CN104992428B - A kind of image falling rocks cubage measuring method based on K mean cluster analysis - Google Patents
A kind of image falling rocks cubage measuring method based on K mean cluster analysis Download PDFInfo
- Publication number
- CN104992428B CN104992428B CN201510195309.5A CN201510195309A CN104992428B CN 104992428 B CN104992428 B CN 104992428B CN 201510195309 A CN201510195309 A CN 201510195309A CN 104992428 B CN104992428 B CN 104992428B
- Authority
- CN
- China
- Prior art keywords
- falling rocks
- image
- frame
- target
- cluster
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30232—Surveillance
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of image falling rocks cubage measuring method based on K mean cluster analysis, including：Falling rocks video image is gathered by the video acquisition system built in advance, and falling rocks detection is carried out to every two field picture of the falling rocks video image of collection, it is determined that the falling rocks bianry image per frame after detection；According to the every frame falling rocks bianry image determined, the area and nominal altitude per corresponding to frame falling rocks bianry image is determined；According to the area and nominal altitude, the volume of calculating falling rocks target, and according to K mean cluster parser, the volume set to falling rocks target clusters；First number data of some clusters after cluster are counted, and according to statistical result, determine the most cluster of number, and according to the cluster, estimate the volume of falling rocks target, wherein, the method for estimation is to take the average of all first numbers in cluster.
Description
Technical field
The present invention relates to digital image processing techniques field, it particularly relates to a kind of figure based on K mean cluster analysis
As falling rocks cubage measuring method.
Background technology
Landslide early-warning system is all monitored using monitoring instrument at present, is traditionally used for the side of mountain landslide supervision
Method probably has following several：Conventional geodesic method；Liquid multi-point leveling measurement, gravimetric method；Water table measure method；Electricity
Survey method, earth drilling inclination etc..Although these methods all serve positive role in terms of landslide early-warning, they have
Some drawbacks, such as by the influence of topography, it is impossible to which Continuous Observation, automaticity is not high, and human input is excessive, and data can not be located in real time
Reason etc..
Above-mentioned part drawback can be then eliminated using the monitoring method based on digital image processing techniques, as non-contact type is surveyed
Amount, intelligence degree is high, and cost is low, can be with real-time early warning etc..In the video frequency monitoring system using falling rocks as warning information, prison
While survey falling rocks whether there is, it is often desirable to provide the specific measurement parameter of falling rocks, such as volume.Varying number and it is different size of fall
Stone characterizes different degrees of early warning, or even minimum falling rocks may be false-alarm.Therefore, while monitoring falling rocks, phase is provided
The cubing parameter meaning answered and its important.
Volume calculation techniques based on image processing algorithm are shot from different perspectives using multiple cameras, are increased into
This, and realize difficult；And the technology based on one camera is largely then the model library for pre-establishing target component to be measured, Ran Houtong
Cross algorithms of different to be matched and be fitted, be not suitable for the monitoring of falling rocks.
The problem of in correlation technique, effective solution is not yet proposed at present.
The content of the invention
The problem of in correlation technique, the present invention propose that a kind of image falling rocks volume based on K mean cluster analysis is surveyed
Determine method, its consideration based on falling rocks random rolling in motion process, utilize the area and nominal altitude of different images sequence
Parameter establishes the set of falling rocks target difference posture lower body volume data, and sub-clustering, number are carried out to set using K mean cluster algorithm
The maximum cluster of amount is the random distribution of true volume, thus estimates the volume of falling rocks.
The technical proposal of the invention is realized in this way：
A kind of image falling rocks cubage measuring method based on K mean cluster analysis, including：
Falling rocks video image is gathered by the video acquisition system built in advance, and to the every of the falling rocks video image of collection
Two field picture carries out falling rocks detection, it is determined that the falling rocks bianry image per frame after detection；
According to the every frame falling rocks bianry image determined, the area and name per corresponding to frame falling rocks bianry image is determined
Adopted height；
According to the area and nominal altitude, the volume of falling rocks target is calculated, and according to K mean cluster parser, it is right
The volume set of falling rocks target is clustered；
The element number of some clusters after cluster is counted, and according to statistical result, determines the most cluster of number,
And according to the cluster, the volume of falling rocks target is estimated, wherein, the method for estimation is to take the average of all elements in cluster.
Further, the above-mentioned image falling rocks cubage measuring method based on K mean cluster analysis also includes：
Before the video acquisition system collection falling rocks video image by building in advance, the video acquisition system is entered
Row camera calibration.
Further, before the video acquisition system collection falling rocks video image by building in advance, to the video
Acquisition system, which carries out camera calibration, to be included：
Two marks being pre-configured with are placed on massif, promote two mark lines to be located at visual field horizontal direction, and survey
The actual range of fixed two marks；
The physical location of two marks is changed, promotes two mark lines to be located at visual field vertical direction, and determines two marks
The actual range of thing；
According to the actual range on the actual range and vertical direction in above-mentioned horizontal direction, video acquisition system is determined
In, the physical length represented by per pixel, and camera calibration is carried out to video acquisition system according to the physical length.
Further, the method for the falling rocks detection includes being based on background subtraction and target signature detection algorithm.
Further, according to the every frame falling rocks bianry image determined, determine described per corresponding to frame falling rocks bianry image
Area include：
The traversal of pixel is carried out to single falling rocks target, wherein, the number of pixel is the area of present frame falling rocks.
Further, according to the every frame falling rocks bianry image determined, determine described per corresponding to frame falling rocks bianry image
Nominal altitude include：
According to the every frame falling rocks bianry image determined, the moment of inertia of calculating falling rocks target image；
According to the moment of inertia of the falling rocks target image, the major axis value and short axle value of falling rocks target image are determined；
According to the major axis value and the short axle value, the average of the major axis value and the short axle value is determined, and this is equal
Value is used as the nominal altitude.
Beneficial effects of the present invention：The present invention is based on digital image processing techniques and mode identification technology, using list
Camera, heed contacted measure technology, falling rocks target difference appearance is established using area and the nominal altitude parameter of different images sequence
The set of state lower body volume data, and the volume that falling rocks is finally estimated in sub-clustering is carried out to set using K mean cluster algorithm.The algorithm
Without establishing the model library of target component to be measured in advance, image processing algorithm is simply efficient.
Embodiment
Below in conjunction with the embodiment of the present invention, the technical scheme in the embodiment of the present invention is clearly and completely described,
Obviously, described embodiment is only part of the embodiment of the present invention, rather than whole embodiments.Based in the present invention
Embodiment, the every other embodiment that those of ordinary skill in the art are obtained, belongs to the scope of protection of the invention.
According to an embodiment of the invention, there is provided a kind of image falling rocks cubage measuring method based on K mean cluster analysis,
Falling rocks video image, and every frame figure of the falling rocks video image to collection are gathered including the video acquisition system by building in advance
As carrying out falling rocks detection, it is determined that the falling rocks bianry image per frame after detection；According to the every frame falling rocks bianry image determined, it is determined that
The area and nominal altitude per corresponding to frame falling rocks bianry image；According to the area and nominal altitude, falling rocks mesh is calculated
Target volume, and according to K mean cluster parser, the volume set to falling rocks target clusters；To some after cluster
First number data of cluster are counted, and according to statistical result, determine the most cluster of number, and according to the cluster, estimate falling rocks mesh
Target volume, wherein, the method for estimation is to take the average of all elements in cluster.
Further, the above-mentioned image falling rocks cubage measuring method based on K mean cluster analysis also includes：By advance
Before the video acquisition system collection falling rocks video image built, camera calibration is carried out to the video acquisition system.
Further, before the video acquisition system collection falling rocks video image by building in advance, to the video
Acquisition system, which carries out camera calibration, to be included：Two marks being pre-configured with are placed on massif, promote two mark lines to be located at
Visual field horizontal direction, and determine the actual range of two marks；The physical location of two marks is changed, promotes two mark lines
Positioned at visual field vertical direction, and determine the actual range of two marks；According to the actual range in above-mentioned horizontal direction and vertically
Actual range on direction, is determined in video acquisition system, the physical length represented by per pixel, and according to the physical length pair
Video acquisition system carries out camera calibration.
Further, the method for the falling rocks detection includes being based on background subtraction and target signature detection algorithm.
Further, according to the every frame falling rocks bianry image determined, determine described per corresponding to frame falling rocks bianry image
Area include：The traversal of pixel is carried out to single falling rocks target, wherein, the number of pixel is the area of present frame falling rocks.
Further, according to the every frame falling rocks bianry image determined, determine described per corresponding to frame falling rocks bianry image
Nominal altitude include：According to the every frame falling rocks bianry image determined, the moment of inertia of calculating falling rocks target image；According to described
The moment of inertia of falling rocks target image, determine the major axis value and short axle value of falling rocks target image；According to the major axis value and described short
Axle value, the average of the major axis value and the short axle value is determined, and using the average as the nominal altitude.
In order to facilitate the above-mentioned technical proposal of the present invention is understood, below by way of concrete operating principle and working method, to this
The above-mentioned technical proposal of invention is described in detail.
First step camera calibration
Video acquisition system is built, two marks are placed on massif, and two mark lines are located at visual field horizontal direction,
Determine the actual range L of two marksw, the center-of-mass coordinate of two marks is respectively (x on video image1, y1) and (x2, y2)；More
The position of two marks is changed, line is allowed to and is located at visual field vertical direction, the actual range of two marks of measure is Lh, video image
The barycenter of upper two mark is respectively (x1', y1') and (x2', y2’)；Then in video image, the physical length d of expression per pixel
It is expressed as：
Second step image procossing and falling rocks detection
Every two field picture is pre-processed using image processing techniques and falling rocks detects.Falling rocks detection, which uses, is based on background subtraction
Point-score technology realizes that the falling rocks bianry image per frame after detection is f (i, j) with the algorithm that target signature technology is combined；
3rd step falling rocks target component calculates
Area s and nominal altitude h calculating is carried out to every frame falling rocks bianry image f (i, j).
The traversal of pixel is carried out to single falling rocks target, the number of pixel is the area s when frame falling rocks.
Nominal altitude h is a virtual value, and it is defined as the volume V of falling rocks and its image projection area s ratio, in this hair
In bright, when the nominal altitude of frame corresponds to a certain projected area of falling rocks, with major axis DmaxWith short axle DminAverage represent.Calculate first
The moment of inertia M of falling rocks target image：
In formula, i and j represent abscissa and ordinate of the falling rocks target on image respectively；∑ represents sum operation symbol；
CxAnd CyThe center of gravity abscissa and ordinate of falling rocks target image are represented respectively, are represented with following formula,
Major axis DmaxWith short axle DminRepresented with such as following formula：
When frame nominal altitude then can using approximate representation as
4th step falling rocks target volume calculation
After same falling rocks target component, which calculates, meets 20 frame, its volume can be calculated.20 frame areas and nominal altitude are deposited
Preserve and be combined into { s1, s2, s3...s20And { h1, h2, h3...h20}.Target volume set V is expressed as：
{s1h2, s1h3...s1h20, s2h1, s2h3...s20h1, s20h19,
Volume collective number is 380.
Target volume set is clustered using K mean algorithms, k values are 3, are specially
By s1h2, s2h1, s3h1As cluster initialization central point, set of computations per volume element point and central point away from
From volume element vegetarian refreshments being categorized into the cluster nearest with center；
The barycenter of each cluster is recalculated, and using barycenter as central point, repeats a steps；
Repeat step a and b, until the error of central point is less than fixed value ε twice；
The element number of 3 clusters is counted, the most cluster of number is expressed as：
{v1, v2, v3...vm}
Wherein m represents number of elements.Carry out the estimation of volume with this cluster, evaluation method is take all elements in cluster equal
It is worth, as a result vgRepresent as follows：
In formula, j represents the sequence number of element in cluster.Result is transformed into actual size Vr
Vr=Vg×d3
So far, the image falling rocks cubage measuring method based on K mean cluster analysis is realized.
In summary, by means of the above-mentioned technical proposal of the present invention, by with digital image processing techniques and pattern-recognition
Based on technology, using one camera, heed contacted measure technology, built using the area and nominal altitude parameter of different images sequence
The set of vertical falling rocks target difference posture lower body volume data, and sub-clustering is carried out to set using K mean cluster algorithm and finally estimated
The volume of falling rocks.Model library of the algorithm without establishing target component to be measured in advance, image processing algorithm are simply efficient.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all essences in the present invention
God any modification, equivalent substitution and improvements made etc., should be included in the scope of the protection with principle.
Claims (6)
- A kind of 1. image falling rocks cubage measuring method based on K mean cluster analysis, it is characterised in that including：Falling rocks video image, and every frame figure of the falling rocks video image to collection are gathered by the video acquisition system built in advance As carrying out falling rocks detection, it is determined that the falling rocks bianry image per frame after detection；According to the every frame falling rocks bianry image determined, determine that the area per corresponding to frame falling rocks bianry image and name are high Degree；According to the area and nominal altitude, the volume of falling rocks target is calculated, and according to K mean cluster parser, to falling rocks The volume set of target is clustered；The element number of some clusters after cluster is counted, and according to statistical result, determines the most cluster of number, and root According to the cluster, the volume of falling rocks target is estimated, wherein, the method for estimation is to take the average of all elements in cluster；The nominal altitude is a virtual value, and it is defined as the volume V of falling rocks and its image projection area s ratio.
- 2. the image falling rocks cubage measuring method according to claim 1 based on K mean cluster analysis, it is characterised in that Also include：Before the video acquisition system collection falling rocks video image by building in advance, phase is carried out to the video acquisition system Machine is demarcated.
- 3. the image falling rocks cubage measuring method according to claim 2 based on K mean cluster analysis, it is characterised in that Before the video acquisition system collection falling rocks video image by building in advance, camera mark is carried out to the video acquisition system Surely include：Two marks being pre-configured with are placed on massif, promote two mark lines to be located at visual field horizontal direction, and determine two The actual range of mark；The physical location of two marks is changed, promotes two mark lines to be located at visual field vertical direction, and determine two marks Actual range；According to the actual range on the actual range and vertical direction in above-mentioned horizontal direction, determine in video acquisition system, often Physical length represented by pixel, and camera calibration is carried out to video acquisition system according to the physical length.
- 4. the image falling rocks cubage measuring method according to claim 1 based on K mean cluster analysis, it is characterised in that The method of the falling rocks detection includes being based on background subtraction and target signature detection algorithm.
- 5. the image falling rocks cubage measuring method according to claim 1 based on K mean cluster analysis,Characterized in that, according to the every frame falling rocks bianry image determined, determine described per corresponding to frame falling rocks bianry image Area includes：The traversal of pixel is carried out to single falling rocks target, wherein, the number of pixel is the area of present frame falling rocks.
- 6. the image falling rocks cubage measuring method according to claim 1 based on K mean cluster analysis, it is characterised in that According to the every frame falling rocks bianry image determined, determine that the nominal altitude per corresponding to frame falling rocks bianry image includes：According to the every frame falling rocks bianry image determined, the moment of inertia of calculating falling rocks target image；According to the moment of inertia of the falling rocks target image, the major axis value and short axle value of falling rocks target image are determined；According to the major axis value and the short axle value, the average of the major axis value and the short axle value is determined, and the average is made For the nominal altitude.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510195309.5A CN104992428B (en) | 2015-04-23 | 2015-04-23 | A kind of image falling rocks cubage measuring method based on K mean cluster analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510195309.5A CN104992428B (en) | 2015-04-23 | 2015-04-23 | A kind of image falling rocks cubage measuring method based on K mean cluster analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104992428A CN104992428A (en) | 2015-10-21 |
| CN104992428B true CN104992428B (en) | 2018-01-02 |
Family
ID=54304236
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510195309.5A Active CN104992428B (en) | 2015-04-23 | 2015-04-23 | A kind of image falling rocks cubage measuring method based on K mean cluster analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104992428B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105547325B (en) * | 2015-12-08 | 2018-06-19 | 北京航天时代光电科技有限公司 | A kind of optical fiber based on K mean cluster is used to a group temperature model coefficient and determines method |
| CN107452021A (en) * | 2016-04-19 | 2017-12-08 | 深圳正谱云教育技术有限公司 | Camera to automatically track system and method based on single-lens image Dynamic Recognition |
| CN109000559B (en) * | 2018-06-11 | 2020-09-11 | 广东工业大学 | Object volume measuring method, device and system and readable storage medium |
| CN113188975B (en) * | 2021-05-07 | 2022-07-15 | 中南大学 | Rock mass fracture and flying rock motion analysis system and method based on image processing technology |
| CN113344964B (en) * | 2021-06-23 | 2024-02-23 | 江苏三恒科技股份有限公司 | Mine robot falling stone monitoring and early warning method based on image processing |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1894720A (en) * | 2003-12-16 | 2007-01-10 | 美国西门子医疗解决公司 | Toboggan-based shape characterization |
| GB2428924B (en) * | 2005-07-22 | 2007-10-31 | Snell & Wilcox Ltd | Method of and apparatus for image analysis |
| CN102768022A (en) * | 2012-07-20 | 2012-11-07 | 长安大学 | Tunnel surrounding rock deformation detection method adopting digital camera technique |
| RU2012129943A (en) * | 2012-07-16 | 2014-01-27 | Институт проблем комплексного освоения недр Российской академии наук (ИПКОН РАН) | METHOD FOR LARGE-SCALE EXPLOSIVE DESTRUCTION OF MOUNTAIN ARRAYS OF COMPLEX STRUCTURE FOR SELECTIVE DIGGING OF USEFUL FOSSIL IN OPEN WORKS |
-
2015
- 2015-04-23 CN CN201510195309.5A patent/CN104992428B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1894720A (en) * | 2003-12-16 | 2007-01-10 | 美国西门子医疗解决公司 | Toboggan-based shape characterization |
| GB2428924B (en) * | 2005-07-22 | 2007-10-31 | Snell & Wilcox Ltd | Method of and apparatus for image analysis |
| RU2012129943A (en) * | 2012-07-16 | 2014-01-27 | Институт проблем комплексного освоения недр Российской академии наук (ИПКОН РАН) | METHOD FOR LARGE-SCALE EXPLOSIVE DESTRUCTION OF MOUNTAIN ARRAYS OF COMPLEX STRUCTURE FOR SELECTIVE DIGGING OF USEFUL FOSSIL IN OPEN WORKS |
| CN102768022A (en) * | 2012-07-20 | 2012-11-07 | 长安大学 | Tunnel surrounding rock deformation detection method adopting digital camera technique |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104992428A (en) | 2015-10-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104992428B (en) | A kind of image falling rocks cubage measuring method based on K mean cluster analysis | |
| CN104173054B (en) | Measuring method and measuring device for height of human body based on binocular vision technique | |
| CN105157608B (en) | A kind of detection method of overrun vehicle, apparatus and system | |
| CN109145696B (en) | Old people falling detection method and system based on deep learning | |
| CN107679503A (en) | A kind of crowd's counting algorithm based on deep learning | |
| CN105868707B (en) | A kind of falling from bed behavior real-time detection method based on deep image information | |
| CN106683070A (en) | Body height measurement method and body height measurement device based on depth camera | |
| US20190294890A1 (en) | Method and apparatus for surveillance | |
| CN107909604A (en) | Dynamic object movement locus recognition methods based on binocular vision | |
| CN106210634A (en) | A kind of wisdom gold eyeball identification personnel fall down to the ground alarm method and device | |
| CN106978774A (en) | A kind of road surface pit automatic testing method | |
| EP3729135A1 (en) | Object tracking | |
| CN104036488A (en) | Binocular vision-based human body posture and action research method | |
| CN108876799A (en) | A kind of real-time step detection method based on binocular camera | |
| CN107295230A (en) | A kind of miniature object movement detection device and method based on thermal infrared imager | |
| CN111062918B (en) | Abnormality detection method and device based on computer vision | |
| CN109271918B (en) | Method for distinguishing people with balance ability disorder based on gravity center shift model | |
| CN105760809A (en) | Method and apparatus for head pose estimation | |
| KR20150010248A (en) | Method and apparatus for surveillance by using 3-dimension image data | |
| CN115082825A (en) | Video-based real-time human body falling detection and alarm method and device | |
| KR20130058526A (en) | Method and apparatus for measuring spin character of rotating body | |
| CN106033614A (en) | Moving object detection method of mobile camera under high parallax | |
| CN101494772A (en) | Method and apparatus for detecting image | |
| CN109558797B (en) | Method for distinguishing human body balance disorder based on gravity center area model under visual stimulation | |
| Russo et al. | A new approach to obtain metric data from video surveillance: Preliminary evaluation of a low-cost stereo-photogrammetric system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |