CN-121214241-B - Method, device, equipment and medium for monitoring occupied state of cultivated land

CN121214241BCN 121214241 BCN121214241 BCN 121214241BCN-121214241-B

Abstract

The application of the invention provides a method, a device, equipment and a medium for monitoring the occupied state of cultivated lands, wherein the method comprises the steps of acquiring multi-temporal video image data of a target monitoring area and a cultivated land distribution reference graph to obtain distribution information of the cultivated lands; extracting boundary lines of cultivated land areas and non-cultivated land areas in video image data of each time phase to obtain spatial position change data of the boundary lines, analyzing boundary feature sets of each time phase according to a time sequence to identify abnormal areas with boundary changes to obtain abnormal area change data, comparing and analyzing the abnormal area change data with a cultivated land distribution reference graph of multiple time phases to obtain a comparison result, extracting morphological features and expansion directions of the abnormal areas, simulating future boundary changes of the abnormal areas to generate a prediction distribution map of potential occupied areas to obtain an influence range of the abnormal areas, and obtaining a monitoring result. The invention can improve the continuity and accuracy of monitoring the occupied state of the cultivated land.

Inventors

CHEN JINGYUN
SUN YING
HU MIN
SU CHUANG
DUAN RUI
Zeng Xueer
PENG TE
TAN FEI
XIE YI
CHEN YUN
ZHONG JIANLI
TAN YIWEN
LUO LONG
OuYang Renbin

Assignees

广州城市规划技术开发服务部有限公司

Dates

Publication Date: 20260508
Application Date: 20250805

Claims (8)

1. A method for monitoring the occupancy state of a cultivated land, comprising: acquiring multi-time-phase video image data of a target monitoring area, acquiring a multi-time-phase cultivated land distribution reference graph, and analyzing based on the cultivated land distribution reference graph to obtain distribution information of cultivated lands; extracting boundary lines of the cultivated land area and the non-cultivated land area of the video image data of each time phase according to the distribution information, generating boundary feature sets of each time phase, and further obtaining spatial position change data of the boundary lines; Analyzing the boundary feature set of each time phase according to the time sequence of the space position change data, identifying an abnormal region of boundary change, and further obtaining abnormal region change data; Comparing and analyzing the abnormal region change data with a multi-phase farmland distribution reference graph to obtain a comparison result, extracting morphological characteristics and an expansion direction of an abnormal region based on the comparison result, and obtaining a spatial expansion trend according to the morphological characteristics and the expansion direction; Simulating future boundary changes of the abnormal region according to the space expansion trend to generate a prediction distribution diagram of a potential occupation range, further obtaining an influence range of the abnormal region according to the prediction distribution diagram, and obtaining a monitoring result of the occupation state of the cultivated land according to the comparison result, the influence range and the space expansion trend; extracting boundary lines of the cultivated land area and the non-cultivated land area in the video image data of each time phase according to the distribution information, generating boundary feature sets of each time phase, and further obtaining spatial position change data of the boundary lines, wherein the method comprises the following steps: According to the distribution information, carrying out image segmentation by adopting a preset U-Net model, extracting to obtain boundary lines of a cultivated land area and a non-cultivated land area of video image data of each time phase, and generating boundary feature sets of each time phase; calculating the space coordinates of boundary lines of each time phase through the boundary feature set to obtain the space positions of the boundary lines; detecting the spatial variation of the boundary line through differential analysis when the difference value of the spatial position information of adjacent time phases exceeds a preset difference threshold value; performing cluster analysis on pixels corresponding to the spatial variation of the boundary line through a K-Means algorithm to obtain spatial position variation data of the boundary line; The monitoring method further comprises the following steps: Extracting image texture features of the area corresponding to the influence range; Analyzing the edge directivity and roughness change of the image texture based on the image texture characteristics to obtain a characteristic data set; according to the characteristic data set, calculating to obtain boundary continuity and diffusion strength information of the occupied behavior; Analyzing a first corresponding relation between the edge directivity and the boundary moving track according to the boundary continuity and the diffusion intensity information and combining the space expansion trend, extracting consistency characteristics of linear extension direction and angle offset according to the first corresponding relation, and further determining an expansion path of the occupied behavior; Aiming at the expansion path, analyzing a second corresponding relation between the roughness change and the expansion speed change, extracting dynamic association of the land coverage heterogeneity and the area growth rate according to the second corresponding relation, and further predicting to obtain a simulation path of a dynamic expansion track; and carrying out matching analysis according to the simulation path and the video image data acquired in real time to obtain a matching analysis result, and updating the monitoring result in real time by utilizing the matching analysis result.
2. The method for monitoring an occupancy state of a cultivated land according to claim 1, wherein the obtaining the influence range of the abnormal area according to the prediction distribution map comprises: dividing the prediction distribution map into a plurality of area units, and identifying and obtaining farmland units adjacent to the abnormal area through space proximity analysis; extracting the topography information and land utilization attributes of the tilling unit; judging the potential risk level of the cultivated land unit according to the terrain information and the land utilization attribute; and determining the cultivated land units with the potential risk levels larger than the preset level threshold as the influence range of the abnormal area.
3. The method for monitoring the occupancy state of a cultivated land according to claim 1, wherein the performing a matching analysis according to the simulated path and video image data acquired in real time to obtain a matching analysis result comprises: performing spatial registration processing on the analog path and video image data acquired in real time by adopting a spatial data processing tool to obtain registered image path data; And performing spatial difference analysis on the registered image path data through superposition analysis to obtain difference data and matching data, and further determining the difference data and the matching data as the matching analysis result.
4. The method for monitoring the occupancy state of a cultivated land according to claim 1, wherein the acquiring the multi-temporal video image data of the target monitoring area comprises: The method comprises the steps of obtaining a multi-phase image sequence, carrying out radiation correction and geometric registration on the image sequence, standardizing data, and carrying out denoising processing based on median filtering on the standardizing data to obtain the multi-phase video image data.
5. The method for monitoring an occupancy state of a cultivated land according to claim 1, wherein simulating future boundary changes of an abnormal area according to the spatial expansion trend to generate a prediction distribution map of a potential occupancy range comprises: determining the category of the spatial expansion trend; the category and the multi-temporal video image data are fused by utilizing a rasterization technology, so that an occupied dynamic change mode is determined; and simulating future boundary changes of the abnormal region on the basis of the dynamic change mode to generate a prediction distribution diagram of the potential occupation range.
6. A monitoring device for the occupied state of cultivated land is characterized by comprising an information acquisition module, a generation module, an analysis module, an extraction module and a monitoring module, wherein, The information acquisition module is used for acquiring multi-time-phase video image data of the target monitoring area, acquiring a multi-time-phase cultivated land distribution reference graph, and analyzing based on the cultivated land distribution reference graph to obtain the distribution information of the cultivated lands; The generation module is used for extracting boundary lines of the cultivated land area and the non-cultivated land area in the video image data of each time phase according to the distribution information, generating boundary feature sets of each time phase, and further obtaining spatial position change data of the boundary lines; the analysis module is used for analyzing the boundary feature set of each time phase according to the time sequence of the space position change data, identifying an abnormal region of boundary change and further obtaining abnormal region change data; The extraction module is used for carrying out comparison analysis on the abnormal region change data and a multi-phase farmland distribution reference graph to obtain a comparison result, extracting morphological characteristics and an expansion direction of the abnormal region based on the comparison result, and obtaining a space expansion trend according to the morphological characteristics and the expansion direction; The monitoring module is used for simulating future boundary changes of the abnormal area according to the space expansion trend to generate a prediction distribution diagram of a potential occupation range, further obtaining an influence range of the abnormal area according to the prediction distribution diagram, and obtaining a monitoring result of the cultivated land occupation state according to the comparison result, the influence range and the space expansion trend; The generation module extracts boundary lines of the cultivated land area and the non-cultivated land area in the video image data of each time phase according to the distribution information, generates boundary feature sets of each time phase, and further obtains spatial position change data of the boundary lines, and the generation module comprises the following steps: The generation module performs image segmentation by adopting a preset U-Net model according to the distribution information, extracts boundary lines of the cultivated land area and the non-cultivated land area of the video image data of each time phase, and generates boundary feature sets of each time phase; calculating the space coordinates of boundary lines of each time phase through the boundary feature set to obtain the space positions of the boundary lines; detecting the spatial variation of the boundary line through differential analysis when the difference value of the spatial position information of adjacent time phases exceeds a preset difference threshold value; performing cluster analysis on pixels corresponding to the spatial variation of the boundary line through a K-Means algorithm to obtain spatial position variation data of the boundary line; The monitoring device further comprises a monitoring result updating module, wherein the monitoring result updating module is used for: Extracting image texture features of the area corresponding to the influence range; Analyzing the edge directivity and roughness change of the image texture based on the image texture characteristics to obtain a characteristic data set; according to the characteristic data set, calculating to obtain boundary continuity and diffusion strength information of the occupied behavior; Analyzing a first corresponding relation between the edge directivity and the boundary moving track according to the boundary continuity and the diffusion intensity information and combining the space expansion trend, extracting consistency characteristics of linear extension direction and angle offset according to the first corresponding relation, and further determining an expansion path of the occupied behavior; Aiming at the expansion path, analyzing a second corresponding relation between the roughness change and the expansion speed change, extracting dynamic association of the land coverage heterogeneity and the area growth rate according to the second corresponding relation, and further predicting to obtain a simulation path of a dynamic expansion track; and carrying out matching analysis according to the simulation path and the video image data acquired in real time to obtain a matching analysis result, and updating the monitoring result in real time by utilizing the matching analysis result.
7. A terminal device comprising a processor, a memory and a computer program stored in the memory and configured to be executed by the processor, the processor implementing the method of monitoring the occupancy state of cultivated land according to any one of claims 1 to 5 when the computer program is executed.
8. A computer readable storage medium, characterized in that the computer readable storage medium comprises a stored computer program, wherein the computer program, when run, controls a device in which the computer readable storage medium is located to perform the method for monitoring the occupancy state of a cultivated land according to any one of claims 1 to 5.

Description

Method, device, equipment and medium for monitoring occupied state of cultivated land Technical Field The present invention relates to the field of image recognition, and in particular, to a method, an apparatus, a device, and a medium for monitoring an occupied state of a cultivated land. Background The farmland protection is an important foundation for guaranteeing the grain safety and ecological balance, and the monitoring and management of the farmland protection system has irreplaceable strategic significance in the field of homeland resources. The phenomenon that the cultivated land is illegally occupied not only threatens the capability of agricultural production, but also can cause a series of social and environmental problems, so that the timely discovery and accurate identification of the illegal occupancy behavior are particularly critical. However, current monitoring methods for illegal occupancy of cultivated land still have shortcomings. Many schemes rely on image data analysis at a single time point too, and it is difficult to capture the dynamic evolution process of the occupation behavior in time and space, and especially when the occupation phenomenon of gradual diffusion is faced, the expansion trend and the influence range of the occupation behavior cannot be accurately judged. This limitation results in a lack of continuity and predictive capability of the monitoring results, and it is difficult to provide an effective decision basis for the relevant departments in time. Disclosure of Invention The application of the invention provides a method, a device, equipment and a medium for monitoring the occupied state of cultivated land, which can improve the continuity of monitoring the occupied state of cultivated land. In order to solve the technical problems, the present invention provides a method for monitoring an occupied state of a cultivated land, including: acquiring multi-time-phase video image data of a target monitoring area, acquiring a multi-time-phase cultivated land distribution reference graph, and analyzing based on the cultivated land distribution reference graph to obtain distribution information of cultivated lands; extracting boundary lines of the cultivated land area and the non-cultivated land area of the video image data of each time phase according to the distribution information, generating boundary feature sets of each time phase, and further obtaining spatial position change data of the boundary lines; Analyzing the boundary feature set of each time phase according to the time sequence of the space position change data, identifying an abnormal region of boundary change, and further obtaining abnormal region change data; Comparing and analyzing the abnormal region change data with a multi-phase farmland distribution reference graph to obtain a comparison result, extracting morphological characteristics and an expansion direction of an abnormal region based on the comparison result, and obtaining a spatial expansion trend according to the morphological characteristics and the expansion direction; Simulating future boundary changes of the abnormal region according to the space expansion trend to generate a prediction distribution diagram of a potential occupation range, further obtaining an influence range of the abnormal region according to the prediction distribution diagram, and obtaining a monitoring result of the occupied state of the cultivated land according to the comparison result, the influence range and the space expansion trend. Preferably, the obtaining the influence range of the abnormal area according to the prediction distribution diagram includes: dividing the prediction distribution map into a plurality of area units, and identifying and obtaining farmland units adjacent to the abnormal area through space proximity analysis; extracting the topography information and land utilization attributes of the tilling unit; judging the potential risk level of the cultivated land unit according to the terrain information and the land utilization attribute; and determining the cultivated land units with the potential risk levels larger than the preset level threshold as the influence range of the abnormal area. Preferably, the monitoring method further comprises: Extracting image texture features of the area corresponding to the influence range; Analyzing the edge directivity and roughness change of the image texture based on the image texture characteristics to obtain a characteristic data set; according to the characteristic data set, calculating to obtain boundary continuity and diffusion strength information of the occupied behavior; Analyzing a first corresponding relation between the edge directivity and the boundary moving track according to the boundary continuity and the diffusion intensity information and combining the space expansion trend, extracting consistency characteristics of linear extension direction and angle offset according to the first co