the land surface temperature (LST) is extracted; based on precipitation data, the surface precipitation is extracted; based on farmland plot vector data, the pixel band eigenvalues, surface temperature LST, and precipitation at the plot scale are calculated respectively, and an irrigation remote sensing monitoring dataset at the plot scale is constructed.

In particular, the optical image time series involved in this embodiment is not limited to Sentinel-2 images, but is also applicable to other multi-temporal optical images (such as NASA Landsat images).

S3. Construct a scatter plot of the NDMI-LST feature space at the plot scale with the Normalized Difference Moisture Index (NDMI) as the horizontal axis and the Land Surface Temperature (LST) as the vertical axis, extract and linearly fit the upper and lower boundary lines (i.e., the dry edge and the wet edge) of the NDMI-LST feature space, and calculate the Temperature Vegetation Dryness Index (TVDI) of the plot, as shown in Figure 3.

Among them, the upper and lower boundaries of the NDMI-LST feature space are as follows:

First, the horizontal axis normalized difference moisture index NDMI is divided into intervals of 0.01;

Secondly, extract the scattered points in each divided interval Q, and calculate the mean, standard deviation and confidence interval of the land surface temperature LST (in this embodiment, a 95% confidence interval is taken to remove abnormal values caused by factors such as noise);

Then, find the maximum and minimum values of LST within the confidence interval, and record the corresponding scattered points as the upper and lower boundary points of crop moisture that divides interval Q;

Finally, the upper boundary points of all divided intervals are extracted to form the upper boundary line of the scatter plot, and the upper boundary line of the NDMI-LST space is generated by using the least squares straight line fitting method:

f _u = a _u × NDMI + b _u

Where _fu is the upper and lower boundary lines of the NDMI-LST space, and _au and _bu are the slope and intercept of the upper boundary line obtained by fitting.

Similarly, obtain the lower boundary line of the NDMI-LST space:

f _b = a _b × NDMI + b _b

Where _fb is the upper and lower boundary lines of the NDMI-LST space, and _ab and _bb are the slope and intercept of the lower boundary line obtained by fitting.

In the NDMI-LST scatter point space, with the upper and lower boundary lines as constraints, the coordinates of each farmland plot are calculated as (NDMI, LST) and the distance Ts between the upper boundary line and the distance MI between the upper and lower boundaries in the vertical direction, so as to obtain the farmland plot temperature drought vegetation index TVDI:

Ts＝ _fu (NDMI)-LST

MI＝ _fu (NDMI) _-fb (NDMI)

TVDI＝Ts/MI

S4. As shown in FIG4 , a time series curve of TVDI at the plot scale is constructed, a plot irrigation detection rule set based on TVDI changes is set, and an adaptive dynamic threshold plot irrigation detection algorithm (Adaptive Threshold for Irrigation Detection Algorithm, ATID) is established.

The rules for setting irrigation detection include:

(1) The maximum number of irrigation times M during the entire crop growing season;

(2) the minimum threshold T of the increase in TVDI before and after irrigation;

(3) The minimum time interval between two consecutive irrigations is D days;

(4) The minimum precipitation m that causes TVDI to increase beyond T.

The parameters such as M, T, D, m, etc. involved in the irrigation detection rules must be set in combination with the regional environment and conditions.

This embodiment selects Zhongning County, Ningxia, to discuss the setting of irrigation detection rule parameters. The summer crops in the Yellow River irrigation area of Zhongning County are mainly corn, with a growth period of about 150 days (mid-to-late April to late September). It is recommended to irrigate 11 times during the corn growing season. The threshold T for the increase in TVDI caused by irrigation is the difference between the mean values of TVDI at all peaks and troughs. Combined with the recommended irrigation period, the minimum time interval between two adjacent irrigations is set to 10 days. Combined with the total water requirement of 1476.6 mm for corn in Zhongning County and a growth period of 150 days, the average daily irrigation water requirement for corn is 9.84 mm; this data provides a benchmark for judging whether daily precipitation is effective precipitation. When daily precipitation exceeds this value, it can be considered that precipitation (non-irrigation) causes the TVDI to rise, that is, erroneous irrigation detection caused by precipitation is excluded. It should be noted that the above rules can be added or modified according to the actual situation in the region in the application to improve the applicability and reliability of irrigation detection.

The plot irrigation detection takes the TVDI event sequence curve as input; first, search for the curve peak point set _Vf and the trough point set _Vg along the TVDI curve time axis; then, starting from any peak point, search forward along the TVDI time axis for trough points that meet the rules, and form a trough-peak combination ( _Vg , _Vf ); finally, select the first M combinations with the largest TVDI interpolation values from the trough-peak combinations as the irrigation cycle of the farmland plot, and use the trough points of the combination as the irrigation time.

S5. Based on the extraction results of irrigation cycles of farmland plots, regional irrigation information is statistically analyzed from multiple dimensions. From the perspective of irrigation area analysis, the monthly (or crop phenological period) irrigation area of the region is statistically analyzed, and a monthly irrigation spatial distribution thematic map is produced; from the perspective of irrigation frequency analysis, the monthly (or crop phenological period) irrigation frequency of the region or a certain plot is statistically analyzed, and a monthly irrigation frequency spatial distribution map is produced, as shown in Figure 5.

From the monthly statistics of irrigation area, as shown in Figure 5 (b), the irrigation of farmland plots reached a peak from March to May, began to decline in June, rebounded in July, reached the lowest point in August, rose again in September, and then gradually decreased from October to the end of the year. The irrigation area is relatively high in spring and summer, while it is relatively low in autumn and winter, which is closely related to the crop growth cycle and climatic conditions. High values of irrigation area generally appear in the period of vigorous crop growth in spring and summer, while low values of irrigation area appear in the slow growth or dormancy stage of crops in autumn and winter; and the smallest irrigation area in August is mainly because the precipitation in that month is sufficient to meet the growth of crops. The low irrigation area at the end of the year and the beginning of the year indicates that it is not the crop growing season and the water demand is small.

From the spatial distribution of irrigation times, as shown in Figure 5 (a), April to May is the peak period of irrigation in Zhongning County. Basically, there are 2 to 3 irrigations for each plot. Among them, the plots irrigated twice are concentrated on both sides of the Yellow River in the middle of the study area, and the plots irrigated three times are concentrated in the south of the Yellow River and on both sides of the Yellow River in the northeast. From June to August, most plots were irrigated only once, and the overall number of irrigated plots was less than that in other months. From September to October, the number of irrigated plots increased, but still mainly irrigated once, and some plots were irrigated twice. These results are consistent with the judgment and understanding of the spatiotemporal distribution of agricultural irrigation in Zhongning County.

Therefore, the present invention adopts the above-mentioned farmland irrigation detection method that integrates multi-source remote sensing data, which can quickly and accurately extract farmland irrigation events and cycle information, and generate regional agricultural irrigation information maps to provide a reference for supporting decisions such as water resources management and agricultural production activities.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims (4)

1. A method for detecting irrigation of farmland plots by integrating multi-source remote sensing data, characterized in that it comprises the following steps: S1. Obtain time series remote sensing data and auxiliary data, and perform preprocessing to obtain a raster dataset with the same geographic spatial reference and spatiotemporal resolution and with pixel spatiotemporal alignment; S2. Based on the acquired time series remote sensing data and auxiliary data, the normalized difference moisture index, surface temperature and precipitation at the plot scale are obtained to construct a plot-scale irrigation remote sensing monitoring dataset; S3, construct a scatter plot of the NDMI-LST feature space with the normalized difference moisture index as the horizontal coordinate and the surface temperature as the vertical coordinate at the plot scale, extract and linearly fit the upper and lower boundary lines of the NDMI-LST feature space, and calculate the temperature vegetation drought index of the plot; S4, construct the time series curve of TVDI at the plot scale, set the plot irrigation detection rule set, establish the plot irrigation detection algorithm with adaptive dynamic threshold, and obtain the plot irrigation cycle; Among them, the plot irrigation detection rule set includes: Maximum number of irrigations during the entire growing season; Minimum threshold for increase in TVDI before and after irrigation; The minimum time interval between two adjacent irrigations; The minimum precipitation that causes the TVDI to increase beyond the minimum threshold; S5. Analyze and count regional irrigation information based on the irrigation cycle of the plot. 2. The method for detecting irrigation of farmland plots by integrating multi-source remote sensing data according to claim 1, characterized in that, in step S3, the upper and lower boundary lines of the NDMI-LST feature space specifically include: The normalized difference moisture index NDMI on the horizontal axis is divided into intervals, and the mean, standard deviation and confidence interval of the surface temperature LST in the divided interval are statistically calculated. The maximum and minimum values of LST in the confidence interval are found, and the corresponding scattered points are recorded as the upper boundary point and lower boundary point of the moisture of the crops in the divided interval; the upper boundary point and lower boundary point of the moisture are fitted to obtain the upper boundary line and lower boundary line of the NDMI-LST space. 3. A method for detecting irrigation of farmland plots by integrating multi-source remote sensing data according to claim 2, characterized in that the upper boundary line and the lower boundary line of the NDMI-LST space are specifically expressed as: f _u = a _u × NDMI + b _u f _b = a _b × NDMI + b _b Wherein, _fu is the upper and lower boundary lines of the NDMI-LST space, _au and _bu are the slope and intercept of the upper boundary line obtained by fitting, _fb is the upper and lower boundary lines of the NDMI-LST space, and _ab and _bb are the slope and intercept of the lower boundary line obtained by fitting. 4. The method for detecting irrigation of farmland plots by integrating multi-source remote sensing data according to claim 3, characterized in that, in step S3, the temperature vegetation drought index TVDI is calculated as follows: Ts＝ _fu (NDMI)-LST MI＝ _fu (NDMI) _-fb (NDMI TVDI＝Ts/MI Where Ts is the distance between the plot coordinates (NDMI, LST) and the upper boundary line in the vertical direction, and MI is the distance between the upper and lower boundaries in the vertical direction.