CN115830804B - Pipeline geological disaster intelligent early warning negative sample sampling method under constraint of easily-generated partition - Google Patents
Pipeline geological disaster intelligent early warning negative sample sampling method under constraint of easily-generated partition Download PDFInfo
- Publication number
- CN115830804B CN115830804B CN202211298958.4A CN202211298958A CN115830804B CN 115830804 B CN115830804 B CN 115830804B CN 202211298958 A CN202211298958 A CN 202211298958A CN 115830804 B CN115830804 B CN 115830804B
- Authority
- CN
- China
- Prior art keywords
- negative sample
- disaster
- geological disaster
- rainfall
- early warning
- 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
Links
Landscapes
- Pit Excavations, Shoring, Fill Or Stabilisation Of Slopes (AREA)
Abstract
The invention discloses an intelligent early warning negative sample sampling method for pipeline geological disasters under the constraint of easily-developed partitions, which comprises the following steps: selecting factors required by early warning of a pipeline geological disaster area, including disaster-tolerant environment and induction factors; drawing a corresponding FR grid chart by using historical geological disaster points along the pipeline and data of each disaster-tolerant environment layer by using a GIS technology; drawing an accumulated rainfall graph of a primary rainfall process up to the occurrence date of each geological disaster point by combining the pipeline geological disaster easy-occurrence subareas; and overlapping the FR layer and the easily-generated partition layer, defining a negative sample acquisition unit area, acquiring the spatial attribute and the time attribute of a negative sample in the negative sample acquisition unit, and extracting the disaster-enriched environmental factor data and the induction factor data of corresponding time of the spatial geographic position corresponding to each negative sample point to form a negative sample set. The invention improves the effectiveness, representativeness and acquisition quality of the intelligent early warning negative sample of the pipeline geological disaster.
Description
Belonging to the field of
The invention belongs to the technical field of oil and gas pipeline geological disaster prevention and control, and particularly relates to an intelligent pipeline geological disaster early warning negative sample sampling method under the constraint of easily-generated partition.
Technical Field
The regional geological disaster monitoring and early warning is an important pipeline geological disaster risk prevention and control means, and can pre-judge the geological disaster risk along the pipeline in advance, so that risk prevention is performed in advance. However, the traditional geological disaster area early warning model has the problems of low early warning accuracy and high false alarm rate all the time, and the regional geological disaster monitoring early warning model needs to be continuously researched and perfected. In recent years, along with the development of artificial intelligence technology, the artificial intelligence technology has been gradually applied to the field of geological disaster prevention and control, but the intelligent early warning of the geological disaster of the oil and gas pipeline is still studied. The development of artificial intelligence technology provides a new technical means for early warning of pipeline geological disaster areas. Although the pipeline geological disaster investigation accumulates a large amount of data of the geological disaster occurring points, non-geological disaster points required by the artificial intelligence algorithm are difficult to obtain, the non-geological disaster points are called negative sample points, and the effective selection of the negative sample points is a key problem encountered in constructing an intelligent pipeline region geological disaster early warning model. In the past, the problem of negative sample selection is solved, the negative samples are selected in a random mode, but the negative samples which are selected randomly are not necessarily all effective, so that the accuracy of the intelligent early warning model is restricted, and the application effect of the early warning model is influenced.
Disclosure of Invention
Aiming at the problems, the invention provides the pipeline geological disaster intelligent early warning negative sample sampling method under the constraint of the easily-generated partition, which can effectively select negative sample points under the constraint of the FR model and the easily-generated partition, overcomes the defect of randomly taking the negative sample points outside the generated geological disaster points, is beneficial to the construction of an artificial intelligence-based pipeline geological disaster area early warning sample set and is beneficial to the improvement of the accuracy of an area early warning model.
The invention is realized by the following technical scheme, which comprises the following technical steps:
(1) And selecting factors required by early warning of a pipeline geological disaster area. The early warning of the geological disaster area of the pipeline is limited by two factors, namely a disaster-pregnant environment and an induction factor. The disaster-tolerant environment comprises stratum, rock-soil body type, elevation, gradient, slope direction, plane curvature, water system, fault, annual average rainfall and the like; the inducing factors mainly refer to rainfall, and the rainfall is the accumulated rainfall in the primary rainfall process.
(2) And dividing early warning analysis units. The pipeline is divided into square early warning analysis units according to grid units with uniform size along the line area, namely the grid units are one by one, and the grid units are also used as the smallest geographic space units in artificial intelligence training. And (4) processing the disaster-enriched environment and the inducing factors into grid files with consistent standards by using a GIS technology.
(3) FR is calculated and an FR grid map is drawn. Using geological disaster points and each disaster-tolerant environmental layer data obtained by pipeline geological disaster investigation, calculating the FR of each disaster-tolerant environmental factor by using an FR model (probability ratio model), wherein the FR comprises stratum FR, rock-soil body type FR, elevation FR, gradient FR, slope FR, plane curvature FR, water system FR, fault FR and annual average rainfall FR, and obtaining the total FR of the disaster-tolerant environment. And drawing a corresponding FR grid graph by using a GIS technology.
The FR calculation formula is as follows:
FR is the ratio of the factors of j to i, and Np (LXi) is the number of geological disaster evaluation units of factor variable X in class i. Np (Xj) is the evaluation unit data of the factor variable Xj, and m is the number of secondary classification factors of the factor variable Xi. n represents the number of factors.
(4) And manufacturing a pipeline geological disaster susceptible partition. The pipeline geological disaster easy-occurrence partition map is drawn by geological disaster investigation technicians, the pipeline geological disaster easy-occurrence partition map is divided into pipeline geological disaster easy-occurrence partitions and drawn according to the principles of similarity and difference between areas in the areas and main induction conditions by the pipeline geological disaster investigation technicians according to the geological environment background, the geological disaster development characteristics, the distribution rules and the main induction conditions along the pipeline, namely, the geological environment background conditions, the main induction conditions and the geological disaster development characteristics are basically similar in the areas of the same type, but the pipeline geological disaster easy-occurrence partition map has obvious differences in the areas of different types.
(5) And counting and drawing a cumulative rainfall graph of a rainfall process which is up to the occurrence date of each geological disaster point. Each geological disaster point corresponds to one accumulated rainfall map, and n geological disaster points correspond to n accumulated rainfall maps. The accumulated rainfall in one rainfall process is stopped until the current day of disaster occurrence, and the one rainfall process is not continuously interrupted for more than two days.
(6) And overlapping the FR layer and the easily-generated partition layer to define a negative sample acquisition unit area. One negative sample acquisition unit comprises a plurality of early warning analysis units, and negative samples are respectively sampled in the negative sample acquisition units.
(7) And counting the total FR value range of each negative sample acquisition unit, counting the total FR value of the early warning analysis unit where each geological disaster point is located, and obtaining the average Mi of the total FR value of the early warning analysis unit where each geological disaster point is located and the minimum total FR value of the negative sample acquisition units.
(8) The spatial and temporal properties of the negative sample are acquired within the negative sample acquisition unit. The geological disaster points that have occurred are called positive sample points, the positive samples have spatial and temporal properties, and the negative sample points have spatial and temporal properties as well. Two negative sample points are acquired corresponding to each geological disaster point, one negative sample point is sampled at the position where the total FR value is equal to Mi and the rainfall is the same, and the other negative sample point is sampled at the position where the total FR value is the same but the rainfall is less. The time attribute of the two negative sample points is equivalent to the date on which the point of the geologic hazard occurred.
(9) A negative set of samples is constructed. Through negative sample point sampling, negative sample points which are 2 times that of positive samples are formed, and each negative sample point is provided with a space attribute and a time attribute. And extracting the data of the disaster-enriched environmental factors (including stratum, rock-soil body type, elevation, gradient, slope direction, plane curvature, water system, fault and annual average rainfall) of the corresponding space and geographic position and the data of the induction factors (the induction factors are the accumulated rainfall in one rainfall process) of the corresponding time of each negative sample point to form a negative sample set. The negative sample set and the positive sample set together form a sample set of the pipeline intelligent geological disaster area early warning model.
Preferably: the negative sample sampling method is based on two major types of data of a disaster-pregnant environment and an induction factor, a GIS technology is used for dividing an early warning analysis unit, an FR model is used for calculating and drawing an FR grid chart of the disaster-pregnant environment by the GIS technology, a GIS technology is used for drawing an accumulated rainfall chart of a rainfall process of each geological disaster point occurrence date, and each geological disaster point corresponds to one accumulated rainfall chart.
Preferably: based on FR model and easy-occurrence partition constraint, the region and boundary of negative sample sampling are determined, the concept of a negative sample acquisition unit is defined, the sampling method of the negative sample at the specific position of the negative sample acquisition unit region is determined, and the acquisition boundary and the acquisition position of the negative sample are defined.
Preferably: two negative sample points are acquired corresponding to each geological disaster point, one negative sample point is sampled at the position where the total FR value is equal to Mi and the rainfall is the same, the other negative sample point is sampled at the position where the total FR value of the geological disaster point is similar but the rainfall is less, the time attribute of the two negative sample points is equal to the occurrence date of the geological disaster point, and the acquired negative sample points are provided with the space attribute and the time attribute.
Preferably: and extracting the disaster-pregnant environment factor data of the corresponding space geographic position and the induction factor data of the corresponding time of each negative sample point, extracting the disaster-pregnant environment and the induction factor data by using the positions of the negative sample points, and giving the stratum, the rock-soil body type, the elevation, the gradient, the slope direction, the plane curvature, the water system, the fault and the annual average rainfall value and the accumulated rainfall value of the rainfall process to each negative sample point to form a complete negative sample set with the time, the space position, the induction factor and the disaster-pregnant environment value.
Preferably: the negative sample collection method fully considers the influence of the pregnancy disaster environment and the induction factors on the geological disaster, avoids collecting potential geological disaster hidden danger points, simultaneously avoids collecting extreme negative sample points, simultaneously avoids the spatial distribution randomness of the negative samples, and defines the sampling position and the sampling proportion, so that the negative samples have higher effectiveness and high representativeness, the collection quality of the negative samples is improved, the negative sample set and the positive sample set jointly form a sample set of the pipeline intelligent geological disaster area early warning model, and the accuracy rate and the leakage rate of the intelligent pipeline geological disaster area early warning model are improved.
The intelligent early warning negative sample sampling method overcomes the defect that undiscovered geological disaster hidden trouble points are selected as negative sample points by randomly selecting the negative sample, and also overcomes the defect that when only FR values are used, the point with a lower FR value in space position is selected as the negative sample point to cause the polar end of the negative sample point to be typical and the defect of regional representativeness is overcome. The intelligent early warning negative sample sampling method for the pipeline geological disasters based on the FR model and the easily-generated partition constraint can avoid selecting potential geological disaster hidden trouble points, avoid extreme negative sample points, and has good non-geological disaster area representativeness. The invention solves the problem of the sampling proportion of the negative sample and the positive sample, and the negative sample and the positive sample acquired by the method form a proportion relation of 2:1.
Drawings
FIG. 1 is a flow chart of an intelligent early warning negative sample sampling method for pipeline geological disasters under the constraint of an FR model and an easily-generated partition;
FIG. 2 is a FR grid view of a disaster recovery environment in an embodiment of the invention;
FIG. 3 is a plot of pipeline geological disaster vulnerability in an embodiment of the present invention;
FIG. 4 is a negative sample acquisition unit area divided along a pipeline in an embodiment of the invention;
FIG. 5 is a graph of the cumulative rainfall for one rainfall process corresponding to one of the geological disaster points in an embodiment of the present invention;
FIG. 6 is a negative sample of one of the geologic hazard points, in accordance with an embodiment of the invention;
DETAILED DESCRIPTION OF EMBODIMENT (S) OF INVENTION
The present invention is further described below in conjunction with embodiments, which are merely some, but not all embodiments of the present invention. Based on the embodiments of the present invention, those of ordinary skill in the art may obtain other embodiments without making any inventive effort, which fall within the scope of the present invention.
The flow of the method of this embodiment is shown in fig. 1. The main flow of the intelligent early warning negative sample sampling method for the pipeline geological disasters under the constraint of the FR model and the easily-occurring partition is as follows: firstly, confirming disaster-enriched environmental factors and induction factors required by geological disaster areas along the pipeline, and processing the factors into standard uniform grid files according to unified early warning analysis units; then, calculating FR by using an FR model, drawing an FR graph, drawing an easily-occurring partition graph according to the pipeline geological disaster investigation result, and drawing an accumulated rainfall graph of a primary rainfall process of each geological disaster point occurrence date; then, overlapping an FR layer and an easily-generated partition layer, defining a negative sample acquisition unit area, counting the total FR value range of each negative sample acquisition unit, and counting the total FR value of an early warning analysis unit where each geological disaster point is; and finally, collecting the spatial attribute and the time attribute of the negative sample in the negative sample collecting unit, extracting the disaster-enriched environmental factor data corresponding to the spatial geographic position and the induction factor data corresponding to the time, and forming a negative sample set of the pipeline intelligent geological disaster area early warning model.
Factors which play a main role in pregnancy and disaster along the pipeline geological disasters in the embodiment are as follows: stratum, rock-soil mass type, elevation, gradient, slope direction, plane curvature, water system, fault and annual average rainfall. Dividing the pipeline into early warning analysis units with the size of 500 multiplied by 500m along the line, and dividing the disaster-tolerant environmental factors: stratum, rock-soil mass type, elevation, gradient, slope direction, plane curvature, water system, fault and annual average rainfall are drawn into standard uniform grid files with the size of 500 multiplied by 500m by using a GIS technology. And (3) calculating the FR of each disaster-tolerant environmental factor by using an FR model (probability ratio model), wherein the FR comprises stratum FR, rock-soil body type FR, elevation FR, gradient FR, slope direction FR, plane curvature FR, water system FR, fault FR and annual average rainfall FR, and obtaining the total FR of the disaster-tolerant environment. And drawing a corresponding FR grid graph by using a GIS technology.
The FR calculation formula is as follows:
FR is the ratio of the factors of j to i, and Np (LXi) is the number of geological disaster evaluation units of factor variable X in class i. Np (Xj) is the evaluation unit data of the factor variable Xj, and m is the number of secondary classification factors of the factor variable Xi. n represents the number of factors.
The FR value is high to represent the high correlation between the pregnant disaster environment and the geological disaster, which is beneficial to the occurrence of the geological disaster, and the FR value is low to represent the low correlation between the pregnant disaster environment and the geological disaster, which is not beneficial to the occurrence of the geological disaster. When a geological disaster occurs, the geological disaster has a great relationship with not only the disaster-pregnant environment but also the induction factors of the period. Drawing an accumulated rainfall graph of a primary rainfall process of an occurrence time period of each historical occurrence geological disaster point, wherein rainfall is a main inducing factor of pipeline geological disasters. The accumulated rainfall in the primary rainfall process is stopped until the day of disaster occurrence, and the primary rainfall process is not continuously interrupted for more than two days. The occurrence time of each geological disaster point is different, and the corresponding rainfall is different, so that each geological disaster point corresponds to one accumulated rainfall map, and n accumulated rainfall maps are drawn for n geological disaster points.
The long oil and gas pipeline is long in conveying distance, long in pipeline line, multiple in geological disasters along the pipeline and large in geological environment difference along the pipeline, professional technicians and specialists are required to conduct pipeline along geological disaster investigation, and according to geological disaster investigation results, the technicians divide pipeline geological disaster easily-occurring subareas according to the geological environment background, geological disaster development characteristics, distribution rules and main induction conditions along the pipeline and the principles of similarity in areas and difference between areas, and draw pipeline geological disaster easily-occurring subareas. The easily-distributed areas are divided into a high easily-distributed area, a medium easily-distributed area, a low easily-distributed area and a non easily-distributed area, and the easily-distributed areas of different levels are generally distributed in a staggered mode, so that pipelines are divided into a plurality of easily-distributed areas along the line, and geological environments in the easily-distributed areas of each geological disaster are similar. The high-susceptibility partitions are not all subjected to geological disasters in a certain future time period, and the low-susceptibility partitions are not subjected to geological disasters in a certain future time period, so that it is necessary to characterize the differences in the same susceptibility partitions. The geological environments in the same geological disaster-prone zone are similar, but not identical, so that FR (field-effect transistor) images are overlapped to reflect the refined difference of the disaster-prone environments in each zone. The negative sample cannot be sampled only in the low-incidence area, and cannot be sampled only in the high-incidence area.
After the total FR diagram and the easily-developed partitions of the disaster-tolerant environment are overlapped, the pipeline is divided into a plurality of negative sample acquisition units along the line, the boundary of each negative sample acquisition unit is used as the boundary of negative sample sampling, and independent sampling is carried out in each negative sample acquisition unit. The negative sample avoids selecting potential geological disaster hidden trouble points as far as possible, has regional representativeness as far as possible, and avoids extreme negative sample points. The extreme negative sample points are extremely dissimilar to the geological disaster points, can only represent specific non-geological disaster areas, and are not regional representative. If only a batch of minimum FR values are selected during sampling, negative sample points are concentrated in a specific non-geological disaster area, so that the area representativeness is lost, all the total FR value range intervals and the total FR values corresponding to geological disaster points need to be counted, and after analysis, the FR values acquired by a proper negative sample are selected. The FR is too small to lose the region representativeness, the correlation between the FR and the negative sample is weakened, and the average value between the FR value corresponding to the geological disaster point and the lower boundary of the FR value of the negative sample acquisition unit region where the geological disaster point is located is selected as one basis for negative sample acquisition after analysis. In addition, the negative sample collection process also needs to represent the influence of the difference of the evoked factors on the geological disasters, and the FR value is similar to the FR value corresponding to the geological disaster point, but the evoked factors have larger difference, so that the difference is used as another basis for negative sample collection.
And counting the total FR value range interval in each negative sample acquisition unit region, counting the total FR value corresponding to the early warning analysis unit where each geological disaster point is located, and calculating the average of the total FR value corresponding to each geological disaster point and the minimum value of the negative sample acquisition unit region where the geological disaster point is located. For example, 7 months and 7 days in 2016, landslide geological disasters occur in three groups of villages under the east city office of Lichuan City, miao nationality, hubei province, and the total FR value corresponding to the landslide disasters is 5.44, and the total FR range of the negative sample acquisition unit area is: the average number of disasters and the minimum value is 4.165, and the central point of an early warning analysis unit with total FR near 4.165 is collected in a negative sample collection unit area where the disasters are located as one negative sample point, and the negative sample point is kept consistent with the accumulated rainfall in the current rainfall process which induces the landslide in 7 months of 2016 during collection; meanwhile, the total FR value is collected in the same negative sample collecting unit area and is near 5.44, but the rainfall is obviously less than the position of accumulated rainfall in the current rainfall process which induces the landslide to occur, and another negative sample point is collected; the spatial location and temporal properties of the negative samples are determined by sampling. The spatial position attribute and the time attribute of the negative sample points corresponding to all the geological disaster points are sampled by using the method, and the ratio of the positive sample points to the negative sample points after sampling is 1:2. And finally, extracting data of a disaster-tolerant environment and an induction factor by using the positions of the negative sample points, and endowing stratum, rock-soil mass type, elevation, gradient, slope direction, plane curvature, water system, fault and annual average rainfall values and accumulated rainfall values in the rainfall process to each negative sample point to form a complete negative sample set with time, space position, induction factor and disaster-tolerant environment values.
Claims (5)
1. An intelligent early warning negative sample sampling method for pipeline geological disasters under the constraint of easily-occurring partitions is characterized by comprising the following steps:
(1) Selecting factors required by early warning of a pipeline geological disaster area, wherein the required factors comprise a disaster-tolerant environment and induction factors;
the disaster-tolerant environment comprises stratum, rock-soil body type, elevation, gradient, slope direction, plane curvature, water system, fault and annual average rainfall;
the inducing factors mainly refer to rainfall, and the rainfall is accumulated rainfall in a primary rainfall process;
(2) Dividing an early warning analysis unit, dividing a pipeline into square early warning analysis units one by one along a line area according to grid units with uniform size, and processing a disaster-enriched environment and an induction factor into standard uniform grid files by using a GIS technology as a minimum geographic space unit in artificial intelligent training;
(3) Calculating FR, drawing an FR grid chart, calculating FR of each disaster-pregnant environment factor by using geological disaster points obtained by pipeline geological disaster investigation and data of each disaster-pregnant environment layer, using an FR model to calculate FR of each disaster-pregnant environment factor, including stratum FR, rock-soil body type FR, elevation FR, gradient FR, slope direction FR, plane curvature FR, water system FR, fault FR and annual average rainfall FR, obtaining total FR of the disaster-pregnant environment, and drawing a corresponding FR grid chart by using a GIS technology;
wherein, the FR model is a probability ratio model;
the FR calculation formula is as follows:
FR is the ratio of the overnumbers of the j-factor i types, np (LXi) is the number of geological disaster evaluation units of the i-type of the factor variable X, np (Xj) is the evaluation unit data of the factor variable Xj, m is the number of secondary classification factors of the factor variable Xi, and n represents the number of factors;
(4) Making pipeline geological disaster easily-occurring subareas, drawing a pipeline geological disaster easily-occurring subarea map by geological disaster investigation technicians, dividing the pipeline geological disaster easily-occurring subareas according to geological environment backgrounds, geological disaster development characteristics, distribution rules and main induction conditions along the pipeline, and drawing the pipeline geological disaster easily-occurring subarea map according to the principle of similarity in the subareas and difference between the subareas, namely that the geological environment backgrounds, the main induction conditions and the geological disaster development characteristics are basically similar in the same type of subareas, but have obvious differences in the different types of subareas;
(5) Counting and drawing an accumulated rainfall map of a primary rainfall process stopping at the occurrence date of each geological disaster point, wherein each geological disaster point corresponds to one accumulated rainfall map, n geological disaster points correspond to n accumulated rainfall maps, the accumulated rainfall of the primary rainfall process stopping at the occurrence date of the disaster, and the primary rainfall process refers to a rainfall process which is not continuously interrupted for more than two days;
(6) Overlapping the FR layer and the easily-partitioned layer to define a negative sample acquisition unit area, wherein one negative sample acquisition unit comprises a plurality of early warning analysis units, and the negative samples are respectively sampled in each negative sample acquisition unit area;
(7) Counting the total FR value range of each negative sample acquisition unit, counting the total FR value of the early warning analysis unit where each geological disaster point is located, and obtaining the average Mi of the total FR value of the early warning analysis unit where each geological disaster point is located and the minimum total FR value of the negative sample acquisition units;
(8) Collecting the spatial attribute and the time attribute of the negative sample in a negative sample collecting unit; the occurred geological disaster points are called positive sample points, the positive samples are provided with spatial attributes and time attributes, the negative sample points are also provided with the spatial attributes and the time attributes, two negative sample points are acquired corresponding to each geological disaster point, one negative sample point is sampled at the position where the total FR value is equal to Mi and the rainfall is the same, the other negative sample point is sampled at the position where the total FR value is the same but the rainfall is less, and the time attributes of the two negative sample points are equal to the date when the geological disaster point occurs;
(9) The method comprises the steps of constructing a negative sample set, sampling by negative sample points to form negative sample points which are 2 times of positive samples, wherein each negative sample point is provided with a spatial attribute and a time attribute, each negative sample point extracts disaster-inducing environmental factor data of a corresponding spatial geographic position and extracts induction factor data of corresponding time of the corresponding negative sample point to form the negative sample set, the negative sample set and the positive sample set jointly form a sample set of an intelligent pipeline geological disaster area early warning model, and the disaster-inducing environmental factors comprise stratum, rock-soil body type, elevation, gradient, slope direction, plane curvature, water system, fault and annual average rainfall, and the induction factors are accumulated rainfall in one rainfall process.
2. The intelligent pipeline geological disaster early warning negative sampling method under the constraint of the easily-distributed areas according to claim 1 is characterized by comprising the following steps: the negative sample sampling method is based on two major types of data of a disaster-pregnant environment and an induction factor, a GIS technology is used for dividing an early warning analysis unit, an FR model is used for calculating and drawing an FR grid chart of the disaster-pregnant environment by the GIS technology, a GIS technology is used for drawing an accumulated rainfall chart of a rainfall process of each geological disaster point occurrence date, and each geological disaster point corresponds to one accumulated rainfall chart.
3. The intelligent pipeline geological disaster early warning negative sampling method under the constraint of the easily-distributed areas according to claim 1 is characterized by comprising the following steps: based on FR model and easy-occurrence partition constraint, the region and boundary of negative sample sampling are determined, the concept of a negative sample acquisition unit is defined, the sampling method of the negative sample at the specific position of the negative sample acquisition unit region is determined, and the acquisition boundary and the acquisition position of the negative sample are defined.
4. The intelligent pipeline geological disaster early warning negative sampling method under the constraint of the easily-distributed areas according to claim 1 is characterized by comprising the following steps: two negative sample points are acquired corresponding to each geological disaster point, one negative sample point is sampled at the position where the total FR value is equal to Mi and the rainfall is the same, the other negative sample point is sampled at the position where the total FR value of the geological disaster point is similar but the rainfall is less, the time attribute of the two negative sample points is equal to the occurrence date of the geological disaster point, and the acquired negative sample points are provided with the space attribute and the time attribute.
5. The intelligent pipeline geological disaster early warning negative sampling method under the constraint of the easily-distributed areas according to claim 1 is characterized by comprising the following steps: and extracting the disaster-pregnant environment factor data of the corresponding space geographic position and the induction factor data of the corresponding time of each negative sample point, extracting the disaster-pregnant environment and the induction factor data by using the positions of the negative sample points, and giving the stratum, the rock-soil body type, the elevation, the gradient, the slope direction, the plane curvature, the water system, the fault and the annual average rainfall value and the accumulated rainfall value of the rainfall process to each negative sample point to form a complete negative sample set with the time, the space position, the induction factor and the disaster-pregnant environment value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211298958.4A CN115830804B (en) | 2022-10-24 | 2022-10-24 | Pipeline geological disaster intelligent early warning negative sample sampling method under constraint of easily-generated partition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211298958.4A CN115830804B (en) | 2022-10-24 | 2022-10-24 | Pipeline geological disaster intelligent early warning negative sample sampling method under constraint of easily-generated partition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115830804A CN115830804A (en) | 2023-03-21 |
| CN115830804B true CN115830804B (en) | 2023-08-22 |
Family
ID=85525225
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211298958.4A Active CN115830804B (en) | 2022-10-24 | 2022-10-24 | Pipeline geological disaster intelligent early warning negative sample sampling method under constraint of easily-generated partition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115830804B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101789120B1 (en) * | 2016-06-23 | 2017-11-15 | (주)해동기술개발공사 | Land Disaster Information System and Method by Parcel using Mobile GIS |
| CN112001435A (en) * | 2020-08-18 | 2020-11-27 | 中国地质环境监测院 | Method and system for constructing training sample set in regional landslide early warning and storage medium |
| CN112735097A (en) * | 2020-12-29 | 2021-04-30 | 中国地质环境监测院 | Regional landslide early warning method and system |
| CN112966722A (en) * | 2021-02-07 | 2021-06-15 | 南昌大学 | Regional landslide susceptibility prediction method based on semi-supervised random forest model |
| CN112991688A (en) * | 2021-03-03 | 2021-06-18 | 中山大学 | Geological disaster space-time combined early warning method and system based on PUL algorithm |
| CN113822522A (en) * | 2021-06-23 | 2021-12-21 | 中国科学院空天信息创新研究院 | Landslide susceptibility assessment method, device and equipment and readable storage medium |
| CN115100464A (en) * | 2022-06-21 | 2022-09-23 | 长安大学 | Landslide susceptibility assessment method and tool based on support vector machine |
-
2022
- 2022-10-24 CN CN202211298958.4A patent/CN115830804B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101789120B1 (en) * | 2016-06-23 | 2017-11-15 | (주)해동기술개발공사 | Land Disaster Information System and Method by Parcel using Mobile GIS |
| CN112001435A (en) * | 2020-08-18 | 2020-11-27 | 中国地质环境监测院 | Method and system for constructing training sample set in regional landslide early warning and storage medium |
| CN112735097A (en) * | 2020-12-29 | 2021-04-30 | 中国地质环境监测院 | Regional landslide early warning method and system |
| CN112966722A (en) * | 2021-02-07 | 2021-06-15 | 南昌大学 | Regional landslide susceptibility prediction method based on semi-supervised random forest model |
| CN112991688A (en) * | 2021-03-03 | 2021-06-18 | 中山大学 | Geological disaster space-time combined early warning method and system based on PUL algorithm |
| CN113822522A (en) * | 2021-06-23 | 2021-12-21 | 中国科学院空天信息创新研究院 | Landslide susceptibility assessment method, device and equipment and readable storage medium |
| CN115100464A (en) * | 2022-06-21 | 2022-09-23 | 长安大学 | Landslide susceptibility assessment method and tool based on support vector machine |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115830804A (en) | 2023-03-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Mwangi et al. | Forecasting droughts in East Africa | |
| Gudmundsson et al. | Observation-based gridded runoff estimates for Europe (E-RUN version 1.1) | |
| CN108875806B (en) | False forest fire hot spot mining method based on space-time data | |
| CN103491351A (en) | Intelligent video monitoring method for illegal buildings | |
| Skok et al. | Analysis of tropical cyclone precipitation using an object-based algorithm | |
| CN106651004B (en) | Flood disaster prediction method based on rainfall flood disaster time-space database | |
| CN110134907B (en) | Rainfall missing data filling method and system and electronic equipment | |
| CN107730496A (en) | A kind of multi-temporal remote sensing image building change detecting method based on image blocks | |
| CN113902580B (en) | Historical farmland distribution reconstruction method based on random forest model | |
| CN111144637A (en) | Regional power grid geological disaster forecasting model construction method based on machine learning | |
| Shi et al. | Development of quantitative methods for detecting climate contributions to boundary shifts in farming-pastoral ecotone of northern China | |
| Cui et al. | Bayesian optimization of typhoon full-track simulation on the Northwestern Pacific segmented by QuadTree decomposition | |
| CN116842435A (en) | SWAP-based field drought three-dimensional measurement and time-space evolution visualization method | |
| Ma et al. | A multisource remotely sensed data oriented method for “ghost city” phenomenon identification | |
| CN110245692B (en) | Hierarchical clustering method for collecting numerical weather forecast members | |
| CN115830804B (en) | Pipeline geological disaster intelligent early warning negative sample sampling method under constraint of easily-generated partition | |
| Clark et al. | An automated, multiparameter dryline identification algorithm | |
| CN105046089A (en) | Method for predicting strong rainfall and flood disasters | |
| CN108182178A (en) | Groundwater level analysis method and system based on event text data mining | |
| Lee et al. | Markov chain random field kriging for estimating extreme precipitation at unevenly distributed sites | |
| CN110208876B (en) | Characterization method for cooperative change of radial positions of subtropical zone torrent and polar front torrent | |
| Chernokulsky et al. | Objectively and manually identified characteristics of mid-latitude storms: a comparison for Siberian region | |
| CN113742929A (en) | Data quality evaluation method for grid weather live | |
| Dastorani et al. | Analysis of the trend of dust changes in Ardestan region, Iran | |
| Jacintho et al. | Brazilian presidential elections: Analysing voting patterns in time and space using a simple data science pipeline |
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 |