CN110532969A - Ramp unit division methods based on multi-scale image segmentation - Google Patents
Ramp unit division methods based on multi-scale image segmentation Download PDFInfo
- Publication number
- CN110532969A CN110532969A CN201910823312.5A CN201910823312A CN110532969A CN 110532969 A CN110532969 A CN 110532969A CN 201910823312 A CN201910823312 A CN 201910823312A CN 110532969 A CN110532969 A CN 110532969A
- Authority
- CN
- China
- Prior art keywords
- unit
- layer
- value
- ramp
- slope aspect
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0637—Strategic management or analysis, e.g. setting a goal or target of an organisation; Planning actions based on goals; Analysis or evaluation of effectiveness of goals
- G06Q10/06375—Prediction of business process outcome or impact based on a proposed change
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
Abstract
The invention discloses a kind of ramp unit division methods based on multi-scale image segmentation, including obtain slope aspect figure layer based on digital elevation model;Be calculated the unit vector of X-direction and the unit vector of Y direction and using as divide ramp unit main body divide figure layer;The limitation figure layer extracting gathering ground figure layer and being divided using it as ramp unit;Ramp unit is divided to obtain final ramp unit division result.The present invention establishes data set, best segmental scale and the partitioning algorithm for being suitble to multi-scale division, generates suitable ramp unit by comprehensively considering topographic(al) feature;Therefore the method for the present invention can be evaluated for geological disaster risk provides reliable map unit support, and the high reliablity of ramp unit division methods of the invention, the dividing precision of ramp unit are higher.
Description
Technical field
Present invention relates particularly to a kind of ramp unit division methods based on multi-scale image segmentation.
Background technique
Regional geology Hazard Risk Assessment can long-range prediction geological disaster occur a possibility that, be national government department
Decision of preventing and reducing natural disasters provides reference, to reducing people life property loss and promoting social harmony to develop there is important reality to anticipate
Justice.And the committed step of regional geology Hazard Risk Assessment is the determination of map unit, reasonably selects and divide map unit
The quality of risk evaluation results will be directly affected.
Most common Evaluation of Geologic Hazards unit includes grid cell and ramp unit.Wherein: the division of grid cell is straight
It connects and survey region is divided into equal-sized square net, regular shape, subdivision is simple and easy, the matrix obtained after discrete
The data of form are also beneficial to further operation, but grid cell is not easy to present the spatial correlation between landforms, also reflects
Do not go out geomorphic feature;The division of ramp unit is then that will close on similar landform to divide same unit into, and unit is by boundary condition and micro-
The influence of topography can be considered a nearly homogeneous geomorphic unit, be the basic unit of the Development of Geological Hazards such as landslide, avalanche, Ke Yiyu
Geological conditions are closely connected, comprehensive to embody all kinds of controls or the effect of influence factor, make evaluation result closer in reality.
Traditional ramp unit division methods are often surveyed and drawn with field investigation or using three-dimensional aviation pictures pair, oblique by manual identified
Slope unit is simultaneously classified, but this method treatment process is many and diverse and time-consuming, therefore some scholars attempt to utilize digital elevation mould
Type exploitation automation landform sort program.Automation divides time saving and energy saving and save the cost, can effectively reduce human error, avoid
The different limitation of the experience and subjective criterion of different operation person itself, it can be ensured that the consistency of division result, than manually drawing
Dividing has apparent advantage.Currently, what self-acting slope dividing elements method was most widely used is traditional gathering ground overlay method, i.e.,
Hydrological analysis modules A rcHydro based on ArcGIS platform carries out a watershed analysis first, obtains valley route;It will count again
After being worth the high low value reversion of elevation model, second of watershed analysis is carried out, then raw water system can be reversed to ridge line;Then by mountain
Valley line crosses with ridge line, former gathering ground can be cut into two ramp units in left and right.But this method partition process is more numerous
Trivial, the determining subjective dependency degree of catchment area threshold values is high in partition process, lacks unified standard, and the gathering ground unit generated holds
Easily there is the more tiny plane of disruption and unreasonable strip face, the later period needs to carry out a large amount of manual modifications.
Summary of the invention
The purpose of the present invention is to provide a kind of high reliablity and the higher slopes based on multi-scale image segmentation of precision
Dividing elements method.
This ramp unit division methods based on multi-scale image segmentation provided by the invention, include the following steps:
S1. slope aspect figure layer is obtained based on digital elevation model;
S2. the unit vector of X-direction and the list of Y direction is calculated according to the slope aspect figure layer data that step S1 is obtained
Bit vector, and using the unit vector of X-direction and the unit vector of Y direction as the main body segmentation figure of division ramp unit
Layer;
S3. gathering ground figure layer is extracted, and the gathering ground figure layer of extraction is used as the limitation figure layer that ramp unit divides;
S4. the limitation figure layer that the main body segmentation figure layer and step S3 obtained according to step S2 obtains, carries out ramp unit
It divides, to obtain final ramp unit division result.
The ramp unit division methods based on multi-scale image segmentation, further include following steps:
S5. segmentation precision evaluation function is constructed using local variance and Moran index, and to the slope of the method for the present invention list
First division result is evaluated.
Slope aspect figure layer is obtained based on digital elevation model described in step S1, is specially based on digital elevation model, is passed through
ArcGIS pre-treatment obtains slope aspect figure layer, as the data set for dividing ramp unit automatically.
The unit vector and Y of X-direction is calculated described in step S2 according to the slope aspect figure layer data that step S1 is obtained
The unit vector of axis direction specially converts Circular measure slope aspect data for the slope aspect figure layer data that step S1 is obtained, then into
Row trigonometric function calculates, to obtain the unit vector of X-direction and the unit vector of Y direction.
The unit vector of the X-direction and the unit vector of Y direction are specially calculated using following steps
It arrives:
A. slope aspect figure layer data are converted to by radian data θ using following formula:
α is slope aspect figure layer data in formula;
B. trigonometric function calculating is carried out to the data that step A is obtained using following formula, to obtain the unit of X-direction
VectorWith the unit vector of Y direction
θ is the radian data that step A is obtained in formula.
Extraction gathering ground figure layer described in step S3 specially extracts gathering ground figure layer using following steps:
A. depression filling is carried out to dem data, to avoid generating flow direction exception;
B. flow direction analysis is carried out, the received integrated flow of grid flow direction and each grid institute is calculated;
C. it using practical water system as reference, determines best integrated flow threshold value, obtains integrated flow and be higher than threshold value
Water system;
D. rivers and creeks segmentation is carried out, water system is divided into different sections, and carry out gathering ground calculating, to obtain final collection
Pool figure layer.
Ramp unit is divided described in step S4, is specially divided using following steps:
(1) initial segmentation scale (it is recommended that one lesser initial segmentation size of setting) is set, and uses following formula meter
Calculate the standard deviation in neighborhood corresponding to the initial segmentation scale of the setting:
σ in formulaiThe standard deviation in neighborhood being sized for i-th;N is the pixel that the neighborhood being sized is included
Number;ciFor the gray value of i-th of pixel;The mean value of pixel gray value in the neighborhood being sized for i-th;
(2) mean value of the standard deviation in addition to edge in each neighborhood being sized is calculated using following formula, and this is
Value is LV (Local-variance) value of object layer:
M is the number that the neighborhood being sized calculated is participated in entire destination layer in formula;
(3) the initial segmentation scale that will be set in step (1) is amplified by the zooming parameter of setting, calculates different neighborhoods
LV value under size;And the dividing layer where different size of neighborhood, as the different target layer;
(4) local variance change rate ROC-LV (ROC-LV, the rates of of different target interlayer is calculated using following formula
Change of LV):
L is the LV value of destination layer in formula, and L-1 is the LV value of next destination layer;
(5) the local variance change rate obtained using step (4) measures the LV from a destination layer to another destination layer
Variable quantity chooses segmentation scale corresponding to the LV variable quantity as best segmental scale, i.e. ROC- when LV variable quantity maximum
Segmentation scale corresponding to the peak point of LV;
(6) use control variate method, be determined by experiment multi-scale division needs other two parameter: form factor,
The compact degree factor;
(7) divide figure layer using the unit vector of the X-direction of slope aspect figure layer and the unit vector of Y direction as main body,
Using gathering ground figure layer as orographic condition restricted boundary, using best segmental scale, form factor and compact degree factor conduct
Partitioning parameters carry out multi-scale division, to obtain final ramp unit division result.
Segmentation precision evaluation function is constructed using local variance and Moran index described in step S5, specially using such as
Lower formula is as segmentation precision evaluation function F (V, I):
V in formulamaxFor the maximum value of V;VminFor the minimum value of V;ImaxFor the maximum value of I;IminFor the minimum value of I;V is
It one intermediate parameters and is defined assnFor the area of n-th of unit, cnFor the slope aspect side inside n-th of unit
Difference andQ is the number of unit interior pel, piFor the corresponding slope aspect value of i-th of pixel,For
The slope aspect value mean value of unit interior pel;I is the second intermediate parameters and is defined asN is unit number, αnFor the slope aspect mean value in n-th of unit, αlFor l
Slope aspect mean value in a unit, ωn,lRelationship index, and the ω when n-th of unit is adjacent with first of unit are closed on for spacen,l
Value is 1, the ω when n-th of unit and non-conterminous first of unitn,lValue is 0,For the slope aspect mean value of figure layer.
This ramp unit division methods based on multi-scale image segmentation provided by the invention, by comprehensively considering landform
Element establishes data set, best segmental scale and the partitioning algorithm for being suitble to multi-scale division, generates suitable ramp unit;Cause
This method for the present invention can be evaluated for geological disaster risk provides reliable map unit support, and ramp unit of the invention
The dividing precision of the high reliablity of division methods, ramp unit is higher.
Detailed description of the invention
Fig. 1 is the method flow schematic diagram of the method for the present invention.
Fig. 2 is the slope aspect signal of the method for the present invention and is converted to unit vector schematic diagram.
Fig. 3 is that the best segmental scale of the method for the present invention calculates schematic diagram.
Fig. 4 is the ramp unit division result contrast schematic diagram of the method for the present invention and conventional method.
Fig. 5 is the accuracy comparison schematic diagram of the ramp unit division result of the method for the present invention and conventional method.
Specific embodiment
It is as shown in Figure 1 the method flow schematic diagram of the method for the present invention: provided by the invention this based on multi-scale image
The ramp unit division methods of segmentation, include the following steps:
S1. slope aspect figure layer is obtained based on digital elevation model;It is specially based on digital elevation model, before ArcGIS
Reason obtains slope aspect figure layer, for dividing the data set of ramp unit automatically;
S2. the unit vector of X-direction and the list of Y direction is calculated according to the slope aspect figure layer data that step S1 is obtained
Bit vector, and using the unit vector of X-direction and the unit vector of Y direction as the main body segmentation figure of division ramp unit
Layer;Circular measure slope aspect data specially are converted by the slope aspect figure layer data that step S1 is obtained, then carry out trigonometric function calculating,
To obtain the unit vector of X-direction and the unit vector of Y direction;
Due to slope aspect data α have period continuity, as shown in Figure 2: 0 ° with 359 ° representated by actual direction only have 1 degree
Difference, numerically but have 359 ° of difference.Number when carrying out block merging in order to avoid Image Segmentation, around 0 degree of slope aspect
Value difference is different excessive and leads to the problem of, therefore uses following steps unit of account vector:
A. slope aspect figure layer data are converted to by radian data θ using following formula:
α is slope aspect figure layer data in formula;
B. trigonometric function calculating is carried out to the data that step A is obtained using following formula, to obtain the unit of X-direction
VectorWith the unit vector of Y direction
θ is the radian data that step A is obtained in formula;
S3. gathering ground figure layer is extracted, and the gathering ground figure layer of extraction is used as the limitation figure layer that ramp unit divides;Specifically
To extract gathering ground figure layer using following steps:
A. depression filling is carried out to dem data, to avoid generating flow direction exception;
B. flow direction analysis is carried out, the received integrated flow of grid flow direction and each grid institute is calculated;
C. it using practical water system as reference, determines best integrated flow threshold value, obtains integrated flow and be higher than threshold value
Water system;
D. rivers and creeks segmentation is carried out, water system is divided into different sections, and carry out gathering ground calculating, to obtain final collection
Pool figure layer;
S4. the limitation figure layer that the main body segmentation figure layer and step S3 obtained according to step S2 obtains, carries out ramp unit
It divides, to obtain final ramp unit division result.Specially divided using following steps:
(1) initial segmentation scale (it is recommended that one lesser initial segmentation size of setting) is set, and uses following formula meter
Calculate the standard deviation in neighborhood corresponding to the initial segmentation scale of the setting:
σ in formulaiThe standard deviation in neighborhood being sized for i-th;N is the pixel that the neighborhood being sized is included
Number;ciFor the gray value of i-th of pixel;The mean value of pixel gray value in the neighborhood being sized for i-th;
(2) mean value of the standard deviation in addition to edge in each neighborhood being sized is calculated using following formula, and this is
Value is LV (Local-variance) value of object layer:
M is the number that the neighborhood being sized calculated is participated in entire destination layer in formula;
(3) the initial segmentation scale that will be set in step (1) is amplified by the zooming parameter of setting, calculates different neighborhoods
LV value under size;And the dividing layer where different size of neighborhood, as the different target layer;
(4) local variance change rate ROC-LV (ROC-LV, the rates of of different target interlayer is calculated using following formula
Change of LV):
L is the LV value of destination layer in formula, and L-1 is the LV value of next destination layer;
(5) the local variance change rate obtained using step (4) measures the LV from a destination layer to another destination layer
Variable quantity chooses segmentation scale corresponding to the LV variable quantity as best segmental scale, i.e. ROC- when LV variable quantity maximum
Segmentation scale corresponding to the peak point of LV;
(6) use control variate method, be determined by experiment multi-scale division needs other two parameter: form factor,
The compact degree factor;
(7) divide figure layer using the unit vector of the X-direction of slope aspect figure layer and the unit vector of Y direction as main body,
Using gathering ground figure layer as orographic condition restricted boundary, using best segmental scale, form factor and compact degree factor conduct
Partitioning parameters carry out multi-scale division, to obtain final ramp unit division result;
Fig. 4 display tradition divides the result of ramp unit and the Comparative result of the method for the present invention based on ArcHydro.It can be with
Find out, traditional ramp unit that area overlapping partitioning obtains that catchments will appear the parallel river valley for the conjunctions that do not conform to the actual conditions largely, and have
A little units cross over two hillside.Method segmentation result shown in the present invention, the case where not occurring parallel river valley and cross over hillside.
S5. segmentation precision evaluation function is constructed using local variance and Moran index, and to the slope of the method for the present invention list
First division result is evaluated;Specially using following formula as segmentation precision evaluation function F (V, I):
V in formulamaxFor the maximum value of V;VminFor the minimum value of V;ImaxFor the maximum value of I;IminFor the minimum value of I;V is
It one intermediate parameters and is defined assnFor the area of n-th of unit, cnFor the slope aspect side inside n-th of unit
Difference andQ is the number of unit interior pel, piFor the corresponding slope aspect value of i-th of pixel,For
The slope aspect value mean value of unit interior pel;I is the second intermediate parameters and is defined as
N is unit number, αnFor the slope aspect mean value in n-th of unit, αlFor the slope aspect mean value in first of unit, ωn,lIt is closed on for space
Relationship index, and the ω when n-th of unit is adjacent with first of unitn,lValue is 1, when n-th of unit and first of unit not phase
ω when adjacentn,lValue is 0,For the slope aspect mean value of figure layer;When by the unit vector of radian data θ and X-directionWith Y
The unit vector of axis directionThe calculating formula of I is substituted into, so that the parameter rewriting in the calculating formula of I is as follows:
F (V, I) value of each ramp unit is calculated according to above formula, and number then can be described according to calculated result drafting
According to the box figure of distribution characteristics, ramp unit dividing precision obtained by traditional gathering ground overlay method and the method for the present invention is carried out with this
Comparison.The evaluation function is according to the division principle of ramp unit, when heterogeneous maximum between homogeney highest, unit in unit,
Ramp unit division result is best.Using local variance as homogeney measurement index inside unit, the index in evaluation function
It is worth smaller, indicates that internal homogeney is higher;Using size heterogeneous between Moran measure unit, the index is smaller, indicates
Have between unit high heterogeneous.Therefore by F (V, I) Computing Principle it is found that F (V, I) value is bigger, show in unit that segmentation obtains
Portion's homogeney is higher, and heterogeneous higher between unit, i.e., segmentation precision is higher.
Example results are as shown in Fig. 5 median and cabinet data profile.As can be seen that multi-scale image split plot design divides
(V, the I) minimum value of F corresponding to ramp unit is 0.3466, median 1.3108, is all larger than traditional gathering ground overlay method and calculates
Corresponding minimum value 0.2174 and median 1.2312 divide for showing the relatively traditional gathering ground overlay method of the method for the present invention
As a result more preferable, segmentation precision is higher.
Claims (8)
1. a kind of ramp unit division methods based on multi-scale image segmentation, include the following steps:
S1. slope aspect figure layer is obtained based on digital elevation model;
S2. according to step S1 obtain slope aspect figure layer data be calculated X-direction unit vector and Y direction unit to
Amount, and divide figure layer using the unit vector of X-direction and the unit vector of Y direction as the main body for dividing ramp unit;
S3. gathering ground figure layer is extracted, and the gathering ground figure layer of extraction is used as the limitation figure layer that ramp unit divides;
S4. the limitation figure layer that the main body segmentation figure layer and step S3 obtained according to step S2 obtains, divides ramp unit,
To obtain final ramp unit division result.
2. the ramp unit division methods according to claim 1 based on multi-scale image segmentation, it is characterised in that also wrap
Include following steps:
S5. segmentation precision evaluation function is constructed using local variance and Moran index, and the ramp unit of the method for the present invention is drawn
Point result is evaluated.
3. the ramp unit division methods according to claim 1 based on multi-scale image segmentation, it is characterised in that step
Slope aspect figure layer is obtained based on digital elevation model described in S1, digital elevation model is specially based on, by ArcGIS pre-treatment
Slope aspect figure layer is obtained, for dividing the data set of ramp unit automatically.
4. the ramp unit division methods according to claim 1 based on multi-scale image segmentation, it is characterised in that step
The unit vector of X-direction and the unit of Y direction is calculated according to the slope aspect figure layer data that step S1 is obtained described in S2
Vector specially converts Circular measure slope aspect data for the slope aspect figure layer data that step S1 is obtained, and then carries out trigonometric function meter
It calculates, to obtain the unit vector of X-direction and the unit vector of Y direction.
5. the ramp unit division methods according to claim 4 based on multi-scale image segmentation, it is characterised in that described
The unit vector that X-direction is calculated and Y direction unit vector, unit specially is calculated using following steps
Vector:
A. slope aspect figure layer data are converted to by radian data θ using following formula:
α is slope aspect figure layer data in formula;
B. trigonometric function calculating is carried out to the data that step A is obtained using following formula, to obtain the unit vector of X-directionWith the unit vector of Y direction
θ is the radian data that step A is obtained in formula.
6. the ramp unit division methods according to claim 1 based on multi-scale image segmentation, it is characterised in that step
Extraction gathering ground figure layer described in S3 specially extracts gathering ground figure layer using following steps:
A. depression filling is carried out to dem data, to avoid generating flow direction exception;
B. flow direction analysis is carried out, the received integrated flow of grid flow direction and each grid institute is calculated;
C. it using practical water system as reference, determines best integrated flow threshold value, obtains the water system that integrated flow is higher than threshold value;
D. rivers and creeks segmentation is carried out, water system is divided into different sections, and carry out gathering ground calculating, to obtain final gathering ground
Figure layer.
7. the ramp unit division methods based on multi-scale image segmentation described according to claim 1~one of 6, feature exist
Ramp unit is divided described in the step S4, is specially divided using following steps:
(1) initial segmentation scale is set, and is calculated in neighborhood corresponding to the initial segmentation scale of the setting using following formula
Standard deviation:
σ in formulaiThe standard deviation in neighborhood being sized for i-th;N is the pixel number that the neighborhood being sized is included;ci
For the gray value of i-th of pixel;The mean value of pixel gray value in the neighborhood being sized for i-th;
(2) mean value of the standard deviation in addition to edge in each neighborhood being sized is calculated using following formula, and the mean value is
For LV (Local-variance) value of object layer:
M is the number that the neighborhood being sized calculated is participated in entire destination layer in formula;
(3) the initial segmentation scale that will be set in step (1) is amplified by the zooming parameter of setting, calculates different Size of Neighborhood
Under LV value;And the dividing layer where different size of neighborhood, as the different target layer;
(4) local variance change rate ROC-LV (ROC-LV, the rates of of different target interlayer is calculated using following formula
Change of LV):
L is the LV value of destination layer in formula, and L-1 is the LV value of next destination layer;
(5) the local variance change rate obtained using step (4) measures the LV variation from a destination layer to another destination layer
Amount chooses segmentation scale corresponding to the LV variable quantity as best segmental scale when LV variable quantity maximum, i.e. ROC-LV's
Segmentation scale corresponding to peak point;
(6) control variate method is used, is determined by experiment other two parameter of multi-scale division needs: form factor, compact
Spend the factor;
(7) divide figure layer using the unit vector of the X-direction of slope aspect figure layer and the unit vector of Y direction as main body, use
Gathering ground figure layer is as orographic condition restricted boundary, using best segmental scale, form factor and the compact degree factor as segmentation
Parameter carries out multi-scale division, to obtain final ramp unit division result.
8. the ramp unit division methods according to claim 2 based on multi-scale image segmentation, it is characterised in that step
Segmentation precision evaluation function is constructed using local variance and Moran index described in S5, is specially used as and is divided using following formula
Cut precision evaluation function F (V, I):
V in formulamaxFor the maximum value of V;VminFor the minimum value of V;ImaxFor the maximum value of I;IminFor the minimum value of I;V is in first
Between parameter and be defined assnFor the area of n-th of unit, cnFor the slope aspect variance inside n-th of unit andQ is the number of unit interior pel, piFor the corresponding slope aspect value of i-th of pixel,For unit
The slope aspect value mean value of interior pel;I is the second intermediate parameters and is defined asN
For unit number, αnFor the slope aspect mean value in n-th of unit, αlFor the slope aspect mean value in first of unit, ωn,lIt is closed on for space
Relationship index, and the ω when n-th of unit is adjacent with first of unitn,lValue is 1, when n-th of unit and first of unit not phase
ω when adjacentn,lValue is 0,For the slope aspect mean value of figure layer.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910823312.5A CN110532969B (en) | 2019-09-02 | 2019-09-02 | Slope unit dividing method based on multi-scale image segmentation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910823312.5A CN110532969B (en) | 2019-09-02 | 2019-09-02 | Slope unit dividing method based on multi-scale image segmentation |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110532969A true CN110532969A (en) | 2019-12-03 |
CN110532969B CN110532969B (en) | 2022-12-27 |
Family
ID=68666186
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910823312.5A Active CN110532969B (en) | 2019-09-02 | 2019-09-02 | Slope unit dividing method based on multi-scale image segmentation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110532969B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113850822A (en) * | 2021-09-18 | 2021-12-28 | 四川大学 | Automatic slope unit dividing method based on confluence segmentation |
CN114329663A (en) * | 2021-12-27 | 2022-04-12 | 中国自然资源航空物探遥感中心 | Slope unit dividing method based on scale of historical geological disasters |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103455711A (en) * | 2013-08-15 | 2013-12-18 | 广州地理研究所 | Small watershed region-oriented landslide hazard risk division method based on mechanism analysis |
CN104820826A (en) * | 2015-04-27 | 2015-08-05 | 重庆大学 | Digital elevation model-based slope form extraction and recognition method |
WO2016032398A2 (en) * | 2014-08-25 | 2016-03-03 | Singapore University Of Technology And Design | Method and device for analysing an image |
-
2019
- 2019-09-02 CN CN201910823312.5A patent/CN110532969B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103455711A (en) * | 2013-08-15 | 2013-12-18 | 广州地理研究所 | Small watershed region-oriented landslide hazard risk division method based on mechanism analysis |
WO2016032398A2 (en) * | 2014-08-25 | 2016-03-03 | Singapore University Of Technology And Design | Method and device for analysing an image |
CN104820826A (en) * | 2015-04-27 | 2015-08-05 | 重庆大学 | Digital elevation model-based slope form extraction and recognition method |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113850822A (en) * | 2021-09-18 | 2021-12-28 | 四川大学 | Automatic slope unit dividing method based on confluence segmentation |
CN113850822B (en) * | 2021-09-18 | 2023-04-25 | 四川大学 | Automatic slope unit dividing method based on confluence division |
CN114329663A (en) * | 2021-12-27 | 2022-04-12 | 中国自然资源航空物探遥感中心 | Slope unit dividing method based on scale of historical geological disasters |
CN114329663B (en) * | 2021-12-27 | 2022-08-05 | 中国自然资源航空物探遥感中心 | Slope unit dividing method based on scale of historical geological disasters |
Also Published As
Publication number | Publication date |
---|---|
CN110532969B (en) | 2022-12-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Niu et al. | Measuring urban poverty using multi-source data and a random forest algorithm: A case study in Guangzhou | |
CN103236086B (en) | One takes the contextual multiple dimensioned DEM modeling method of the earth's surface hydrology into account | |
CN103092572B (en) | The parallel method of distributing numerical control under a kind of cluster environment | |
CN101840581B (en) | Method for extracting profile of building from satellite remote sensing image | |
CN107220754B (en) | County scale mountain torrent disaster risk assessment method | |
CN104899562B (en) | Radar remote sensing image culture's recognizer based on Texture Segmentation fusion | |
CN106815850A (en) | The method that canopy density forest reserves very high is obtained based on laser radar technique | |
Qin et al. | Mapping soil organic matter in small low-relief catchments using fuzzy slope position information | |
Zhi et al. | Urban flood risk assessment and analysis with a 3D visualization method coupling the PP-PSO algorithm and building data | |
CN104751478A (en) | Object-oriented building change detection method based on multi-feature fusion | |
CN102122395A (en) | Adaptive scale DEM (digital elevation model) modeling method capable of keeping terrain features | |
CN107330422A (en) | A kind of method for carrying out mima type microrelief classification to semiarid zone based on high accuracy number elevation model | |
CN103679675A (en) | Remote sensing image fusion method oriented to water quality quantitative remote sensing application | |
CN110532969A (en) | Ramp unit division methods based on multi-scale image segmentation | |
CN105335965A (en) | High-resolution remote sensing image multi-scale self-adaptive decision fusion segmentation method | |
Zhu et al. | Loess terrain segmentation from digital elevation models based on the region growth method | |
CN116151610A (en) | Method for simulating risk exposure space of disaster-bearing body on underlying surface of heterogeneous city | |
Wang et al. | Mapping of regional soil salinities in Xinjiang and strategies for amelioration and management | |
Huang et al. | Modelling landslide susceptibility prediction: A review and construction of semi-supervised imbalanced theory | |
CN109190593A (en) | Slope stability principium identification method along mountain road based on concave-convex category division | |
CN103377477A (en) | Method and system for multilevel division of high-resolution remote sensing image | |
Dong et al. | Pixel-level intelligent segmentation and measurement method for pavement multiple damages based on mobile deep learning | |
Wang et al. | Urban road information extraction from high resolution remotely sensed image based on semantic model | |
Zhang et al. | Application and evaluation of object-oriented technology in high-resolution remote sensing image classification | |
CN113850822B (en) | Automatic slope unit dividing method based on confluence division |
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 |