WO2016192516A1 - 视频监控布点 - Google Patents

视频监控布点 Download PDF

Info

Publication number
WO2016192516A1
WO2016192516A1 PCT/CN2016/081736 CN2016081736W WO2016192516A1 WO 2016192516 A1 WO2016192516 A1 WO 2016192516A1 CN 2016081736 W CN2016081736 W CN 2016081736W WO 2016192516 A1 WO2016192516 A1 WO 2016192516A1
Authority
WO
WIPO (PCT)
Prior art keywords
camera
spatial
monitoring
sample points
distribution
Prior art date
Application number
PCT/CN2016/081736
Other languages
English (en)
French (fr)
Inventor
何伟魏
刘常积
叶倩燕
柴亚琴
Original Assignee
浙江宇视科技有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Priority claimed from CN201510288095.6A external-priority patent/CN104899368B/zh
Priority claimed from CN201510540787.5A external-priority patent/CN105159978B/zh
Application filed by 浙江宇视科技有限公司 filed Critical 浙江宇视科技有限公司
Priority to US15/577,466 priority Critical patent/US10445348B2/en
Publication of WO2016192516A1 publication Critical patent/WO2016192516A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/29—Geographical information databases
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/38—Electronic maps specially adapted for navigation; Updating thereof
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T11/00—2D [Two Dimensional] image generation
    • G06T11/40—Filling a planar surface by adding surface attributes, e.g. colour or texture
    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B13/00—Burglar, theft or intruder alarms
    • G08B13/18—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength
    • G08B13/189—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength using passive radiation detection systems
    • G08B13/194—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength using passive radiation detection systems using image scanning and comparing systems
    • G08B13/196—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength using passive radiation detection systems using image scanning and comparing systems using television cameras
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N7/00—Television systems
    • H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast

Definitions

  • the invention relates to video surveillance distribution.
  • the commonly used camera layout method is to abstract the objective world into points (such as intersections, ATM machines, etc.), lines (roads), planes (CBD, large shopping malls, etc.), and then calculate the coverage area of the monitoring points, according to whether the coverage area contains The objective world or the extent of the objective world is included to judge the rationality of the monitoring point layout, and then adjust and obtain the final camera layout demand map according to the judgment result, and arrange the camera.
  • points such as intersections, ATM machines, etc.
  • lines roads
  • planes CBD, large shopping malls, etc.
  • the layout method focuses on the coverage of the surveillance area by the camera, for example, by calculating the coverage of the camera and the coverage of the surveillance area by the camera. This process is usually complicated.
  • the coverage of the camera involves the type of camera, focal length, corner, lens, resolution, and installation height.
  • the monitoring target involves various entities in the objective world. If you consider the consideration of entities in the objective world, the requirements for the data will be multiplied several times, because there are references between the entities and entities in the objective world, but there is no reproducibility.
  • Another shortcoming of this method is that there is no judgment or distinction on the importance of entities in the objective world, that is, the distribution of important areas and non-essential areas is the same, which is likely to cause insufficient monitoring points in important areas. There are too many monitoring points in non-critical areas.
  • a very important indicator is whether to cover key areas, that is, to arrange as many monitoring points as possible in key areas to ensure that all key areas can be fully Efficient monitoring of azimuth coverage.
  • key areas that is, to arrange as many monitoring points as possible in key areas to ensure that all key areas can be fully Efficient monitoring of azimuth coverage.
  • the layout of the entire monitoring point it is generally given clear requirements based on the monitored objects, such as covering the densely populated areas such as major banks and supermarkets.
  • abstracting objective objects into point, line and polygon data structures based on vector data operations may neglect the essential properties of objective things.
  • the cameras arranged in the key areas with high frequency of video capture are insufficient due to the fact that the area is not considered, and cameras of the same density are arranged in some non-critical areas. This has led to a waste of resources to some extent, and may also result in poor monitoring of critical areas due to insufficient camera placement.
  • the operating frequency of the camera can be counted. For example, it can be calculated which cameras are used more frequently than the average, and it is considered that the cameras above the average are properly utilized, and the cameras below the average are unreasonable.
  • this analysis method may lead to misjudgment. For example, for a focus area, multiple cameras are usually deployed at the same time to achieve full coverage of the area. The monitoring of the area by the user is performed by the camera group, and the monitoring video of the camera group can be frequently viewed, but the frequency of operation may not be high for a single camera. At this time, if the single camera operation frequency is counted according to the above method, the analysis result of the lower utilization rate of the group camera will be obtained, and the analysis result is not in accordance with the actual situation.
  • the raster data structure is a data organization form based on the grid model, which divides the space into regular grids (hereinafter referred to as grid cells), and gives corresponding attribute values on each grid unit. Represents a specific attribute of a corresponding geographic entity.
  • the raster data structure can visually represent the impact of the objective entity on the surroundings.
  • a video monitoring distribution method is provided to generate camera distribution position data in a specific spatial range based on sample point data capable of reflecting an objective monitoring distribution requirement.
  • the method may include: determining a plurality of sample points on the map; determining a spatial range to be analyzed based on the distribution area of the plurality of sample points, and rasterizing the spatial range to Obtaining a grid unit of the spatial extent; performing weight assignment on each of the sample points, wherein the weight assignment represents a size of a monitoring requirement at the sample point; and a weight value according to each of the sample points and each sample The positional relationship between the point and each of the grid cells is calculated, the weight value of each of the grid cells is calculated, and the camera distribution location data of the spatial range is generated according to the weight values of the grid cells.
  • the camera distribution position data in the spatial range is generated according to the weight value of each sample point and the geographical relationship between each sample point and each spatial position in the spatial range.
  • the camera can distribute the position data with high reliability.
  • the method may further include: acquiring, according to the spatially distributed camera position data, a monitoring intensity matrix that can represent camera monitoring strength at different spatial locations within the spatial range; a range of camera operation log data, an operation intensity matrix that can represent a camera operation strength at different spatial locations within the spatial range; and a correlation coefficient between the monitoring intensity matrix and the operation intensity matrix, and The correlation coefficient analyzes the degree of rationality of the camera distribution within the spatial extent.
  • the correlation coefficient between the monitoring intensity matrix and the operation intensity matrix in a specific spatial range it can be more accurately judged whether the camera distribution in the spatial range is reasonable.
  • the monitoring intensity matrix and the operational strength matrix are usually positively correlated. If it is determined according to the correlation coefficient that the two do not conform to the proper relationship, it can be determined that the camera distribution of the spatial range may need to be rationalized and improved.
  • a video monitoring distribution device comprising a processor executable by reading and executing a machine stored on a storage medium corresponding to a video monitoring point control logic Directing to: determine a plurality of sample points on the map; determining a spatial extent to be analyzed based on the distribution area of the plurality of sample points, and rasterizing the spatial extent to obtain the spatial extent a grid unit; performing weight assignment on each of the sample points, wherein a weight value of each of the sample points represents a monitoring requirement size at the sample point; and a weight value according to each of the sample points and each network Calculating the weight value of each of the grid cells, and generating camera distribution location data of the spatial extent according to the weight values of the grid cells.
  • FIG. 1 is a rendering effect diagram of rendering DEM data by using a rendering color band based on regional heat according to an embodiment of the invention
  • FIG. 2 is an effect diagram of superimposing the rendering effect diagram shown in FIG. 1 with an actual map
  • FIG. 3A is a schematic flowchart of a method for determining a video monitoring distribution according to an embodiment of the invention
  • FIG. 3B is a schematic diagram of functional modules of an apparatus for determining a video monitoring point according to an embodiment of the invention.
  • FIG. 4 is a schematic flowchart of a method for analyzing the rationality of camera distribution according to an exemplary embodiment of the present application
  • FIG. 5 is a hardware structural diagram of a monitoring device according to an exemplary embodiment of the present application.
  • FIG. 6 is a schematic diagram showing spatial mesh division according to an exemplary embodiment of the present application.
  • FIG. 7 is a schematic diagram showing an effect of rendering a preset spatial region based on a monitoring intensity factor according to an exemplary embodiment of the present application.
  • FIG. 8 is a schematic flowchart of a method for evaluating the rationality of camera distribution according to a correlation coefficient between a monitoring intensity matrix and an operational intensity matrix, according to an exemplary embodiment of the present application;
  • FIG. 9 is a schematic flowchart of a method for evaluating the rationality of camera distribution according to another exemplary embodiment of the present application.
  • FIG. 10 is a schematic diagram showing a superposition operation according to an exemplary embodiment of the present application.
  • FIG. 11 is a functional block diagram of an analysis apparatus for plausibility of camera distribution according to an exemplary embodiment of the present application.
  • FIG. 12 is a functional block diagram of an analysis apparatus for plausibility of camera distribution according to another exemplary embodiment of the present application.
  • the video monitoring distribution method according to an embodiment of the present invention may be as shown in FIG. 3A, and includes the following steps 310-340.
  • step 310 a plurality of sample points on the map are determined.
  • Step 310 can be performed by sample point determination module 1310 shown in FIG. 3B.
  • the weight of the sample points reflects the data heat of the sample points, that is, the degree of importance.
  • the sample points may be POI points (Point of Interesting) or points generated according to other distribution rules. For example, according to the location data of the case, the distribution law of the case can be obtained; according to the existing camera points, the law of the location of the camera can be obtained. According to these distribution rules, Sample points that reflect the needs of the objective world monitoring points to a certain extent can be obtained.
  • sample points there are many ways to select sample points.
  • the location of the case, the distribution of existing camera points, etc. can reflect the monitoring requirements of the distribution, and the points corresponding to the maps can be extracted as sample points.
  • the distribution of POI can more comprehensively reflect the flow of people in the monitored area. For example, the denser the POI density, the higher the heat of the area, the more important the monitoring will be, and the key deployment is required. Otherwise, the importance is low.
  • sample points are points of interest on the map.
  • Existing network maps generally have points of interest, and the points of interest can more fully and accurately reflect the distribution of human traffic in the map area.
  • a spatial extent to be analyzed is determined based on a distribution area of the plurality of sample points, and the spatial extent is rasterized to obtain a grid unit of the spatial extent.
  • the rasterization process may include generating raster data corresponding to the spatial extent on the map to enable processing of latitude and longitude information and attribute value information at respective spatial locations within the spatial extent based on the grid unit.
  • the raster data may be a two-dimensional array, and each element in the two-dimensional array may have a one-to-one correspondence with each grid unit in the map area.
  • the one-to-one correspondence may include: an index of each element in the two-dimensional array may represent a latitude and longitude value of the corresponding grid unit, and a value of each element in the two-dimensional array may represent a weight value of the corresponding grid unit.
  • Step 320 can be performed by rasterization processing module 1320.
  • the Digital Elevation Model is an entity ground model that represents the ground elevation in an array with an ordered set of values.
  • raster data in the form of DEM (hereinafter may be referred to simply as DEM data) may be employed, the index of the grid unit may be used to represent the latitude and longitude of the corresponding block on the map, and the value of the grid unit may represent a custom attribute.
  • the value of the grid cell can be defined as the weight value of the grid cell.
  • the latitude and longitude can be converted into an index of the DEM data to facilitate subsequent assignment and display operations.
  • a specific method of converting latitude and longitude into an index of DEM data may include the following steps 320a and steps 320b, to reduce the area of the map that needs to be rasterized to a reasonable range to reduce the complexity of the operation, and include all sample points at the same time.
  • the latitude and longitude values of the sample points can be converted into coordinate values in the Mercator projection coordinate system.
  • step 320a it is determined that all of the sample points are distributed over the longitude and the minimum and the maximum and minimum at the latitude.
  • X is the longitude of the POI and Y is the latitude of the POI.
  • the worldwide Mercator projection coordinate range is (-20037508, -20037508, 20037508, 20037508.34).
  • Width x max-x min
  • width represents the width of the POI distribution area and height represents the height of the POI distribution area.
  • determining the minimum value x min and the maximum value x max of the POI in the x direction correspondingly determine the minimum and maximum values of the POI in the longitude, and similarly, determining the POI in the y direction.
  • the minimum value y min and the maximum value y max correspondingly determine the minimum and maximum values of the POI in latitude.
  • the maximum and minimum values on the latitude and longitude can be respectively converted into corresponding coordinate values in the Mercator projection coordinate system, including: the maximum and minimum values in the longitude correspond to the maximum value in the x direction in the Mercator projection coordinate system. And the minimum value, the maximum and minimum values on the latitude correspond to the y direction in the Mercator projection coordinate system The maximum and minimum values above.
  • the rectangular region formed by the four vertices at the maximum and minimum latitude and longitude map positions determined in step 320a is determined as the map region to be rasterized, that is, the spatial extent to be analyzed.
  • an area such as a rectangle may be formed, which is a map area that needs to be rasterized.
  • the raster data generating module can apply a two-dimensional array DemData in the memory to the same width and height ratio of the POI distribution area according to the memory size and the calculation precision requirement of the computer.
  • DemData can represent DEM data, where the two-dimensional index of the array (that is, the position of the row and column in the array) corresponds to the specific Mercator projection coordinate value, and each value in the array indicates the importance of the region.
  • the POI data used in the current embodiment is from the main urban area of Hangzhou, and the DEMWidth of the set DEM data can be set to 5000.
  • the DEM data width DemWidth is set to 5000, and the memory size is above 2G.
  • the height DemHeight of the DEM data can be:
  • Xindex and yindex are the indices of the elements in the two-dimensional array.
  • Each element in the DEM data in the form of a two-dimensional array has a one-to-one correspondence with the grid unit, and the index of each element in the two-dimensional array corresponds to the latitude and longitude values of the grid unit, and the numerical value of each element represents the network.
  • the weight value of the cell is the weight value of the cell.
  • a weight assignment is performed for each of the sample points, wherein the weight assignment represents a size of the monitoring requirement at the sample point.
  • Step 330 can be completed by sample point assignment module 1330. It should be noted that the step may be performed after step 310 and before step 320, or after step 320 and before step 340, and may also be performed simultaneously with step 320. In the current embodiment, step 330 may be performed after the rasterization process is completed in step 320. In fact, as long as step 330 can be completed before step 340.
  • sample points are different in the geographical location reflected in the map, for example, representing a municipal government, school, or entertainment venue, the monitoring needs of each sample point may be different. Therefore, sample points can be classified to give different weights to sample points depending on the importance of the category.
  • the factors affecting the layout requirements may be considered, including but not limited to: the geographical location of the sample points, the flow of people at each sample point, and the public The size of the sample point.
  • the geographical location includes whether the administrative division is at the center or the suburb. In general, the central location is more weighted than the suburbs.
  • the central location is more weighted than the suburbs.
  • the weight For people traffic, the greater the traffic, the greater the weight.
  • the flow of people at the sample points in categories such as shopping malls and hospitals is generally large, while the traffic at the sample points in farms, fields, and the like is generally small.
  • the public's demand for various sample points refers to the necessity of a certain type of sample points. The greater the necessity, the greater the weight. For example, if the public demand at the sample point of the main road of urban traffic is large, the weight is also relatively large. However, some alternative locations, such as lottery sales outlets, have less weight and less weight.
  • the classification of sample points can have multiple levels.
  • the analytic hierarchy process can be used to assign corresponding weights to sample points.
  • the geographic locations reflected by some sample points may belong to one larger category, while the larger categories are subdivided into smaller categories at multiple levels.
  • the larger category is government agencies, and the larger category includes central government, local government, and grassroots organizations. Small category.
  • the analytic hierarchy process can be performed by assigning values to each of the minimum categories multiple times, so that a weight that is more reflective of the actual importance can be obtained.
  • weighting the sample points may specifically include classifying the POIs and assigning weights by category.
  • the POIs are classified according to the existing POI classification comparison table, and the weights of the POIs are assigned by the analytic hierarchy process. As shown in Table 1, Table 1 shows the classification and impact values of POI:
  • the weight value of the POI can be assigned by, for example, three experts according to their own knowledge of the importance of the POI.
  • x1, x2, and x3 represent the weight assignments of the three experts to the sample points, respectively.
  • the maximum weight value may be limited to, for example, 30. Then, under the A-M big classification, the analytic hierarchy process is performed on each small classification, and the value of each factor is obtained.
  • the scale a ij is introduced, and the meanings of the different values of the scale a ij can be as shown in Table 2.
  • the i factor is as important as the j factor 3
  • the i factor is slightly more important than the j factor 5
  • the i factor is more important than the j factor 7
  • the i factor is very important to the j factor 9
  • the i factor is absolutely important than the j factor 2, 4, 6, 8
  • the two-two comparison matrices of the four factors a1, a2, a3, a4 can be as shown in Table 3.
  • the four factors a1, a2, a3, and a4 represent the four secondary classifications under the primary classification A, respectively.
  • the consistency check of the two-factor comparison matrix of the four factors can be performed at the first level.
  • the judgment matrix A can be obtained according to Table 3, and normalization processing such as normalization can be performed before the judgment matrix A is subjected to the consistency check.
  • Operation 1 is to perform column vector normalization on the judgment matrix A; operation 2 is summation by row; and 3 is the finalized normalization matrix.
  • the maximum eigenvalue of the judgment matrix A can be obtained.
  • the feature vector W (0) can be obtained.
  • A1 represents a pairwise comparison matrix between the assignments x1, x2 and x3 of each expert to the secondary category a1.
  • A2, A3, and A4 respectively represent the two-two comparison matrix between the assignments x1, x2, and x3 of the secondary classes a2, a3, and a4 by each expert.
  • W (1) , W (2) , W (3), and W (4) represent the eigenvectors of the judgment matrices A1, A2, A3, and A4.
  • the consistency indicator can be used for testing:
  • the corresponding RI value can be as follows:
  • CR is less than 0.1, it can be said that the degree of inconsistency of the judgment matrix A is within the allowable range.
  • the sample points may be reclassified or the sample points may be re-assigned directly, and the feature vector of the judgment matrix A may be used instead of the weight vector.
  • the consistency check can also be performed by using the above principles.
  • each POI can find a unique element index xindex and yindex in the DEM data corresponding to its latitude and longitude.
  • POI represents people's points of interest, and there are areas that people may care about at and near the POI point. Therefore, the closer the distance POI is, the more important the area is, the stronger the monitoring demand is; and the denser the POI density in the area represents the higher the heat of the area, for example, it may be the traffic flow area, and the monitoring needs to be deployed, and vice versa.
  • the effect of the POI on the perimeter has the property of attenuating with distance. Therefore, in the current embodiment of the present invention, the influence attenuation with distance can be simulated by setting the influence range and the maximum influence value.
  • a weight value of each of the grid cells is calculated according to a weight value of each of the sample points and a positional relationship with each of the grid cells, and is generated according to weight values of each of the grid cells.
  • the spatial range of camera distribution location data is calculated according to a weight value of each of the sample points and a positional relationship with each of the grid cells, and is generated according to weight values of each of the grid cells.
  • This step can be accomplished by the grid unit assignment module 1340. Since a sample point has an influence on the surrounding mesh, and the influence can be attenuated as the distance increases, the weight of each mesh unit can be obtained according to the weight value of each sample point and the positional relationship with each mesh unit. Value to be able to more realistically reflect the effect of sample points on the surrounding mesh. For example, assume that the impact value of the POI is the Influence determined in step 330, and the maximum distance affected is MaxDis tan ce. Through the x max, x min, y max, y min in step 320, the area range xwidth and ywidth (the unit can be meters) corresponding to the grid unit where the current POI is located can be calculated. In this way, by combining the DEMWidth of the DEM data and the height DemHeight of the DEM data, the resolution of the DEM data can be obtained as follows:
  • the resolution of the DEM data can represent the size of each grid unit. In this way, you can specify the maximum influence distance of a single sample point on the map, and you can determine the weight component of each grid unit within a range of its influence by a single sample point according to the following formula:
  • Dis tan ce represents the distance from the sample point to the grid unit.
  • the distance Dis tan ce is 0.
  • Influence indicates the weight value of the sample point.
  • the weight value of each grid unit is the sum of the weight components of all the sample points for the grid unit.
  • the weight value of the grid unit can be represented by the value of the corresponding element in the DEM data.
  • the weight value of each grid unit reflects the monitoring requirements of the map area corresponding to the grid unit.
  • the DEM data can be rendered.
  • step 340 may further include the following operations: performing dimensionless processing on the weight values of the grid cells to obtain weightless dimensionless raster data.
  • the dimensionless processing of the values of the elements corresponding to the grid cells in the DEM data can be done by the dimensionless processing module 1350. Since the weight value of the grid unit is calculated based on the weight value of the artificially set sample point, the weight value of the grid unit can be dimensionlessly processed to obtain the dimensionless weight of the dimensionless weight including all grid elements. Raster data to reflect the monitoring needs of the objective world relatively accurately. For example, considering the need for rendering, the dimensionless processing can be selected to be minimized so that the weight of ownership is between 0-1 after processing.
  • the specific processing method may be: dividing all the values in the DEM data by the maximum value thereof, and the obtained values represent the importance degree relative to the region with the strongest monitoring demand in the study area, wherein the closer to 1 is, the more important The closer to 0, the less important it is.
  • the pre-generated rendering ribbon can be used to render the rendered DEM data by rendering the display module 1360 to obtain a rendered image.
  • the corresponding color can be used to render according to the importance of the data in the grid unit, and the rendering result can be as shown in FIG. 1 .
  • the color of the rendered ribbon can be varied to distinguish different weight values. In the current embodiment, white is the starting color and black is the ending color, which is drawn on a 100*100 wide and wide picture.
  • the rendering display module 1360 can superimpose and display the rendering map and the map to obtain a monitoring layout requirement map.
  • Each grid element in the DEM data is capable of corresponding to the real latitude and longitude, so it can be displayed in an existing map. For example, you can calculate the four-corner control points for each grid cell and then plot them in an existing map.
  • the result of superimposing the rendered image and the map, that is, the monitoring layout requirement map can be as shown in FIG. 2 .
  • the monitoring layout requirement map can be as shown in FIG. 2 .
  • the spatial distribution of camera distribution location data may then be generated by the rendering display module 1360 based on the monitoring placement requirement map.
  • the weight value of the sample points reflects the monitoring demand of the objective world at the sample point.
  • the weight values of all sample points in the area are taken into account, which makes the monitoring points higher. Credibility.
  • the present application also provides an analysis method for the rationality of camera distribution to evaluate whether the distribution of the surveillance cameras within a predetermined spatial range is reasonable.
  • the preset spatial range may be the Binjiang area of Hangzhou, the Haidian District of Beijing, the Pudong District of Shanghai, etc., and the spatial scope of the application can be set by the user.
  • the "reasonability of the camera distribution" to be evaluated in the method may be: arranging more cameras in key areas where monitoring demand is strong, and placing relatively few cameras in non-critical areas where monitoring demand is weak, so that the camera It is reasonably utilized and can meet the needs of monitoring.
  • the present application is based on the following principle of camera distribution: there is a high positive correlation between the monitoring intensity of the camera and the operational strength of the camera. For example, in key areas where users are more concerned, there are more cameras, and the supervision of this area. The control intensity is large, and the user views the video recording of the area more frequently; and for the non-critical areas that the user does not care about, the camera arrangement is less, the corresponding monitoring intensity is low, and the user rarely views the video recording of the area. . Therefore, it is possible to judge whether the arrangement of the camera is reasonable by checking the relationship between the camera monitoring intensity and the operating intensity in a certain spatial range. This analysis method combines the operation of the camera with the spatial information.
  • FIG. 4 illustrates the flow of an analysis method of the camera distribution rationality of the present application. As shown in FIG. 4, the method can include steps 410-430.
  • a monitoring intensity matrix for indicating camera monitoring intensity at different spatial locations within the preset spatial range is acquired according to camera distribution position data of a preset spatial range.
  • step 420 an operation intensity matrix for indicating camera operation intensity at different spatial locations within the preset spatial range is acquired according to the camera operation log data of the preset spatial range.
  • step 430 a correlation coefficient between the monitoring intensity matrix and the operation intensity matrix is calculated, and a degree of rationality of the camera distribution within the preset spatial range is analyzed according to the correlation coefficient.
  • the above method can be implemented by software, for example, it can be analyzed whether the distribution of the monitoring camera is reasonable by a certain monitoring software.
  • the monitoring software can be running on a physical device, which can be, for example, a monitoring device.
  • the monitoring device may include a processor 510, a memory 520, a non-volatile storage 530, and a network interface. 540.
  • the hardware can be connected to each other through an internal bus 550.
  • the processor 510 can read the video monitoring point control logic stored in the non-volatile storage medium 530 into the memory 520 to perform the method flow for determining the camera distribution position shown in FIG. 3 and/or The method flow for analyzing the rationality of camera distribution shown in FIG.
  • step 410 camera distribution location data (also referred to as monitoring point data) of a preset spatial range may be imported into the software for acquiring a monitoring intensity matrix.
  • Figure 6 illustrates a partial area of a preset spatial extent.
  • the preset spatial extent may be rasterized to divide the preset spatial extent into a regular grid (hereinafter also referred to as a grid unit).
  • a grid unit As shown in Fig. 6, the partially divided space is shown.
  • Each grid may be assigned an attribute value, and in step 410 of the present application, the attribute value of the grid may be referred to as a "monitoring strength factor.”
  • the grid w1 represented by a cross line
  • FIG. 6 can calculate a corresponding monitoring intensity factor indicating the camera monitoring intensity at the grid w1.
  • the camera is distributed in the spatial range of the grid form, and three cameras can be distributed around the grid w1, namely CA1, CA2, and CA3, and the grid
  • the monitoring intensity factor of w1 can be used to indicate how much monitoring effect the three surrounding cameras CA1, CA2, and CA3 have on the grid w1.
  • each grid may correspond to a latitude and longitude coordinate in the actual map space position, and the latitude and longitude coordinates may be converted into a two-dimensional array index in a DEM (Digital Elevation Model).
  • the grid w1 can be identified by a two-dimensional array (x, y) index corresponding to the latitude and longitude coordinates of the actual spatial position represented by the grid w1.
  • the processing of the preset spatial extent in this example includes, after rasterizing the preset space, each grid is marked with a two-dimensional array index, which is the actual spatial latitude and longitude of the grid. From the coordinate conversion, for example, a conversion method including the steps 320a and 320b as described above can be utilized.
  • Each grid can be assigned an attribute value, which is a monitoring intensity factor of a corresponding grid as described above, used to represent the camera monitoring strength at the grid.
  • the influence force F is a representation function considering the characteristics of the monitoring effect of the camera as a function of distance attenuation. For example, the camera monitors the surrounding area. The closer the camera is to the camera, the better the monitoring effect. The farther away from the camera, the worse the monitoring effect.
  • the formula for calculating the influence F can be expressed as follows:
  • F is the influence and Dis is the distance from the camera.
  • Dis is the distance from the camera.
  • the distance from the center of the grid w1 to the camera may be.
  • Value is the influence of the center point.
  • the location of the camera is the center point, because it is usually the best location for monitoring.
  • the center point influence value can be set to 1000.
  • DisMax is the range of influence of the camera, for example, it can be set to 500 meters. Among them, the values of DisMax and Value can be changed according to different cameras.
  • the distance Dis is the distance between the mesh w1 and the camera CA1; when calculating the influence force F2 of the camera CA2 on the mesh w1, the distance Dis is the mesh w1 The distance from the camera CA2; when calculating the influence force F3 of the camera CA3 on the mesh w1, the distance Dis is the distance between the mesh w1 and the camera CA3.
  • the monitoring strength factor Y1 of the final grid w1 can be calculated as follows:
  • the monitoring intensity factors of other grids within the spatial extent of Figure 6 can be calculated.
  • the monitoring intensity factor Y2 of the grid w2 the monitoring intensity factor Y3 of the grid w3, and the like can be calculated.
  • the whole of the monitoring intensity factors corresponding to the respective grids in FIG. 6 may be referred to as a “monitoring intensity matrix”, and the monitoring intensity matrix includes a plurality of monitoring intensity factors, and each factor is used to represent one of the grid cameras. Monitor the intensity.
  • the monitoring intensity factor is a dimensionless value, and the monitoring intensity factor can be standardized, for example, minimizing the minimum value, thereby converting the values of the respective monitoring intensity factors in the monitoring intensity matrix to 0 to 1 Value.
  • the monitoring intensity matrix may be rendered in different colors according to different values of the monitoring intensity factor, for example, a grid with a high intensity factor is monitored with a dark color, and a grid with a low intensity factor is monitored with a light color, and Gradient color rendering. In this way, a monitoring intensity map of the preset spatial range as illustrated in FIG. 7 can be generated, wherein the dark color area indicates that the monitoring intensity of the camera is high, that is, the camera arrangement density of these areas may be Higher.
  • an operational strength matrix can be generated from the camera operation log data.
  • the generation of the operational strength matrix may still be based on the preset spatial extent after the rasterization process mentioned in step 410, except that the attribute values of the mesh are replaced by the monitored intensity factor in step 410 to an operational strength factor.
  • the camera operation log data may be a record stored in the log server.
  • the log server When the user operates the camera, such as viewing the video recording of the camera, calling the live video of the camera, and taking a live video capture of the camera, the log server usually records the user's operation on the camera, including the operation time, the operator, and the type of operation.
  • This embodiment can import the data recorded by the log server into the software for executing the analysis method of the embodiment.
  • the log server's log data can be categorized.
  • the log server stores the operation records of multiple cameras, and the data classification can sort the operation records of each camera to obtain the operation times of each camera.
  • the software in this example can send a data acquisition request to the log server requesting to obtain a camera operation record in the most recent month. In this way, after receiving the data transmitted by the log server, the number of operations of each camera in the month can be obtained by performing data classification, and the operation strength matrix is calculated by using the operation count Count.
  • the calculation principle of the operation intensity matrix can be similar to the monitoring intensity matrix.
  • the calculation formula of the operation intensity factor in the operation intensity matrix can be seen as follows:
  • the operation intensity factor corresponding to each grid of the preset spatial range can be calculated. For example, taking the grid w1 in FIG. 6 as an example, the influences f of the three cameras CA1, CA2, and CA3 around the w1 on the grid w1 can be separately calculated, and the three influential forces f are added as a grid.
  • the operational strength factor of w1 used to represent the camera operating strength at the grid.
  • the position of the camera itself can be considered to have the highest operational intensity; and as the grid is farther away from the camera, the operational intensity at the grid is attenuated.
  • the operating intensity factor can be normalized to convert the value of the operating intensity factor at each grid to between 0 and 1.
  • each mesh of the preset spatial range can be rendered in different colors according to different operating strength factors, and a camera operation intensity map of a preset spatial range is obtained.
  • the user can intuitively see which area of the camera has a higher operating frequency within the preset space range, and the area with high operational intensity indicates a key area for key monitoring, so the operation demand is relatively strong.
  • the intensity map it can be intuitively seen which area is equipped with a dense camera, and the area with strong monitoring intensity is a key area for key monitoring.
  • the correlation strength matrix and the operation intensity matrix obtained above may be correlated and analyzed to obtain a correlation coefficient.
  • the same two-dimensional array index in the monitoring intensity matrix and the operation intensity matrix can be set to represent the same geographical location. For example, for the same two-dimensional array index value of the index of the two-dimensional array represented by the monitoring intensity matrix and the operation intensity matrix, the two correspond to the same regional location within the preset spatial range.
  • the calculation method of the correlation coefficient can be detailed as follows.
  • X is the monitoring intensity matrix and Y is the operation intensity matrix.
  • the covariance Cov(X, Y) between the monitoring intensity matrix X and the operation intensity matrix Y is calculated as follows:
  • sample mean values are used to represent the expected values E(X), E(Y).
  • the standard deviations ⁇ X and ⁇ Y of the monitoring intensity matrix X and the operation intensity matrix Y can be calculated according to the standard deviation calculation formula as follows:
  • u is the sample mean and N is the number of samples.
  • the covariance Cov(X, Y) between the two matrices can be calculated as follows between the two matrices Correlation coefficient:
  • the correlation coefficient ⁇ calculated here is a dimensionless value.
  • the degree of rationality of the camera distribution within the preset spatial range can be evaluated based on the calculated correlation coefficient. For example, the degree of rationality may indicate that the camera distribution within the preset spatial range is very reasonable, reasonable, moderately reasonable, unreasonable, and very unreasonable.
  • FIG. 8 illustrates a schematic flow chart of a method for evaluating the degree of rationality of a camera distribution based on a correlation coefficient between a monitoring intensity matrix and an operational intensity matrix.
  • a correlation strength matrix of the preset spatial range and a correlation coefficient of the operational intensity matrix may be calculated.
  • the correlation coefficient can be calculated in the manner described in the above embodiments, and will not be described in detail.
  • the correlation coefficient may be compared with a preset at least two coefficient intervals to determine a coefficient interval to which the correlation coefficient belongs.
  • the at least two coefficient intervals respectively correspond to different degrees of rationality.
  • several coefficient intervals can be set as follows.
  • the correlation coefficient ⁇ is located in the interval, it means that the distribution of the surveillance camera has a complete correlation with the distribution of the operational intensity, and the camera distribution is very reasonable.
  • Interval 2 0.8 ⁇ ⁇ ⁇ 1.
  • Interval 3 0.3 ⁇ ⁇ ⁇ 0.8.
  • Interval 4 0 ⁇ ⁇ ⁇ 0.3.
  • Interval 5 ⁇ ⁇ 0.
  • the correlation coefficient ⁇ is in the interval, it can be considered that the distribution of the camera is very unreasonable. For example, in areas where the camera is densely distributed, the operation is infrequent, and the camera is basically idle; or, in areas where the camera is less distributed, it is an important area of interest, and the camera is used at a high frequency. Therefore, the position layout of the camera needs to be improved.
  • the correlation coefficient calculated in step 810 is compared with the above several coefficient intervals, and it is determined which coefficient interval the correlation coefficient is located.
  • step 830 the degree of rationality corresponding to the coefficient interval to which the correlation coefficient belongs is determined as the degree of rationality of the camera distribution in the preset spatial range.
  • the calculated correlation coefficient is 0.6
  • the 0.6 is located in the interval 3 listed in step 820, and the degree of rationality corresponding to the interval 3 is moderately reasonable, it is considered that the camera distribution is moderately reasonable and needs to be improved.
  • the calculated correlation coefficient is 0.9, and the 0.9 is located in the interval 2 listed in step 820, it can be considered that the camera distribution is reasonable, and the distribution and operation of the camera substantially conform to objective requirements.
  • the analysis method of the camera distribution rationality in this embodiment combines the operation intensity of the camera with the spatial monitoring distribution of the camera, and can evaluate whether the distribution of the camera is reasonable based on the correlation between the operation intensity and the spatial distribution. This method is consistent with the characteristics of the monitoring strength and the operational strength, so that more accurate and reasonable analysis results can be obtained. Moreover, by comparing the correlation coefficient with a preset number of coefficient intervals to analyze the rationality, the evaluation of the rationality of the camera distribution can be made more detailed and accurate.
  • step 910 a monitoring intensity matrix and an operational intensity matrix of the preset spatial extent are calculated.
  • the monitoring intensity matrix may include monitoring intensity factors respectively for indicating camera monitoring strengths at respective spatial locations;
  • the operational strength matrix may include operating intensity factors respectively for indicating camera operating strength at respective spatial locations.
  • step 920 the monitoring intensity factor and the operation intensity factor of the same spatial location are superimposed to obtain a superposition factor, and the superposition factors of all spatial locations within the preset spatial range constitute a camera distribution sparse matrix.
  • the superposition operation of this step can be realized by spatial superposition analysis, that is, superimposing the monitoring intensity factor and the operation intensity factor of the same spatial position in the preset spatial range, and obtaining the camera distribution at the same spatial position relative to the camera.
  • the result of the overlay of the operation is also referred to as a superposition factor.
  • the preset spatial extent may be divided in a grid form.
  • both the monitoring intensity matrix and the operating intensity factor in the monitoring intensity matrix and the operating intensity matrix are minimized, that is, by scaling the data into a specific interval, such as [-1, +1] , [0, 1], etc.
  • the value of each factor reflects the relative amount of the current position relative to the maximum value of the region and is a dimensionless value.
  • the two-dimensional array index of the two matrices represents the same geographic location. Therefore, when performing spatial superposition analysis, the data of the two matrices can be subtracted.
  • the same spatial position ie, corresponding to the preset space range
  • the monitoring intensity factor of the position of the same two-dimensional array index is subtracted from the operating intensity factor to obtain a superposition factor.
  • the monitoring intensity factor and the operating intensity factor are both greater than 0 and less than 1.
  • the calculations of other grids are similar, and each grid can get a corresponding superposition factor.
  • the matrix of the overall composition of each superposition factor can be referred to as a camera distribution sparse matrix, that is, the point distribution sparse graph in FIG.
  • step 930 the degree of rationality of the camera distribution of each spatial position of the preset spatial range is analyzed according to the camera distribution sparse matrix.
  • the analysis when analyzing the degree of rationality of the camera distribution at each spatial position, the analysis may be performed according to the superposition factor at the corresponding grid position of each two-dimensional array index. Moreover, a threshold may be preset, and the superposition factor obtained in step 920 is compared with the preset threshold to analyze the rationality of the camera distribution at each grid position.
  • a superposition factor in the camera distribution sparse matrix is greater than a preset threshold, it may indicate that the camera distribution density at the corresponding spatial location is too dense relative to the camera operation strength. If a superposition factor in the camera distribution sparse matrix is less than a preset threshold, it may indicate that the camera distribution density at the corresponding spatial position is too thin relative to the camera operation strength.
  • the superposition factor is greater than 0, it indicates that the camera distribution density at the spatial position is too dense with respect to the camera operation intensity; if the superposition factor is less than 0, it indicates the camera distribution at the spatial position.
  • the density is too thin relative to the camera's operating intensity.
  • step 940 may be further performed, according to the superposition factor in the camera distribution sparse matrix, The spatial position in the map of the spatial range is rendered by different colors, and a sparse and reasonable map of the monitoring points of the preset spatial range is obtained.
  • the sparsely reasonable map of the monitoring points can be obtained by respectively rendering the grids corresponding to the superposition factors of >0 and ⁇ 0 into different colors.
  • the red dot area may be indicated in red and the blue area may be over-densified.
  • the corresponding monitoring intensity matrix and the operation intensity matrix are obtained, and different colors are rendered according to the data in the matrix, so that the distribution of the cameras in the map can be clearly displayed and Usage.
  • the data standardization method in the embodiment of the present application is not limited to the minimum value processing method, and other methods may be employed.
  • the influence force F is an empirical function of attenuation with distance, and can also be supplemented by other factors to optimize the calculation formula of the influence force F.
  • the manner of the superposition operation is not limited to subtracting the monitoring intensity factor from the corresponding operation intensity factor.
  • the embodiment further provides an analysis device for reasonable camera distribution, the device can be implemented as a monitoring software, and the device can be the video shown in FIG.
  • the function module integration corresponding to the monitoring point control logic is monitored.
  • the apparatus may include a sample point determination module 1310, a rasterization processing module 1320, a sample point assignment module 1330, a grid unit processing module 1340, a dimensionless processing module 1350, and a rendering display module 1360.
  • the apparatus can include a monitor strength module 1110, an operational strength module 1120, and a distribution analysis module 1130.
  • the monitoring strength module 1110 is configured to acquire, according to the camera distribution location data of the preset spatial range, a monitoring strength matrix for indicating camera monitoring strength at different spatial locations within the preset spatial range.
  • the operation strength module 1120 is configured to acquire, according to the camera operation log data of the preset spatial range, an operation intensity matrix for indicating a camera operation intensity of different spatial locations within the preset spatial range.
  • the distribution analysis module 1130 is configured to calculate a correlation coefficient between the monitoring intensity matrix and the operation intensity matrix, and analyze a degree of rationality of the camera distribution in the preset spatial range according to the correlation coefficient.
  • the distribution analysis module 1130 may be configured to perform the correlation coefficient and the preset at least two coefficient intervals when analyzing the degree of rationality of the camera distribution in the preset spatial range according to the correlation coefficient. The comparison determines the coefficient interval to which the correlation coefficient belongs, and determines the degree of rationality corresponding to the coefficient interval to which the correlation coefficient belongs as the degree of rationality of the camera distribution in the preset spatial range.
  • the analysis apparatus may further include a superposition analysis module 1210, configured to superimpose a monitoring intensity factor and an operation intensity factor for the same spatial location in the preset spatial range to obtain a superposition factor.
  • the superposition factors of all the spatial positions in the preset spatial range constitute a camera distribution sparse matrix, and the degree of rationality of the camera distribution of each spatial position of the preset spatial range is analyzed according to the camera distribution sparse matrix.
  • the monitoring strength matrix may include a monitoring intensity factor for respectively indicating camera monitoring intensity of each spatial position
  • the operating intensity matrix may include an operating intensity factor for indicating camera operating intensity of each spatial position.
  • the overlay analysis module 1210 may include a data operation unit 1211 and a comparison analysis unit 1212.
  • the data operation unit 1211 is configured to subtract the operation intensity factor from the monitoring intensity factor for the same spatial location in the preset space range to obtain a superposition factor.
  • Comparative analysis Element 1212 is configured to analyze a degree of rationality of camera distribution at each spatial location based on the overlay factor.
  • a superposition factor in the camera distribution sparse matrix is greater than a preset threshold, it may indicate that the camera distribution density at the corresponding spatial position is too dense relative to the camera operation strength; if the camera distributes a superposition factor in the sparse matrix If it is less than the preset threshold, it may indicate that the camera distribution density at the corresponding spatial position is too thin with respect to the camera operation intensity.
  • the monitoring software for performing the analysis method of the camera distribution rationality of the present application may include only the data operation unit 1211, so that the superposition factor can be displayed in the software.
  • the overlap factor obtained by subtracting the monitoring intensity factor and the operation intensity factor corresponding to a certain grid is 0.213, and the result data may be displayed on the map on the map, or the range of the result data is greater than 0,
  • the user himself determines the sparseness of the camera distribution based on the superposition factor.
  • the software may include a data operation unit 1211 and a comparison analysis unit 1212, and the software obtains a sparse distribution result of the camera according to the superposition factor, such as the distribution is too thin, and displays the sparse result in the corresponding grid area. .
  • the analyzing device may further include a graphic display module 1220, configured to perform, by using different colors, corresponding spatial positions in the map of the preset spatial range according to different overlapping factors in the camera distribution sparse matrix Rendering, obtaining a sparse and reasonable graph of the monitoring points of the preset spatial range.
  • a graphic display module 1220 configured to perform, by using different colors, corresponding spatial positions in the map of the preset spatial range according to different overlapping factors in the camera distribution sparse matrix Rendering, obtaining a sparse and reasonable graph of the monitoring points of the preset spatial range.

Abstract

一种视频监控布点方法和装置，包括：确定地图上的多个样本点（310），并基于所述多个样本点的分布区域确定待分析的空间范围；对所述待分析的空间范围进行栅格化处理，得到所述空间范围的网格单元（320）；对各所述样本点进行权重赋值，其中所述权重赋值表示所述样本点处的监控需求大小（330）；以及，根据各所述样本点的权重值以及各所述样本点与各所述网格单元的位置关系，计算得出各所述网格单元的权重值，并根据各所述网格单元的权重值生成所述空间范围的摄像机分布位置数据（340）。

Description

视频监控布点 技术领域
本发明涉及视频监控布点。
背景技术
目前视频监控已经广泛地应用于城市日常管理中。
常用的摄像机布点方法为，将客观世界抽象为点(例如路口、ATM机等)、线(道路)、面(CBD、大型商场等)，然后计算监控点位的覆盖区域，根据覆盖区域是否包含客观世界或者对客观世界包含的程度来对监控点位布局的合理性进行评判，再根据评判的结果调整并得到最终的摄像机布点需求图，对摄像机进行布点。
这种布点方法在评价时侧重于摄像机对监控区域的覆盖度等，例如通过计算摄像机的覆盖度、摄像机对监控区域的覆盖情况等建立评价布点的指标。这个过程通常很复杂，例如摄像机的覆盖情况涉及到摄像机的类型、焦距、转角、镜头、分辨率以及安装高度等，而监控目标又涉及到客观世界中的各种实体。如果再加上对客观世界中的实体的考虑，那么对数据的要求将会翻好几倍，因为客观世界的实体与实体之间有可借鉴性、但是没有可复制性。
这种方式的另外一个缺点就是没有对客观世界的实体的重要性进行判断、区分，即重要区域与非重要区域的布点方式是一样的，这极有可能会造成重要区域的监控点位不足、而非重要区域的监控点位太多。
在评价视频监控布局的合理性时，一个十分关键的指标是是否覆盖关键区域，即在关键的区域尽可能多地布置监控点，以保证对关键区域能够全 方位覆盖地高效监控。在整个监控点位的布局中，一般都会基于被监控的对象给出明确的要求，例如覆盖主要的银行、超市等人流较密集区域。但是将客观事物抽象成点、线和面的数据结构进行基于矢量数据的运算，有可能忽略了客观事物的本质属性。
例如，在监控点位布局中，有可能由于未考虑区域实际情况而导致在视频调取频次高的关键区域布置的摄像机不足，而在一些非关键区域却布置了同样密度的摄像机。这在一定程度上导致了资源浪费，还可能使得关键区域由于摄像机布置不足而监控效果不佳。
相关技术中，对于摄像机分布合理性的分析，可以统计摄像机的操作频次。例如，可计算哪些摄像机的利用频次高于平均值，并认为高于平均值的摄像机是被合理利用的，低于平均值的摄像机则分布不合理。但是这种分析方法有可能出现误判。比如，对于一个重点区域，通常同时部署多个摄像机以实现该区域的全方位覆盖。用户对该区域的监控是通过摄像机群组完成的，可经常查看该摄像机群组的监控视频，但是对于单个摄像机而言可能操作频次并不高。此时如果按照上述的方法统计单个摄像机操作频次，将得出这群组摄像机的利用率较低的分析结果，该分析结果是不符合实际情况的。
发明内容
由于网络地图的应用，目前基础数据，如POI数据、道路网络数据等较过去已经有了十分大的进步，这些数据在客观上都能够映射出监控的需求。在地理信息系统中，有一种栅格数据结构。栅格数据结构是基于栅格模型的一种数据组织形式，是指将空间分割成有规则的网格(以下可称为网格单元)，在各个网格单元上给出相应的属性值来表示对应地理实体的特定属性。栅格数据结构可以很形象的将客观实体对周围的影响表现出来。再结合POI数据，就可以将基于POI数据的监控需求形象的表示出来。
根据本发明的一个方面，提供了一种视频监控布点方法，以基于能够反映客观监控布点需求的样本点数据，生成特定空间范围内的摄像机分布位置数据。
根据本发明的一个示例，该方法可包括：确定地图上的多个样本点；基于所述多个样本点的分布区域确定待分析的空间范围，并对所述空间范围进行栅格化处理以得到所述空间范围的网格单元；对各所述样本点进行权重赋值，其中所述权重赋值表示所述样本点处的监控需求大小；根据各所述样本点的权重值以及各所述样本点与各所述网格单元的位置关系，计算得出各所述网格单元的权重值，并根据各所述网格单元的权重值生成所述空间范围的摄像机分布位置数据。
由于样本点的权重大小反映了客观世界的监控需求，根据各样本点的权重值以及各样本点与空间范围内各空间位置之间的地理位置关系来生成所述空间范围内的摄像机分布位置数据，可使得所述摄像机分布位置数据的具有较高的可信度。
进一步，根据另一个示例，该方法还可包括：根据所述空间范围的摄像机分布位置数据，获取可表示所述空间范围内的不同空间位置处的摄像机监控强度的监控强度矩阵；根据所述空间范围的摄像机操作日志数据，获取可表示所述空间范围内的不同空间位置处的摄像机操作强度的操作强度矩阵；以及，计算所述监控强度矩阵与所述操作强度矩阵的相关系数，并根据所述相关系数分析在所述空间范围内的摄像机分布的合理性程度。
通过计算特定空间范围内的监控强度矩阵和操作强度矩阵之间的相关系数，可以更加准确的判断该空间范围内的摄像机分布是否合理。比如，监控强度矩阵和操作强度矩阵通常是正相关的，如果根据相关系数确定两者不符合应有的关系，则可以确定该空间范围的摄像机分布可能需要进行合理化改善。
此外，根据本发明的另一个方面，提供了一种视频监控布点装置，其包括处理器，所述处理器通过读取并执行存储介质上所存储的与视频监控布点控制逻辑对应的机器可执行指令，来执行以下：确定地图上的多个样本点；基于所述多个样本点的分布区域确定待分析的空间范围，并对所述空间范围进行栅格化处理以得到所述空间范围的网格单元；对各所述样本点进行权重赋值，其中各所述样本点的权重值表示在所述样本点处的监控需求大小；根据各所述样本点的权重值以及与各所述网格单元的位置关系，计算得出各所述网格单元的权重值，并根据各所述网格单元的权重值生成所述空间范围的摄像机分布位置数据。
附图说明
图1为根据本发明一实施例基于区域热度采用渲染色带对DEM数据进行渲染的渲染效果图；
图2为将图1所示渲染效果图与实际地图叠加显示的效果图；
图3A为根据本发明一实施例的确定视频监控布点的方法的示意性流程图；
图3B为根据本发明一实施例的确定视频监控布点的装置的功能模块示意图；
图4是根据本申请一示例性实施例示出的一种分析摄像机分布合理性的方法的示意性流程图；
图5是根据本申请一示例性实施例示出的一种监控设备的硬件结构图；
图6是根据本申请一示例性实施例示出的空间网格划分示意图；
图7是根据本申请一示例性实施例示出的基于监控强度因子对预设空间区域进行渲染的效果示意图；
图8是根据本申请一示例性实施例示出的一种根据监控强度矩阵与操作强度矩阵之间的相关系数评价摄像机分布合理性的方法的示意性流程图；
图9是根据本申请另一示例性实施例示出的评价摄像机分布合理性的方法的示意性流程图；
图10是根据本申请一示例性实施例示出的叠合运算的示意图；
图11是根据本申请一示例性实施例示出的一种摄像机分布合理性的分析装置的功能模块框图；
图12是根据本申请另一示例性实施例示出的一种摄像机分布合理性的分析装置的功能模块框图。
具体实施方式
这里将结合附图详细地对示例性实施例进行说明。下面的描述涉及附图时，除非另有表示，不同附图中的相同数字表示相同或相似的要素。以下示例性实施例中所描述的实施方式并不代表与本申请相一致的所有实施方式。相反，它们仅是与如所附权利要求书中所详述的、本申请的一些方面相一致的装置和方法的例子。
根据本发明一实施例的视频监控布点方法可如图3A所示，包括如下步骤310-340。
在步骤310：确定地图上的多个样本点。
步骤310可由图3B中所示的样本点确定模块1310进行。
样本点的权重大小反映了样本点的数据热度，即重要程度。其中样本点可以为POI点(Point of Interesting，兴趣点)，或者是根据其他分布规律生成的点。例如，根据有案件发生的位置数据，可以得到案件发生分布规律；根据已有的摄像机点位，可以得到摄像机点位布点规律。根据这些分布规律， 可以得到一定程度上能够反映客观世界监控布点需求的样本点。
换言之，样本点的选取可有多种方式。例如，案件发生的地点分布、已有的摄像机点位分布等等均能够反映监控的布点需求，可以以这类地点对应在地图上的点作为样本点进行提取。这样，POI的分布能较全面地反映监控地区的人流量，例如，POI密度越密，则代表该区域热度越高，将是监控的重点，需要重点部署，反之则表示重要程度较低。
进一步而言，所述的样本点为地图上的兴趣点。现有的网络地图一般都带有兴趣点，兴趣点能够比较全面准确地反映人流量在地图区域中的分布。
在步骤320：基于所述多个样本点的分布区域确定待分析的空间范围，并对所述空间范围进行栅格化处理以得到所述空间范围的网格单元。
其中，栅格化处理可包括，生成与地图上的所述空间范围对应的栅格数据，以能够基于网格单元对所述空间范围内的各个空间位置处的经纬度信息和属性值信息进行处理。例如，所述栅格数据可为二维数组，所述二维数组中的各个元素可与地图区域中的各个网格单元一一对应。所谓一一对应，可包括：所述二维数组中的各个元素的索引可表示对应网格单元的经纬度值，所述二维数组中的各个元素的数值可表示对应网格单元的权重值。
步骤320可由栅格化处理模块1320进行。
数字高程模型(Digital Elevation Model，简称DEM)是用一组有序数值以阵列形式表示地面高程的一种实体地面模型。为了便于操作，可采用DEM形式的栅格数据(以下可简称为DEM数据)，其网格单元的索引可用于表示地图上对应区块的经纬度，而网格单元的值可以表示自定义的属性。例如，为了显示布点需求，网格单元的值可定义为网格单元的权重值。
在以DEM数据为基础将样本点进行栅格化处理之前，可将经纬度转换为DEM数据的索引，以方便进行后续的赋值和显示操作。根据一个例子，将经纬度转换为DEM数据的索引的具体方法可包括如下的步骤320a和步骤 320b，以将需要进行栅格化处理的地图区域缩小到一个合理的范围以减少操作的复杂度，并同时包含所有的样本点。其中，为了便于操作，可将样本点的经纬度值转换为墨卡托投影坐标系中的坐标值。
在步骤320a：确定所有的样本点分布在经度上最大值和最小值以及在纬度上的最大值及最小值。
提取POI，将POI的经纬度转换为墨卡托投影坐标系的坐标，转换公式可具体如下：
Figure PCTCN2016081736-appb-000001
y1＝ln(tan((90+Y)*π/360))/(π/180)；
Figure PCTCN2016081736-appb-000002
其中，X为POI的经度，Y为POI的纬度。全世界墨卡托投影坐标范围是(-20037508，-20037508，20037508，20037508.34)。
计算整体POI在x方向上的最小值x min和最大值x max以及在y方向上的最小值y min和最大值y max，得到POI数据区域的分布跨度：
width＝x max-x min；
height＝y max-y min。
其中，width表示POI分布区域的宽度，height表示POI分布区域的高度。在墨卡托投影坐标系中，确定POI在x方向上的最小值x min及最大值x max对应地确定了POI在经度上的最小值及最大值，同理，确定POI在y方向上的最小值y min及最大值y max对应地确定了POI在纬度上的最小值及最大值。这样，经纬度上的最大值及最小值可分别转换为墨卡托投影坐标系中对应的坐标值，包括：经度上的最大值及最小值对应墨卡托投影坐标系中x方向上的最大值及最小值，纬度上的最大值及最小值对应墨卡托投影坐标系中y方向 上的最大值及最小值。
在步骤320b：将以步骤320a中所确定的经纬度的最大值和最小值所在地图位置处为四个顶点形成的矩形区域确定为需要栅格化的地图区域，也即待分析的空间范围。
基于经纬度上的最大值及最小值在墨卡托投影坐标系中对应的坐标值所确定的四个顶点，可以形成一个呈例如矩形的区域，该区域即为需要栅格化的地图区域。
然后，栅格数据生成模块可根据计算机的内存大小和计算精度需求在内存中申请一个与POI分布区域的宽高比例相同的二维数组DemData。DemData可表示DEM数据，其中数组的二维索引(即数组中的行列位置)对应具体的墨卡托投影坐标值，数组中的各个值表示该区域的重要程度。当前实施例中使用的POI数据来自于杭州主城区，所设置的DEM数据的宽度DemWidth可设置为5000。一般来说，将DEM数据的宽度DemWidth设置为5000，内存大小在2G以上的电脑都可以接受。此外，DEM数据的高度DemHeight可为：
Figure PCTCN2016081736-appb-000003
将所有的POI的墨卡托投影坐标转换为DEM数据中的二维数组索引，具体公式可为如下：
Figure PCTCN2016081736-appb-000004
Figure PCTCN2016081736-appb-000005
式中，方括号表示取整运算。xindex、yindex分别为元素在二维数组中的索引。二维数组形式的DEM数据中的各个元素与网格单元一一对应，每个元素在二维数组中的索引对应网格单元的经纬度值，各个元素的数值表示网 格单元的权重值。
在步骤330：对各所述样本点进行权重赋值，其中所述权重赋值表示所述样本点处的监控需求大小。
步骤330可由样本点赋值模块1330完成。需要说明的是，该步骤既可以在步骤310之后并且在步骤320之前进行，也可以在步骤320之后并且在步骤340之前进行，还可以与步骤320同时进行。当前实施例中，可在步骤320完成栅格化处理之后进行步骤330。事实上，只要步骤330能够在步骤340之前完成即可。
由于样本点在地图中所反映的地理位置不同，例如表示市政府、学校或者娱乐场所，各个样本点的监控需求可能不同。因此，可对样本点进行分类，以能够根据类别的重要程度对样本点赋予不同的权重。
在对不同类别的样本点赋予不同的权重值时，可考虑对布点需求的影响因素，所述影响因素包括但不限于：样本点的地理位置、各类别样本点的人流量以及公众对各类样本点的需求大小。其中，地理位置包括行政划分中处于中心还是郊区。一般来说，中心位置与郊区相比，权重更大。对于人流量而言，人流量越大，则权重越大。例如，商场、医院等类别的样本点处的人流量普遍较大，而农场、田地等类别的样本点处的人流量普遍较小。此外，公众对各类样本点的需求是指某类样本点的必要性，必要性越大，则权重越大。例如，表示城市交通主干道的样本点处的公众需求较大，则权重也对应较大。然而，一些可替代的营业点，例如彩票售卖处，公众需求较小，则权重也就较小。
进一步而言，样本点的分类可具有多个层次。对应地，可利用层次分析法为样本点赋予对应的权重。例如，一些样本点所反映的地理位置可能同属一个较大类别，而较大类别下面细分为多个层次的较小类别。例如，较大类别为政府机构，该较大类别下还包括中央政府、地方政府及基层组织等较 小类别。在分为多个层次后，可通过对每个最小类别多次赋值来进行层次分析，从而可得出一个较能反映实际重要程度的权重。
根据一个例子，对样本点进行权重赋值可具体包括，对POI进行分类，并按类别进行权重赋值。
例如，按照现有的POI分类对照表将POI进行分类，并利用层次分析法对POI进行权重的赋值。如表1所示，其中表1表示POI的分类及影响值：
表1
Figure PCTCN2016081736-appb-000006
Figure PCTCN2016081736-appb-000007
其中，POI的权重值可由例如三位专家分别根据自身对该POI的重要程度的认识进行赋值。例如，x1、x2及x3分别表示三位专家对样本点的权重赋值。此外，可限定最大的权重值为例如30。然后，在A-M大分类下，分别对各小分类进行层次分析法，得出每个因子的值。
以A类为例，为了构造判断矩阵，引入标度aij，标度aij的不同取值对应的含义可如表2所示。
表2
标度aij 定义
1 i因素与j因素同等重要
3 i因素比j因素略重要
5 i因素比j因素较重要
7 i因素比j因素非常重要
9 i因素比j因素绝对重要
2、4、6、8 为以上判断之间的中间状态对应的标度值
倒数 若i因素与j因素比较，得到判断值为：aji＝1/aij，aii＝1
对于矩阵A而言，四个因子a1、a2、a3、a4的两两比较矩阵可如表3所示。其中，四个因子a1、a2、a3、a4分别表示一级分类A下的四个二级分类。
表3
G a1 a2 a3 a4
a1 1 3 5 7
a2 1/3 1 3 5
a3 1/5 1/3 1 3
a4 1/7 1/5 1/3 1
此外，不同专家针对同一因子的重要程度赋值的两两比较矩阵可如表4所示：
表4
a1 x1 x2 x3 a2 x1 x2 x3 a3 x1 x2 x3 a4 x1 x2 x3
x1 1 1 1 x1 1 1 3 x1 1 1 2 x1 1 1/2 1
x2 1 1 1 x2 1 1 3 x2 1 1 3 x2 2 1 2
x3 1 1 1 x3 1/3 1/3 1 x3 1/2 1/3 1 x3 1 1/2 1
在得到层次排序后，可进行如下操作：
(1)得到各层次的判断矩阵，并进行一致性检验。
首先，可在第一层次上对四个因子的两两比较矩阵进行一致性检验。
例如，根据表3可以得到判断矩阵A，在对判断矩阵A进行一致性检验之前，可对其进行例如归一化的标准化处理。
Figure PCTCN2016081736-appb-000008
Figure PCTCN2016081736-appb-000009
Figure PCTCN2016081736-appb-000010
其中操作①是对判断矩阵A进行列向量归一化；操作②是按行求和；③为最终得到的归一化矩阵。
在得到归一化矩阵后，可获得判断矩阵A的最大特征值
Figure PCTCN2016081736-appb-000011
以及特征向量W(0)。
Figure PCTCN2016081736-appb-000012
Figure PCTCN2016081736-appb-000013
W(0)＝(0.558，0.263，0.122，0.057)T
接着，可在第二层次上对相同因子之间的两两比较矩阵进行一致性检验。同理，可计算出根据表4所得的判断矩阵A1、A2、A3、A4对应的最大特征值与特征向量，依次为：
Figure PCTCN2016081736-appb-000014
Figure PCTCN2016081736-appb-000015
Figure PCTCN2016081736-appb-000016
Figure PCTCN2016081736-appb-000017
其中，A1表示各个专家对二级类别a1的赋值x1、x2及x3之间的两两比较矩阵。同理，A2、A3、A4分别表示各个专家对二级类别a2、a3、a4的赋值x1、x2及x3之间的两两比较矩阵。对应地，
Figure PCTCN2016081736-appb-000018
以及
Figure PCTCN2016081736-appb-000019
分别为判断矩阵A1、A2、A3及A4的最大特征值，W(1)、W(2)、W(3)以及W(4)表示判断矩阵A1、A2、A3及A4的特征向量。
可采用一致性指标进行检验：
Figure PCTCN2016081736-appb-000020
其中，对于各个矩阵阶数n，对应的RI值可如下表：
n 1 2 3 4 5 6 7 8 9
RI 0 0 0.58 0.90 1.12 1.24 1.32 1.41 1.49
对于判断矩阵A，阶数n为4，对应的RI值为0.90，从而可得到CR＝0.044。CR小于0.1，可表示判断矩阵A的不一致程度在容许范围内。如未通过一致性验证，则可对样本点进行重新分类或者直接对样本点进行重新赋值，并可用判断矩阵A的特征向量代替权向量。同理，对于判断矩阵A1、A2、A3、A4也可各自利用上述原理进行一致性检验。
(2)求出权向量矩阵，对各个POI进行权重的赋值。
Figure PCTCN2016081736-appb-000021
Figure PCTCN2016081736-appb-000022
根据权向量矩阵，结合三位专家给出的影响值X1、X2、X3，就可以算出一个综合的值：
a1＝30*0.360+30*0.382+30*0.258＝30；
a2＝25*0.360+25*0.382+20*0.258＝23.71；
a3＝20*0.360+20*0.382+15*0.258＝18.71；
a4＝10*0.360+15*0.382+10*0.258＝11.91。
根据这个方法即可求出B-M中各个因子的值。
由步骤310可知，每个POI都能够在DEM数据中找到唯一的元素索引xindex、yindex与其经纬度对应。POI表示人们的兴趣点，在POI点及其附近都是人们可能关心的区域。因此，距离POI越近代表该区域越重要，监控需求越强烈；并且区域内的POI密度越密代表该区域热度越高，例如可能是人流量聚集区域，需要重点部署监控，反之亦然。
在空间上，POI对周边的影响具有随距离衰减的特性。所以，本发明当前实施例中，可通过设置影响范围以及最大影响值来模拟影响随距离衰减。
在步骤340，根据各所述样本点的权重值以及与各所述网格单元的位置关系，计算得出各所述网格单元的权重值，并根据各所述网格单元的权重值生成所述空间范围的摄像机分布位置数据。
该步骤可由网格单元赋值模块1340完成。由于一个样本点对周围网格均有影响，且影响可随着距离的增加而衰减，因此，可根据各样本点的权重值以及与各网格单元的位置关系得出各网格单元的权重值，以能够较真实地反映样本点对周围网格的影响。例如，假设POI的影响值为在步骤330所求出的Influence，影响最大距离为MaxDis tan ce。通过步骤320中的x max、x min、y max、y min可以计算得出当前POI所在网格单元所对应的区域范围xwidth、ywidth(单位可为米)。这样，通过与DEM数据的宽度DemWidth、DEM数据的高度DemHeight结合进行计算，可以得出DEM数据的分辨率如下：
Figure PCTCN2016081736-appb-000023
Figure PCTCN2016081736-appb-000024
其中，DEM数据的分辨率可表示每个网格单元的大小。这样，可指定单个样本点在地图上的最大影响距离，并可根据以下公式来确定单个样本点对其影响范围内的各网格单元的权重分量：
Figure PCTCN2016081736-appb-000025
其中，Dis tan ce表示样本点到网格单元的距离。对于样本点所在的网格单元，距离Dis tan ce为0。Influence表示样本点的权重值。各网格单元的权重值为所有样本点对该网格单元的权重分量之和。网格单元的权重值可在DEM数据中可用所对应元素的数值表示。
各个网格单元的权重值反映了网格单元所对应的地图区域的监控需求。为了将网格单元的权重值能够直观地显示在地图上，可以对DEM数据进行渲染。例如，步骤340还可包括以下操作：将网格单元的权重值进行无量纲处理，得到权重无量纲化栅格数据。
对DEM数据中网格单元所对应的元素的数值进行无量纲处理，可由无量纲处理模块1350完成。由于网格单元的权重值是基于人为设置的样本点的权重值计算得到的，可对网格单元的权重值进行无量纲处理，以得到包含所有网格单元的无量纲权重的权重无量纲化栅格数据，从而能够相对准确地反映客观世界的监控需求。例如，考虑到渲染的需要，可选择无量纲处理为极小化处理，使所有权重值在处理后处于0-1之间。具体处理方式可为，将DEM数据中的所有数值除以其中的最大值，所得到的各个数值表示相对于本研究区域中的监控需求最强烈的区域的重要程度，其中越接近1表示越重要，越接近0表示越不重要。
可利用预生成的渲染色带，通过渲染显示模块1360将极小化处理后的DEM数据进行渲染得到渲染图。例如，可根据网格单元中数据的重要程度使用相应的颜色进行渲染，渲染结果可如图1所示。渲染色带的颜色可以是多种，用于区分不同权重值，在当前实施例中，以白色为起始颜色，黑色为终止颜色，绘制在100*100长宽的图片上。
渲染显示模块1360可将渲染图与地图进行叠合显示，得到监控布点需求图。DEM数据中的每一个网格单元都是能够与真实的经纬度相对应的，所以可以在现有地图中展示。例如，可计算每个网格单元的四角控制点，然后绘制在现有地图中。渲染图与地图的叠合结果、即监控布点需求图可如图2所示。从图中可以看出，在繁华区域，监控的需求会十分强烈，以较深色标识出，而在非主城区，以白色透明色为主。然后，可由渲染显示模块1360基于所述监控布点需求图生成所述空间范围的摄像机分布位置数据。
综上所述，样本点的权重值大小反映了客观世界在该样本点处的监控需求大小，在进行视频监控布点时将区域内所有样本点的权重值考虑进去，可使得监控布点具有较高的可信度。
此外，本申请还提供了一种摄像机分布合理性的分析方法，以评价一个预设空间范围内的监控用摄像机的分布是否合理。
例如，预设的空间范围可以是杭州滨江地区，也可以是北京市海淀区、上海市浦东区等，该方法应用的空间范围用户可以自主设定。
本方法中所要评价的“摄像机分布的合理性”，可以是：在监控需求旺盛的关键区域布置较多的摄像机，而在监控需求较弱的非关键区域布置相对较少的摄像机，以使得摄像机被合理利用且能满足监控的需求。
为了实现对摄像机分布合理性的相对准确的评价，本申请基于如下摄像机分布的原理：摄像机的监控强度与摄像机的操作强度存在着较高的正相关性。例如，在用户更加关心的关键区域，摄像机布置较多，这个区域的监 控强度较大，而且用户对该区域的视频录像的查看次数较多；而对于用户不关心的非关键区域，摄像机布置较少，相应的监控强度低，用户也很少查看该区域的视频录像。因此，可以通过查看某个空间范围内的摄像机监控强度和操作强度的关系，来判断摄像机的布置是否合理，这种分析方式是将摄像机的操作与空间信息进行结合。
图4示例了本申请的摄像机分布合理性的分析方法的流程。如图4所示，该方法可以包括步骤410-步骤430。
在步骤410中，根据预设空间范围的摄像机分布位置数据，获取用于表示所述预设空间范围内的不同空间位置处的摄像机监控强度的监控强度矩阵。
在步骤420中，根据所述预设空间范围的摄像机操作日志数据，获取用于表示所述预设空间范围内的不同空间位置处的摄像机操作强度的操作强度矩阵。
在步骤430中，计算所述监控强度矩阵与所述操作强度矩阵的相关系数，并根据所述相关系数分析在所述预设空间范围内的摄像机分布的合理性程度。
上述方法可以实现为由软件执行，例如可以通过某个监控软件分析监控用摄像机的分布是否合理。该监控软件可以是运行在物理设备上，该物理设备例如可以是监控设备。以监控设备为例，如图5所示，该监控设备可以包括处理器(Processor)510、内存(Memory)520、非易失性存储介质(Non-volatile storage)530以及网络接口(Network interface)540。其中，这些硬件可通过总线(Internal bus)550相互连接。在这个例子中，处理器510可以将非易失性存储介质530中存储的视频监控布点控制逻辑读取到内存520中运行，以执行图3所示的确定摄像机分布位置的方法流程和/或图4所示的分析摄像机分布合理性的方法流程。
在步骤410中，可以将预设空间范围的摄像机分布位置数据(也可以称为监控点位数据)导入到该软件中，用于获取监控强度矩阵。
图6示例了一个预设空间范围的部分区域。如上所述，可以对预设空间范围进行栅格化处理，以将该预设的空间范围分割成有规则的网格(以下也可以称为栅格单元)。如图6所示，示出了部分分割后的空间。可以对各个网格赋予属性值，在本申请的步骤410中，网格的属性值可以称为“监控强度因子”。比如，图6中的网格w1(用交叉线表示)可以计算出一个对应的监控强度因子，表示在该网格w1的摄像机监控强度。例如，在将摄像机分布位置数据导入到监控软件后，摄像机就分布在该网格形式的空间范围中，在网格w1的周围可以分布有三个摄像机，分别为CA1、CA2和CA3，而网格w1的监控强度因子可以用于表示周边的三个摄像机CA1、CA2和CA3对网格w1产生多大的监控效果。
并且，每个网格可以对应实际地图空间位置中的一个经纬度坐标，可以将该经纬度坐标转换为DEM(Digital Elevation Model，数字高程模型)中的二维数组索引。比如，网格w1可以用一个二维数组(x，y)索引来标识，而该索引对应着网格w1所代表的实际空间位置的经纬度坐标。总之，本例子对预设空间范围所做的处理包括，将预设空间进行栅格化分割以后，每个网格都用一个二维数组索引标示，该索引是由该网格的实际空间经纬度坐标转换而来，例如可利用包括如上所述的步骤320a和步骤320b的转换方法。而每个网格可以赋予属性值，属性值即上述的一个对应网格的监控强度因子，用于表示该网格处的摄像机监控强度。
以网格w1为例，来说明监控强度因子的计算：可以先分别计算其周边的三个摄像机中的每个摄像机对于该网格的影响力F，然后将这三个摄像机的影响力相加即为网格对应的摄像机监控强度。其中，影响力F是考虑到摄像机的监控效果随距离衰减的特性的表示函数。比如，摄像机是对周边区域进行监控的，距离摄像机越近监控效果越好，距离摄像机越远监控效果越差。 影响力F的计算公式可表示如下：
Figure PCTCN2016081736-appb-000026
其中，F为影响力，Dis为距离摄像机的距离，本例子中可以是网格w1的中心距离摄像机的距离。Value为中心点的影响力。一般来说，摄像机所在的位置就是中心点，因为这里通常是监控效果最好的位置。本例子中，可以将中心点影响力Value设置为1000。DisMax为摄像机的影响范围，例如可以设置为500米。其中，DisMax和Value的数值可以根据不同的摄像机更改。
例如，在计算摄像机CA1对于网格w1的影响力F1时，距离Dis为网格w1与摄像机CA1之间的距离；在计算摄像机CA2对于网格w1的影响力F2时，距离Dis为网格w1与摄像机CA2之间的距离；在计算摄像机CA3对于网格w1的影响力F3时，距离Dis为网格w1与摄像机CA3之间的距离。最终网格w1的监控强度因子Y1可计算如下：
Y1＝F1+F2+F3。
类似的方式可以计算图6的空间范围内的其他网格的监控强度因子。例如，可以计算网格w2的监控强度因子Y2，网格w3的监控强度因子Y3等。那么由图6中的各个网格对应的监控强度因子组成的整体可以称为“监控强度矩阵”，该监控强度矩阵中包括多个监控强度因子，每个因子用于表示其中一个网格的摄像机监控强度。
此外，监控强度因子是一个无量纲的值，可以将监控强度因子进行标准化处理，例如进行极小值标准化，从而将监控强度矩阵中的各个监控强度因子的取值都转换为0～1范围内的值。并且，还可以将监控强度矩阵中根据监控强度因子的取值不同用不同的颜色进行渲染，例如，监控强度因子高的网格用深颜色，监控强度因子低的网格用浅颜色，并按照渐变色的方式渲染。这样，就可以生成如图7示例的预设空间范围的监控强度图，其中颜色深的区域表示摄像机的监控强度较高，也即表示这些区域的摄像机布置密度可能 较高。
在步骤420中，可以根据摄像机操作日志数据生成操作强度矩阵。该操作强度矩阵的生成可仍然是依据在步骤410中提到的栅格化处理后的预设空间范围，只是将网格的属性值由步骤410中的监控强度因子更换为操作强度因子。预设空间范围的各个网格的操作强度因子的整体组成操作强度矩阵。其中，需要说明的是，步骤410和步骤420并没有严格的执行顺序，这两个步骤可以并行操作，也可以先执行步骤410或者步骤420。
摄像机操作日志数据可以是存储在日志服务器中的记录。当用户对摄像机进行操作时，例如查看摄像机的录像视频、调用摄像机的实时视频、进行摄像机实时视频抓拍等，日志服务器通常会记录用户对摄像机的操作，包括操作时间、操作人以及操作类型。本实施例可以将日志服务器记录的这些数据导入用于执行本实施例分析方法的软件中。
在本例子中，可以将日志服务器的记录数据进行数据归类。例如，日志服务器中保存了多个摄像机的操作记录，通过数据归类可以将每个摄像机的操作记录分类整理，得到每个摄像机的操作次数。示例性的，本例子中的软件可以向日志服务器发送数据获取请求，请求获取最近一个月内的摄像机操作记录。这样，待收到日志服务器传输的数据后，可以通过进行数据归类来分别得到各个摄像机在这一个月内的操作次数，并利用操作次数Count计算操作强度矩阵。
本步骤420中，操作强度矩阵的计算原理可与监控强度矩阵类似。具体地，操作强度矩阵中的操作强度因子的计算公式可参见如下：
Figure PCTCN2016081736-appb-000027
按照以上公式可以计算预设空间范围的每个网格对应的操作强度因子。比如，仍以图6中的网格w1为例，可以分别计算w1周边的三个摄像机CA1、CA2和CA3对网格w1的影响力f，并将三个影响力f相加即为网格 w1的操作强度因子，用于表示该网格处的摄像机操作强度。同样的道理，摄像机本身所在的位置可以认为是具有最高的操作强度；并且，随着网格距离摄像机渐远，网格处的操作强度衰减。
本例子中，与监控强度矩阵类似，可以对操作强度因子进行标准化处理，以将各网格处的操作强度因子的取值转换为在0～1之间。并且，类似地，可以将预设空间范围的各个网格按照操作强度因子的不同进行不同颜色的渲染，得到预设空间范围的摄像机操作强度图。
通过上述的摄像机操作强度图，用户可以直观的看到在预设空间范围内哪个区域的摄像机具有较高的操作频率，操作强度大的区域表示是重点监控的关键区域，所以操作需求较为旺盛。而同样的，通过监控强度图可以直观的看到哪个区域布置的摄像机较密，监控强度大的区域表示是重点监控的关键区域。
在步骤430中，可以将上述得到的监控强度矩阵和操作强度矩阵进行相关性分析，得到相关系数。在进行相关计算时，可以设置监控强度矩阵和操作强度矩阵中的相同的二维数组索引代表相同的地理位置。比如，对于监控强度矩阵的某一个二维数组表示的索引与操作强度矩阵的相同的二维数组索引值，两者对应的是预设空间范围内的同一个区域位置。
相关系数的计算方法可详述如下。假设X为监控强度矩阵，Y为操作强度矩阵，可根据协方差计算公式，如下计算得出监控强度矩阵X与操作强度矩阵Y之间的协方差Cov(X，Y)：
Cov(X，Y)＝E{(X-E(X))(Y-E(Y))}。
此处，使用样本均值来表示期望值E(X)、E(Y)。
可根据如下的标准差计算公式计算监控强度矩阵X、操作强度矩阵Y的标准差σX、σY：
其中，u为样本均值，N为样本数。
在分别计算出监控强度矩阵X、操作强度矩阵Y的标准差σX、σY之后，则可以基于这两个矩阵之间的协方差Cov(X，Y)如下计算这两个矩阵之间的相关系数：
Figure PCTCN2016081736-appb-000029
此处计算的相关系数ρ是一个无量纲的值。可以根据计算得到的相关系数来评价在预设空间范围内的摄像机分布的合理性程度。例如，该合理性程度可表示该预设空间范围内的摄像机分布非常合理、较为合理，中度合理、不合理、非常不合理等。
图8示例了一种根据监控强度矩阵与操作强度矩阵之间的相关系数评价摄像机分布的合理性程度的方法的示意性流程图。如图8所示，在步骤810中，可计算预设空间范围的监控强度矩阵和操作强度矩阵的相关系数。该相关系数的计算方式可以如上面实施例描述的方式，不再详述。
在步骤820中，可将所述相关系数与预设的至少两个系数区间进行比较，以确定所述相关系数所属的系数区间。
所述至少两个系数区间分别对应不同的合理性程度。示例性的，可以设置如下几个系数区间。
区间1：相关系数ρ＝1。当相关系数ρ位于该区间，可意味着监控用摄像机的分布与操作强度的分布具有完全相关性，摄像机分布非常合理。
区间2：0.8＜ρ＜1。当相关系数ρ位于该区间，可以认为监控用摄像机的分布与操作强度两者之间高度相关，摄像机分布较为合理，摄像机的分布与操作基本上符合客观要求。
区间3：0.3＜ρ＜0.8。当相关系数ρ位于该区间，可以认为摄像机的分布属于中度合理，可能需要改善。
区间4：0＜ρ＜0.3。当相关系数ρ位于该区间，可认为摄像机的布点不太合理，可能需要改善。
区间5：ρ＜0。当相关系数ρ位于该区间，可以认为摄像机的分布非常不合理。比如，在摄像机分布较密的区域操作不频繁，摄像机基本处于闲置；或者，在摄像机分布较少的区域反而是重要关注区域，摄像机被高频次使用。因此，摄像机的点位布局有待改善。
上述的几个系数区间只是举例，实际实施中也可以有其他划分方式。本步骤中，将在步骤810中计算得到的相关系数与上述几个系数区间比较，判断相关系数位于哪个系数区间。
在步骤830中，将所述相关系数所属系数区间对应的合理性程度，确定作为所述预设空间范围内的摄像机分布的合理性程度。
例如，假设计算出的相关系数为0.6，该0.6位于步骤820中列举的区间3，区间3对应的合理性程度是中度合理，则认为摄像机分布中度合理，待改善。
又例如，假设计算的相关系数为0.9，该0.9位于步骤820中列举的区间2，则可以认为摄像机分布较为合理，摄像机的分布与操作基本上符合客观要求。
本实施例的摄像机分布合理性的分析方法，通过将摄像机的操作强度与摄像机的空间监控分布情况结合起来，可基于操作强度和空间分布的相关性来评价摄像机的分布是否合理。这种方式符合监控强度与操作强度正相关的特性，从而能够得到较为准确的合理性分析结果。并且，通过将相关系数与预设的几个系数区间进行比较来分析合理性，可以使得对摄像机分布合理性的评价更加细化和准确。
在上述的例子中，可基于监控强度矩阵与操作强度矩阵之间的相关系数，从整体上衡量预设空间范围内的监控布局是否合理。然而，更进一步的，还可以分析在预设空间范围内的各个区域的摄像机分布的合理性。比如，通过图4所示的流程，可得到杭州滨江地区整体上的监控布点是合理的。但是，具体到滨江地区的各个区域，摄像机分布是否合理，则可以通过图9所示的流程来进行分析。
在步骤910中，计算预设空间范围的监控强度矩阵和操作强度矩阵。
例如，该步骤中的两个矩阵的计算方式，可以结合参见图4所示的流程中的步骤410和步骤420，不再详述。其中，监控强度矩阵可以包括分别用于表示各个空间位置处的摄像机监控强度的监控强度因子；操作强度矩阵可以包括分别用于表示各个空间位置处的摄像机操作强度的操作强度因子。
在步骤920中，将同一空间位置的监控强度因子和操作强度因子进行叠合运算以得到叠合因子，该预设空间范围内所有空间位置的叠合因子组成摄像机分布稀疏矩阵。
本步骤的叠合运算可以通过空间叠合分析实现，即将预设空间范围内对于同一空间位置的监控强度因子和操作强度因子进行叠合运算，得到在该同一空间位置处的摄像机分布相对于摄像机操作的叠合结果数据。其中，以下将叠合结果数据又称为叠合因子。
例如，如图6所示的，预设空间范围可以以网格形式划分。并且，监控强度矩阵和操作强度矩阵中的监控强度因子和操作强度因子都经过了极小化处理，即通过将数据按照比例缩放而使之落入一个特定区间，如[-1，+1]、[0，1]等。这样，各因子的数值是反映当前位置相对于区域最大值的相对量，是一个无量纲的值。此外，两个矩阵的二维数组索引代表着相同的地理位置。因此，在进行空间叠合分析时，可以对两个矩阵的数据进行相减运算。
参见图10的示例，可以将预设空间范围内对于同一空间位置(即对应 同一二维数组索引的位置)的监控强度因子减去操作强度因子，得到叠合因子。其中，所述监控强度因子和所述操作强度因子均大于0小于1。例如，在图10示例的网格中，监控强度矩阵和操作强度矩阵中的对应同一网格位置的数据相减，以左上角第一个网格为例，“0.2436-0.23123＝0.01237”，0.01237可以称为叠合因子。其他网格的计算类似，每个网格都可以得到一个对应的叠合因子。
各个叠合因子的整体组成的矩阵可以称为摄像机分布稀疏矩阵，也就是图10中的点位分布稀疏图。
在步骤930中，根据摄像机分布稀疏矩阵分析预设空间范围的各个空间位置的摄像机分布的合理性程度。
例如，在分析各个空间位置的摄像机分布的合理性程度时，可以分别根据各个二维数组索引对应的网格位置处的叠合因子来进行分析。并且，可以预先设定一个阈值，将在步骤920中得到的叠合因子与该预设阈值进行比较，来分析各个网格位置处的摄像机分布合理性。
具言之，若摄像机分布稀疏矩阵中的一个叠合因子大于预设阈值，则可表明对应空间位置处的摄像机分布密度相对于摄像机操作强度过密。若摄像机分布稀疏矩阵中的一个叠合因子小于预设阈值，则可表明对应空间位置处的摄像机分布密度相对于摄像机操作强度过稀。
例如，假设预设阈值为0，叠合因子处于-1～+1之间。在这种情况下，如果叠合因子大于0，则表明该空间位置处的摄像机分布密度相对于摄像机操作强度过密；若所述叠合因子小于0，则表明所述空间位置处的摄像机分布密度相对于摄像机操作强度过稀。比如，仍以左上角第一个网格为例，“0.2436-0.23123＝0.01237”，叠合因子0.01237大于0，因此在该网格区域内的摄像机分布密度相对于摄像机操作强度过密，可以适当减少该网格区域内的摄像机数量。
可选的，在步骤930中根据摄像机分布稀疏矩阵分析预设空间范围的各个空间位置的摄像机分布的合理性程度之后，还可以执行步骤940，根据摄像机分布稀疏矩阵中的叠合因子，对预设空间范围的地图中的空间位置用不同的颜色进行渲染，得到所述预设空间范围的监控点位稀疏合理图。
例如，可以通过将＞0和＜0的叠合因子对应的网格分别渲染成不同的颜色，来得到监控点位稀疏合理图。比如，可以用红色表示点位密度欠缺区域，蓝色表示过密区域。在实际的分析中可以看到，蓝色区域有较多摄像机未被操作过，一直处于闲置状态，而越是红色区域，摄像机的视频被调取越是频繁。
通过将摄像机的空间分布和操作日志数据在地图中进行量化，得到对应的监控强度矩阵和操作强度矩阵，并根据矩阵中的数据渲染不同的颜色，可清晰的展示地图中的摄像机的分布情况以及使用情况。并且，可以将地图中的不同网格区域以不同的颜色显示摄像机分布的稀疏情况，清晰的告诉决策者目前的摄像机分布的合理性程度，非常直观。
此外，在本申请实施例中的数据标准化方法，不限于极小值处理法，也可以采用其他方式。影响力F是一个随距离衰减的经验函数，也可以辅以其他因素来优化影响力F的计算公式。并且，在叠合运算中，叠合运算的方式也不限于将监控强度因子和对应的操作强度因子相减。
为了实现上述的摄像机分布合理性的分析方法，本实施例还提供了一种摄像机分布合理性的分析装置，该装置可以实现为一个监控软件，并且该装置可以是与图5中所示的视频监控布点控制逻辑对应的功能模块集成。如图3B所示，该装置可以包括样本点确定模块1310、栅格化处理模块1320、样本点赋值模块1330、网格单元处理模块1340、无量纲处理模块1350和渲染显示模块1360。如图11所示，该装置可以包括监控强度模块1110、操作强度模块1120和分布分析模块1130。
其中，监控强度模块1110，用于根据所述预设空间范围的摄像机分布位置数据，获取用于表示所述预设空间范围内的不同空间位置处的摄像机监控强度的监控强度矩阵。
操作强度模块1120，用于根据所述预设空间范围的摄像机操作日志数据，获取用于表示所述预设空间范围内的不同空间位置的摄像机操作强度的操作强度矩阵。
分布分析模块1130，用于计算所述监控强度矩阵与所述操作强度矩阵的相关系数，并根据所述相关系数分析在所述预设空间范围的摄像机分布的合理性程度。
进一步的，分布分析模块1130可被配置为，在根据所述相关系数分析在所述预设空间范围的摄像机分布的合理性程度时，将所述相关系数与预设的至少两个系数区间进行比较以确定所述相关系数所属的系数区间，并将所述相关系数所属的系数区间对应的合理性程度确定作为所述预设空间范围内的摄像机分布的合理性程度。
参见图12所示，该分析装置还可以包括叠合分析模块1210，用于将所述预设空间范围内对于同一空间位置的监控强度因子和操作强度因子进行叠合运算以得到叠合因子，所述预设空间范围内所有空间位置的叠合因子组成摄像机分布稀疏矩阵，并根据所述摄像机分布稀疏矩阵分析预设空间范围的各个空间位置的摄像机分布的合理性程度。其中，所述监控强度矩阵可包括分别用于表示各个空间位置的摄像机监控强度的监控强度因子，所述操作强度矩阵可包括分别用于表示各个空间位置的摄像机操作强度的操作强度因子。
进一步的，叠合分析模块1210可以包括数据运算单元1211和比较分析单元1212。其中，数据运算单元1211，用于将所述预设空间范围内对于同一空间位置的监控强度因子减去操作强度因子，得到叠合因子。比较分析单 元1212，用于基于所述叠合因子分析各空间位置处的摄像机分布的合理性程度。例如，若所述摄像机分布稀疏矩阵中一个叠合因子大于预设阈值，则可表明对应空间位置处的摄像机分布密度相对于摄像机操作强度过密；若所述摄像机分布稀疏矩阵中一个叠合因子小于预设阈值，则可表明对应空间位置处的摄像机分布密度相对于摄像机操作强度过稀。
以该装置实现为软件为例，用于执行本申请的摄像机分布合理性的分析方法的监控软件，可以只包括数据运算单元1211，这样在软件中可以显示叠合因子。比如，对应某个网格的监控强度因子和操作强度因子相减得到的叠合因子是0.213，可以在地图上该网格处显示该结果数据，或者显示该结果数据的范围是大于0，由用户自己根据叠合因子判断摄像机分布的稀疏。或者还可以是，该软件中包括数据运算单元1211和比较分析单元1212，是由软件根据叠合因子得出摄像机的分布稀疏结果，比如分布过稀，并将稀疏结果显示在对应的网格区域。
进一步的，该分析装置还可以包括图形显示模块1220，用于根据所述摄像机分布稀疏矩阵中不同大小的叠合因子，对所述预设空间范围的地图中对应的空间位置用不同的颜色进行渲染，得到所述预设空间范围的监控点位稀疏合理图。
以上所述仅为本公开的较佳实施例而已，凡在本公开的精神和原则之内进行的任何修改、等同替换、改进等，均应包含在本公开请求保护的范围之内。

Claims (15)

  1. 一种视频监控布点方法，其特征在于，包括：
    确定地图上的多个样本点；
    基于所述多个样本点的分布区域确定待分析的空间范围，并对所述空间范围进行栅格化处理以得到所述空间范围的网格单元；
    对各所述样本点进行权重赋值，其中各所述样本点的权重值表示在各所述样本点处的监控需求大小；
    根据各所述样本点的权重值以及各所述样本点与各所述网格单元的位置关系，计算得出各所述网格单元的权重值，并根据各所述网格单元的权重值生成所述空间范围的摄像机分布位置数据。
  2. 如权利要求1所述的方法，其特征在于，对各所述样本点进行权重赋值，包括：
    将各所述样本点进行分类，
    按各所述样本点的类别为各所述样本点赋予对应的权重，
    其中，在所述样本点的分类具有多个层次的情况下，利用层次分析法为各所述样本点赋予对应的权重。
  3. 如权利要求1所述的方法，其特征在于，根据各所述样本点的权重值以及与各所述网格单元的位置关系，计算得出各所述网格单元的权重值，包括：
    指定单个所述样本点在地图上的最大影响距离，
    根据以下公式确定单个所述样本点对其影响范围内的各网格单元的权重分量，
    Figure PCTCN2016081736-appb-100001
    其中，MaxDis tan ce表示单个所述样本点的最大影响距离，Dis tan ce表示所述样本点到网格单元的距离，Influence表示样本点的权重值；
    针对各所述网格单元，计算所有所述样本点对该网格单元的权重分量之 和作为该网格单元的权重值。
  4. 如权利要求1所述的方法，其特征在于，基于所述样本点的分布区域确定待分析的空间范围，包括：
    确定所有的所述样本点在经度上的最大值和最小值以及在纬度上的最大值和最小值；
    将以所确定的经度上的最大值和最小值以及维度上的最大值和最小值所在地图位置处为四个顶点形成的区域，确定为所述待分析的空间范围。
  5. 如权利要求1所述的方法，其特征在于，根据各所述网格单元的权重值生成所述空间范围的摄像机分布位置数据，包括：
    将各所述网格单元的权重值进行无量纲处理，得到包括各所述网格单元的无量纲权重的权重无量纲化栅格数据；
    对所述权重无量纲化栅格数据进行渲染得到渲染图，其中不同的无量纲权重用不同颜色表示；
    将所述渲染图叠合至地图上的所述空间范围，生成所述空间范围的监控布点需求图，并基于所述监控布点需求图生成所述空间范围的摄像机分布位置数据。
  6. 根据权利要求1所述的方法，其特征在于，还包括：
    根据所述空间范围的摄像机分布位置数据，获取表示所述空间范围内的不同空间位置处的摄像机监控强度的监控强度矩阵；
    根据所述空间范围的摄像机操作日志数据，获取表示所述空间范围内的不同空间位置处的摄像机操作强度的操作强度矩阵；
    计算所述监控强度矩阵与所述操作强度矩阵的相关系数，并根据所述相关系数分析在所述空间范围内的摄像机分布的合理性程度。
  7. 根据权利要求6所述的方法，其特征在于，根据所述相关系数分析在所述空间范围内的摄像机分布的合理性程度，包括：
    将所述相关系数与预设的至少两个系数区间进行比较，以确定所述相关系数所属的系数区间；
    将所述相关系数所属的系数区间对应的合理性程度，确定作为所述空间范围内的摄像机分布的合理性程度。
  8. 根据权利要求6所述的方法，其特征在于，
    所述监控强度矩阵包括分别用于表示所述空间范围内的各个空间位置处的摄像机监控强度的监控强度因子；
    所述操作强度矩阵包括分别用于表示所述空间范围内的各个空间位置处的摄像机操作强度的操作强度因子；
    所述方法还包括：将对应于同一空间位置的所述监控强度因子和所述操作强度因子进行叠合运算，得到所述空间位置处的叠合因子，并根据所述叠合因子分析所述空间范围内的各个空间位置处的摄像机分布的合理性程度。
  9. 根据权利要求8所述的方法，其特征在于，
    所述叠合运算包括：将所述监控强度因子减去所述操作强度因子；
    根据所述叠合因子分析所述空间范围内的各个空间位置处的摄像机分布的合理性程度，包括：
    若所述叠合因子大于预设的阈值，则表明所述空间范围内的对应空间位置处的摄像机分布密度相对于摄像机操作强度而言过密；
    若所述叠合因子小于所述预设的阈值，则表明所述空间范围内的对应空间位置处的摄像机分布密度相对于摄像机操作强度而言过稀。
  10. 一种视频监控布点装置，其特征在于，包括处理器，所述处理器通过读取并执行存储介质上所存储的与视频监控布点控制逻辑对应的机器可执行指令，来执行以下：
    确定地图上的多个样本点；
    基于所述多个样本点的分布区域确定待分析的空间范围，并对所述空间范围进行栅格化处理以得到所述空间范围的网格单元；
    对各所述样本点进行权重赋值，其中各所述样本点的权重值表示在所述样本点处的监控需求大小；
    根据各所述样本点的权重值以及与各所述网格单元的位置关系，计算得 出各所述网格单元的权重值，并根据各所述网格单元的权重值生成所述空间范围的摄像机分布位置数据。
  11. 如权利要求10所述的装置，其特征在于，在根据各所述样本点的权重值以及各所述样本点与各所述网格单元的位置关系，计算得出各所述网格单元的权重值时，所述机器可执行指令促使所述处理器：
    指定单个所述样本点在地图上的最大影响距离，
    根据以下公式确定单个所述样本点对其影响范围内的各网格单元的权重分量，
    Figure PCTCN2016081736-appb-100002
    其中，MaxDis tan ce表示单个所述样本点的最大影响距离，Dis tan ce表示所述样本点到网格单元的距离，Influence表示样本点的权重值；
    针对各所述网格单元，计算所有所述样本点对该网格单元的权重分量之和作为该网格单元的权重值。
  12. 如权利要求10所述的装置，其特征在于，在基于所述样本点的分布区域确定待分析的空间范围时，所述机器可执行指令促使所述处理器：
    确定所有的所述样本点在经度上的最大值和最小值以及在纬度上的最大值和最小值；
    将以所确定的经度上的最大值和最小值以及维度上的最大值和最小值所在地图位置处为四个顶点形成的区域，确定为所述待分析的空间范围。
  13. 如权利要求10所述的装置，其特征在于，在根据各所述网格单元的权重值生成所述空间范围的摄像机分布位置数据时，所述机器可执行指令促使所述处理器：
    将各所述网格单元的权重值进行无量纲处理，得到包括各所述网格单元的无量纲权重的权重无量纲化栅格数据；
    对所述权重无量纲化栅格数据进行渲染得到渲染图，其中不同的无量纲权重用不同颜色表示；
    将所述渲染图叠合至地图上的所述空间范围，生成所述空间范围的监控布点需求图，并基于所述监控布点需求图生成所述空间范围的摄像机分布位置数据。
  14. 根据权利要求10所述的装置，其特征在于，所述机器可执行指令还促使所述处理器：
    根据所述空间范围的摄像机分布位置数据，获取表示所述空间范围内的不同空间位置处的摄像机监控强度的监控强度矩阵；
    根据所述空间范围的摄像机操作日志数据，获取表示所述空间范围内的不同空间位置处的摄像机操作强度的操作强度矩阵；
    计算所述监控强度矩阵与所述操作强度矩阵的相关系数，并根据所述相关系数分析在所述空间范围内的摄像机分布的合理性程度。
  15. 根据权利要求14所述的装置，其特征在于，在根据所述相关系数分析在所述空间范围内的摄像机分布的合理性程度时，所述机器可执行指令促使所述处理器：
    将所述相关系数与预设的至少两个系数区间进行比较，以确定所述相关系数所属的系数区间；
    将所述相关系数所属的系数区间对应的合理性程度，确定作为所述空间范围内的摄像机分布的合理性程度。
PCT/CN2016/081736 2015-05-29 2016-05-11 视频监控布点 WO2016192516A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US15/577,466 US10445348B2 (en) 2015-05-29 2016-05-11 Video monitor layout

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
CN201510288095.6 2015-05-29
CN201510288095.6A CN104899368B (zh) 2015-05-29 2015-05-29 基于数据热度的监控布点需求图生成方法及装置
CN201510540787.5 2015-08-28
CN201510540787.5A CN105159978B (zh) 2015-08-28 2015-08-28 一种相机分布合理性的分析方法和装置

Publications (1)

Publication Number Publication Date
WO2016192516A1 true WO2016192516A1 (zh) 2016-12-08

Family

ID=57440051

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/081736 WO2016192516A1 (zh) 2015-05-29 2016-05-11 视频监控布点

Country Status (2)

Country Link
US (1) US10445348B2 (zh)
WO (1) WO2016192516A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109194930A (zh) * 2018-10-25 2019-01-11 广州益牛科技有限公司 楼宇环境一体式智能管理系统
CN113961827A (zh) * 2021-10-26 2022-01-21 中国人民公安大学 一种区域防控风险的检测方法及装置

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113052486B (zh) * 2021-04-08 2023-02-07 平安国际智慧城市科技股份有限公司 监控点位布局评价方法、装置、电子设备及可读存储介质
CN114037723B (zh) * 2022-01-07 2022-03-29 成都国星宇航科技有限公司 基于dem数据提取山顶点的方法、装置及存储介质
CN114238346B (zh) * 2022-02-25 2022-05-17 浙江力嘉电子科技有限公司 监控密度分析方法、装置、存储介质和电子设备

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101593457A (zh) * 2008-10-23 2009-12-02 中国科学院地理科学与资源研究所 基于区域划分的屏幕饱和度实现方法及装置
CN101673282A (zh) * 2009-10-15 2010-03-17 杭州北海信息系统有限公司 电子地图自动定位系统方法
CN101808230A (zh) * 2009-02-16 2010-08-18 杭州恒生数字设备科技有限公司 一种用于数字视频监控的统一坐标系
CN104899368A (zh) * 2015-05-29 2015-09-09 浙江宇视科技有限公司 基于数据热度的监控布点需求图生成方法及装置
CN105159978A (zh) * 2015-08-28 2015-12-16 浙江宇视科技有限公司 一种相机分布合理性的分析方法和装置

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP5269002B2 (ja) * 2010-06-28 2013-08-21 株式会社日立製作所 カメラ配置決定支援装置
CN101968912B (zh) 2010-10-28 2012-11-07 重庆市科学技术研究院 森林火灾监测与防控系统及其控制方法
CN102867086B (zh) 2012-09-10 2014-06-25 安科智慧城市技术（中国）有限公司 一种监控摄像机的自动部署方法、系统及电子设备
CN104537068B (zh) 2014-12-29 2018-11-13 浙江宇视科技有限公司 一种电子地图接入方法和装置

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101593457A (zh) * 2008-10-23 2009-12-02 中国科学院地理科学与资源研究所 基于区域划分的屏幕饱和度实现方法及装置
CN101808230A (zh) * 2009-02-16 2010-08-18 杭州恒生数字设备科技有限公司 一种用于数字视频监控的统一坐标系
CN101673282A (zh) * 2009-10-15 2010-03-17 杭州北海信息系统有限公司 电子地图自动定位系统方法
CN104899368A (zh) * 2015-05-29 2015-09-09 浙江宇视科技有限公司 基于数据热度的监控布点需求图生成方法及装置
CN105159978A (zh) * 2015-08-28 2015-12-16 浙江宇视科技有限公司 一种相机分布合理性的分析方法和装置

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109194930A (zh) * 2018-10-25 2019-01-11 广州益牛科技有限公司 楼宇环境一体式智能管理系统
CN113961827A (zh) * 2021-10-26 2022-01-21 中国人民公安大学 一种区域防控风险的检测方法及装置
CN113961827B (zh) * 2021-10-26 2022-08-09 中国人民公安大学 一种区域防控风险的检测方法及装置

Also Published As

Publication number Publication date
US20180150479A1 (en) 2018-05-31
US10445348B2 (en) 2019-10-15

Similar Documents

Publication Publication Date Title
CN104899368B (zh) 基于数据热度的监控布点需求图生成方法及装置
WO2016192516A1 (zh) 视频监控布点
Taubenböck et al. Measuring morphological polycentricity-A comparative analysis of urban mass concentrations using remote sensing data
Xu et al. A spatio–temporal pixel-swapping algorithm for subpixel land cover mapping
Jin et al. A super-resolution mapping method using local indicator variograms
CN109493119B (zh) 一种基于poi数据的城市商业中心识别方法及系统
Freitag et al. Testing for potential survey bias: the effect of roads, urban areas and nature reserves on a southern African mammal data set
Can Residential quality assessment: Alternative approaches using GIS
Qi et al. Development and application of 3D spatial metrics using point clouds for landscape visual quality assessment
Fecht et al. A GIS-based urban simulation model for environmental health analysis
De Bono et al. A global exposure model for GAR 2015
Chen et al. Subpixel land cover mapping using multiscale spatial dependence
JP2020166739A (ja) 位置情報に関連付けられたデータを利用してメッシュ統計を生成するためのシステム、方法、およびプログラム
Castilla et al. Uncertainties in land use data
Makido et al. Weighting function alternatives for a subpixel allocation model
Biljecki et al. Sensitivity of measuring the urban form and greenery using street-level imagery: A comparative study of approaches and visual perspectives
Chen et al. The effect of spatial autocorrelation and class proportion on the accuracy measures from different sampling designs
CN109740678A (zh) 基于多级不均匀空间抽样的遥感影像分类精度检验方法
Walter et al. Measuring the wisdom of the crowd: How many is enough?
Kim et al. Incorporating spatial context and fine-grained detail from satellite imagery to predict poverty
CN105159978B (zh) 一种相机分布合理性的分析方法和装置
Bhatti et al. 1.14 Spatial Metrics: The Static and Dynamic Perspectives
Mao et al. Population spatialization at building scale based on residential population index—A case study of Qingdao city
Antoni Urban sprawl modelling: a methodological approach
Ouyang et al. A Bayesian approach to mapping the uncertainties of global urban lands

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16802439

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 15577466

Country of ref document: US

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16802439

Country of ref document: EP

Kind code of ref document: A1