CN111536973A - Indoor navigation network extraction method - Google Patents
Indoor navigation network extraction method Download PDFInfo
- Publication number
- CN111536973A CN111536973A CN202010223992.XA CN202010223992A CN111536973A CN 111536973 A CN111536973 A CN 111536973A CN 202010223992 A CN202010223992 A CN 202010223992A CN 111536973 A CN111536973 A CN 111536973A
- Authority
- CN
- China
- Prior art keywords
- indoor
- track
- navigation network
- data
- trajectory
- 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.)
- Pending
Links
- 238000000605 extraction Methods 0.000 title claims abstract description 28
- 238000000034 method Methods 0.000 claims abstract description 65
- 238000012545 processing Methods 0.000 claims abstract description 21
- 230000006740 morphological transformation Effects 0.000 claims abstract description 7
- 239000011159 matrix material Substances 0.000 claims description 14
- 230000008569 process Effects 0.000 claims description 14
- 230000000877 morphologic effect Effects 0.000 claims description 10
- 238000001914 filtration Methods 0.000 claims description 9
- 238000004891 communication Methods 0.000 claims description 7
- 238000009499 grossing Methods 0.000 claims description 7
- 238000009826 distribution Methods 0.000 claims description 4
- 238000011985 exploratory data analysis Methods 0.000 claims description 2
- 238000005260 corrosion Methods 0.000 claims 1
- 230000007797 corrosion Effects 0.000 claims 1
- 238000005070 sampling Methods 0.000 abstract description 5
- 238000010586 diagram Methods 0.000 description 30
- 238000003860 storage Methods 0.000 description 10
- 238000004422 calculation algorithm Methods 0.000 description 6
- 238000011960 computer-aided design Methods 0.000 description 6
- 238000004590 computer program Methods 0.000 description 5
- 238000010276 construction Methods 0.000 description 5
- 230000003628 erosive effect Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 238000013507 mapping Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 3
- 238000007726 management method Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- HPTJABJPZMULFH-UHFFFAOYSA-N 12-[(Cyclohexylcarbamoyl)amino]dodecanoic acid Chemical compound OC(=O)CCCCCCCCCCCNC(=O)NC1CCCCC1 HPTJABJPZMULFH-UHFFFAOYSA-N 0.000 description 1
- 241000282412 Homo Species 0.000 description 1
- 241000282414 Homo sapiens Species 0.000 description 1
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000001174 ascending effect Effects 0.000 description 1
- 230000004888 barrier function Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000012733 comparative method Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- DMBHHRLKUKUOEG-UHFFFAOYSA-N diphenylamine Chemical compound C=1C=CC=CC=1NC1=CC=CC=C1 DMBHHRLKUKUOEG-UHFFFAOYSA-N 0.000 description 1
- 238000005530 etching Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000013467 fragmentation Methods 0.000 description 1
- 238000006062 fragmentation reaction Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- 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
- G01C21/20—Instruments for performing navigational calculations
- G01C21/206—Instruments for performing navigational calculations specially adapted for indoor navigation
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Automation & Control Theory (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Navigation (AREA)
Abstract
The invention discloses an indoor navigation network extraction method, which comprises the following steps: extracting any layer of indoor track data from the indoor track data set; rasterizing the indoor track data of any layer to obtain an indoor track raster image; carrying out binarization processing on the indoor track grid image to obtain a binarization indoor track grid image of the layer; and selecting a series of structural elements with preset shapes to perform morphological transformation on the binary indoor track grid image to obtain an indoor two-dimensional navigation network. The indoor navigation network extraction method provided by the embodiment of the invention can be used for obtaining the two-dimensional navigation network, and the method can be used for rasterizing the indoor track points to extract the indoor navigation network, so that the problem of large indoor track positioning error can be effectively solved, the method is better suitable for the situation of track data with large positioning error and sparse sampling, and can be used for more rapidly providing indoor path planning.
Description
Technical Field
The invention relates to the field of mobile internet, in particular to an indoor navigation network extraction method.
Background
Indoor spaces are the main spaces in which human beings move, such as office buildings, shopping centers, hospitals, airports, subway stations, etc., and studies have shown that about 87% of the time is spent by humans in indoor spaces. The indoor navigation network is the basis for indoor pedestrian navigation, indoor path planning, indoor personalized information service recommendation and the like. Compared with an outdoor space navigation network, the indoor space navigation network is more difficult to construct, on one hand, the indoor space belongs to a three-dimensional space, the indoor three-dimensional space has a complex structure, such as a closed space, a semi-closed space and the like, the indoor three-dimensional space has a plurality of entities, such as rooms, walls, doors, windows, corridors, elevators, stairs and the like, the indoor three-dimensional space has various constraints, such as communication constraints, barrier constraints and the like, and the construction of a three-dimensional topological communication relationship is one of the difficulties in the construction of the indoor navigation network; on the other hand, the outdoor road network structure is relatively fixed, the indoor corridor structure is narrow and changeable, and the corridor structure change frequency is higher in certain specific shopping plazas and other indoor spaces. The indoor navigation network is constructed by adopting methods such as manual field mapping or semi-manual CAD plan drawing extraction, and although the method can ensure the extraction precision of the navigation road network, the method extracts the indoor three-dimensional skeleton structure with thicker granularity and can not obtain the detailed indoor structure of the building.
Disclosure of Invention
Objects of the invention
The invention aims to provide an indoor navigation network extraction method, a storage medium and electronic equipment, wherein a binary indoor track raster image is obtained by rasterizing and binarizing an indoor track data set of any floor, and an indoor two-dimensional navigation network is obtained by performing morphological transformation on the image.
(II) technical scheme
According to a first aspect of the present invention, there is provided an indoor navigation network extraction method, including: extracting any layer of indoor track data from the indoor track data set; rasterizing and binarizing the indoor track data of any layer in sequence to obtain a binarized indoor track grid image of the layer; and selecting a series of structural elements with preset shapes to perform morphological transformation on the binary indoor track grid image to obtain an indoor two-dimensional navigation network.
According to the second aspect of the present invention, there is also provided an indoor navigation network extraction method, including: extracting the two-dimensional navigation network of two adjacent floors by adopting the method provided by the first aspect; identifying a connected point set of the two adjacent floors from the two-dimensional navigation network of the two adjacent floors; and connecting lines based on the two-dimensional navigation networks of the two adjacent floors and the connected point sets of the two adjacent floors to obtain indoor three-dimensional navigation networks of the two adjacent layers.
According to a third aspect of the present invention, there is also provided a storage medium having a computer program stored thereon, which when executed by a processor, implements the steps of the indoor navigation network extraction method provided by the first aspect.
According to a fourth aspect of the present invention, there is also provided an electronic device, including a memory, a processor and a computer program stored on the memory and executable on the processor, where the processor implements the steps of the indoor navigation network extracting method provided in the first aspect when executing the program.
(III) advantageous effects
The technical scheme of the invention has the following beneficial technical effects:
(1) the indoor navigation network extraction method provided by the embodiment of the invention can be used for obtaining the two-dimensional navigation network, and the method can be used for rasterizing the indoor track points to extract the indoor navigation network, so that the problem of large positioning error of indoor tracks can be effectively solved, the method is better suitable for the situation of track data with large positioning error and sparse sampling, and can be used for more rapidly providing indoor path planning.
(2) According to the indoor navigation network extraction method provided by the embodiment of the invention, the indoor three-dimensional navigation network is obtained, the indoor navigation network generation efficiency can be effectively improved, the navigation network generation cost is reduced, the problems of high cost and low efficiency of the conventional manual surveying and mapping are solved, data support is provided for real-time navigation of pedestrians in an indoor environment, the indoor position service is further promoted to be popularized and applied, and support is provided for the fields of mobile position service, urban traffic, mobile internet technology and the like.
Drawings
FIG. 1(a) is a CAD drawing of a prior art planar structure within a room;
FIG. 1(b) is a diagram of the distribution of gallery skeleton and the locations of the solid in the chamber in the prior art;
FIG. 1(c) is a schematic diagram of a moving track of a moving object on a floor in a room;
FIG. 1(d) is a schematic diagram of a moving track of a moving object in an indoor three-dimensional space;
fig. 2 is a schematic diagram of statistics of the number of track points of different floors in the first embodiment of the present invention;
FIG. 3 is a statistical schematic diagram of indoor trace point records in the first embodiment of the present invention;
fig. 4 is a schematic flow chart of an extraction method of an indoor navigation network according to a first embodiment of the present invention;
FIG. 5 is a schematic diagram of a movement trajectory of a user in an indoor three-dimensional space according to a first embodiment of the present invention;
FIG. 6 is a schematic diagram of trace point processing according to the first embodiment of the present invention;
fig. 7 is a schematic diagram illustrating indoor trajectory data of a certain floor being rasterized according to the first embodiment of the present invention;
FIG. 8(a) schematically provides a grid map of a binarized indoor trajectory for a floor in a first embodiment of the present invention;
FIG. 8(b) is a diagram illustrating the result of the morphological erosion process applied to the grid map of the tracks in the binarization chamber shown in FIG. 8 (a);
FIG. 8(c) is a diagram illustrating the result of an ON operation applied to FIG. 8 (b);
FIG. 9(a) is a schematic diagram of the mathematical morphology refinement algorithm performed on FIG. 8(c) to extract grid skeleton lines of the indoor navigation network;
FIG. 9(b) is a two-dimensional navigation network diagram obtained by superimposing FIG. 9(a) with a plan view of a room;
FIG. 10 is a schematic illustration of a series of predetermined shaped structural elements provided in accordance with a first embodiment of the present invention;
fig. 11 is a flowchart illustrating an extraction method of an indoor navigation network according to a second embodiment of the present invention;
fig. 12(a) is a schematic view of a floor connected zone acquired in the second embodiment of the present invention;
FIG. 12(b) is a schematic center view of the connected component extracted from FIG. 12 (a);
FIG. 13 is a schematic diagram illustrating identification of a floor connectivity point in a connectivity area in accordance with a second embodiment of the present invention;
FIG. 14 is a schematic diagram of an indoor three-dimensional navigation network according to a second embodiment of the present invention;
FIG. 15 is a schematic diagram of a decision graph provided in accordance with a second embodiment of the present invention;
FIG. 16 shows a second embodiment of the present invention for γiSchematic arrangement of (a).
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the accompanying drawings in conjunction with the following detailed description. It should be understood that the description is intended to be exemplary only, and is not intended to limit the scope of the present invention. Moreover, in the following description, descriptions of well-known structures and techniques are omitted so as to not unnecessarily obscure the concepts of the present invention.
Before describing in detail the method provided by the first embodiment of the present invention, a brief description of the prior art will be provided.
In the prior art, the construction of an indoor navigation network requires acquiring an indoor structure diagram, and then constructing the indoor navigation network based on the indoor structure diagram. There are two methods for obtaining the indoor structure chart, and one method is to obtain the indoor structure chart by adopting a manual field mapping mode. Another method is to obtain the indoor structure diagram by using a semi-artificial CAD plan. Although the indoor navigation network constructed by the method can ensure the extraction precision of the navigation network, the indoor three-dimensional skeleton structure extracted by the two methods has coarse granularity, and the detailed structure of the building cannot be obtained, for example, a CAD (computer aided design) drawing is basically impossible to contain a movable indoor entity. If a CAD plan of a floor as shown in fig. 1(a) is obtained, it is not known how each room is configured, for example, 2302, 2304, 2305, 2307, 2309 and 2311 contain student stations and other devices. Fig. 1(b) is a schematic diagram of the indoor entity location and the skeleton distribution of the corridor in the 2309 chamber in fig. 1(a), in this case, a navigation network needs to be constructed in combination with fig. 1(a) and 1(b), so this approach has the disadvantages of tedious steps and low accuracy.
With the rapid development of indoor positioning technology, many indoor positioning methods, such as WiFi positioning, Radio Frequency Identification (RFID) positioning, bluetooth or nfc (near field communication) positioning, pseudolite positioning, ZigBee positioning, UWB (Ultra-wide) positioning, ultrasonic positioning, image matching and barcode positioning, and geomagnetic positioning, have appeared. Or positioning by using a mobile terminal with a built-in positioning module, for example, indoor positioning by using a smart phone, a tablet computer, a PDA, and the like. With the continuous development of the mobile internet, the application of indoor location services is increasing, such as online navigation, a location-based social network, advertisement push based on location, and the like, a great amount of track data of a mobile object, such as the track data shown in fig. 1(b) and 1(c), is generated in indoor space, and by analyzing the track data of the mobile object, a new possibility is provided for the automatic construction of an indoor navigation network, and effective supplement and rapid change update detection of an indoor space structure can be provided for a traditional indoor navigation network construction method.
In the prior art, the research for extracting and updating the indoor navigation network structure by using the indoor crowdsourcing trajectory data is relatively less, and the following technical difficulties mainly exist:
(1) different from the single road network constraint of the outdoor space, the indoor space is a mixed space mode of 'free space + road network space', and the movement of an indoor mobile user can freely move in the indoor space except running along an indoor corridor, so that the difficulty of extracting the indoor navigation network is increased.
(2) Indoor positioning error is great, and like Wifi fingerprint locate mode error at 3 ~ 5 meters, compare with outdoor GPS or big dipper's positioning error, the gap is great, causes indoor moving object orbit very irregular, as shown in the figure, causes indoor crowd's orbit update method degree of difficulty bigger.
Therefore, the present invention provides the following method for extracting an indoor navigation network under the above technical background.
The method for extracting an indoor navigation network according to the first embodiment of the present invention is discussed in detail below.
Indoor track data adopted by the method are WIFI positioning data within 2 days from the Jinan harmonious square, the related tracks are 4000, the number of track points is over 400 thousands, and the data sampling interval is different from 1-10 seconds.
Fig. 2 is a statistical schematic diagram of the number of track points of different floors in the first embodiment of the present invention.
As shown in FIG. 2, in the indoor track data, track points with sampling intervals of 1-2 s account for more than 80%, and the data comprises five fields of user equipment numbers, positioning time, x coordinates, y coordinates and floor numbers. Table 1 below is a data table of an indoor trajectory moving object of a certain moving object, and as can be clearly seen in table 1, mac refers to a device number of a user, time refers to a positioning time of the user, x refers to an x coordinate of the user at the positioning time, and y refers to a y coordinate of the user at the positioning time.
TABLE 1
The data processing is started next.
Generally, the positioning error of the indoor track data is 3-5 m, and the indoor track data contains a large amount of redundant information, which is rough error information of the same position at more or more times at the same time. Therefore, it is preferable to remove this part of redundant information to reduce the subsequent data processing amount and improve the extraction speed of the indoor navigation network.
In a preferred embodiment of the present invention, the raw data is pre-processed. The data preprocessing steps are as follows:
first, the original trajectory data of the indoor moving object, InTraj, is acquired, and the acquired original trajectory data can be stored and managed.
The definition of the raw trajectory data is:
InTraj＝{itj1，itj2，...，itjn}
itji＝{ipt1，ipt2，...，iptm}
ipti＝{midi，lati，loni，timei，floori}
wherein, itjiIndoor trajectory data of day, mid, for user iiId unique identification of user, (lat)i，loni) Time at time for user iiLocation information of (1), flooriThe information of the floor where the user i is located.
Alternatively, the storage management may employ a relational database, such as Oracle, MySQL, or PostgreSQL, etc. Unstructured databases may also be used for data logic storage and management, such as NoSQL databases, MongoDB, CouchDB, or HBase.
Optionally, the indoor trajectory data may be preprocessed by centralized high-performance processing, such as CUDA or MPI, or may be processed by distributed clustering, such as MapReduce or BSP. The preprocessed data has the characteristic of mass sparseness.
For reference, in the present embodiment, the database employs a MongoDB cluster, and performs logical storage and management of data in an unstructured or structured manner. In an actual storage scheme, an automatic fragmentation strategy of a database is adopted, and a MapReduce model is adopted in a distributed cluster processing model, so that the physical storage and processing of indoor user trajectory data are realized by utilizing a distributed computing environment and a storage structure.
Secondly, exploratory analysis is carried out on the original track data, and tracks with points smaller than a first preset value are deleted to obtain track data of the indoor moving object.
Fig. 3 is a statistical diagram of indoor trace point records in the first embodiment of the present invention.
As shown in fig. 3, in this embodiment, the first preset value is 100, that is, the track less than 100 points in the indoor track is deleted, and the track less than 100 points shows that the moving object moves less indoors, and the value of extracting the indoor navigation network is not great, so that deleting the part of data can simplify the data, and facilitate subsequent processing of the data.
Finally, the indoor moving object track data are layered according to floors to obtain an indoor track data set InTraj containing each floorfloor。
The indoor trajectory data set InTrajfloorIt can be defined as:
InTrajfloor＝{Intrajf1，Intrajf2，…，Intrajfp}
wherein, Intrajf1For the trajectory data of all indoor mobile users at floor f1, fp is the number of floors.
After the indoor trajectory data set for each floor is extracted, extraction of the indoor navigation network may be performed.
Fig. 4 is a schematic flow chart of an indoor navigation network extraction method according to a first embodiment of the present invention.
As shown in fig. 4, the method includes steps S101 to S104:
and step S101, extracting indoor track data of any layer from the indoor track data set.
The indoor track data set of any layer is represented as Intrajfj。
In a preferred embodiment, after step S101, the simplified processing is performed on the indoor trajectory data of the layer. The steps of the simplified process are as follows:
a time threshold Etm and a distance threshold Eps are set.
Acquiring a space-time neighborhood set STN (ipt) of each indoor pedestrian trajectory point ipt, wherein the space-time neighborhood set is acquired by the following method:
STN(ipt)＝{iqt∈IS|dists(ipt，iqt)≤Eps，distt(ipt，iqt)≤Etm}
combining two space-time neighborhood sets with n common repeated track points to obtain a new space-time neighborhood set;
repeatedly executing the steps until the number of the repeated track points in any two space-time neighborhood sets is less than n, and further obtaining a clustered set of clustered high-density track points;
clustering a set aiming at each high-density track point, and extracting a centroid point as a newly generated point IC (x, y, t), wherein the centroid point IC (x, y, t) is obtained by the following method:
deleting high-density track point clusters outside the track point centroid; and/or
And deleting the trace point with the speed of 0 in continuous time.
It should be noted that, in this embodiment, the purpose of the simplification process is to remove two kinds of redundant track points, where the first redundant track point is a track point that appears in the track in a large amount of time and continuously but at a constant position due to a moving object staying at the same position for a long time. The second redundant trace point is a situation that a moving object wanders around a position, a small range of aggregation occurs in a trace, and a high-density trace point cluster occurs.
Specifically, the first redundant trace point is removed by deleting the trace point with the speed of 0 in continuous time.
And removing the second redundant track points, namely identifying high-density track point clusters formed by each track under a continuous time sequence, and clustering the high-density track point clusters into track point centroids. After determining the track point centroid of the high-density track point cluster, replacing all track points in the cluster set with the centroid point, wherein the calculation formula is as follows:
fig. 6 is a schematic diagram of track point processing in the first embodiment of the present invention.
As shown in fig. 6, the left graph is an original indoor movement track of a user, and the positions of centroid points of track points in cluster clusters are identified through an algorithm. And deleting the remaining points except the trace point in the high-density point cluster.
And S102, rasterizing any layer of indoor track data in sequence to obtain an indoor track raster image.
In one embodiment, any layer of indoor track data is rasterized, and in the rasterization process, a statistical threshold value during rasterization is adjusted according to the number and density of points in a grid neighborhood.
More specifically, the trajectory data of each floor is generated into an M × M matrix. And traversing the squares in each M × M matrix, and taking the number of the trace points in each square as the pixel value of the square to obtain the indoor trace matrix M × M filled with the pixel values. Smoothing each square by adopting Gaussian filtering to obtain an optimized matrix of M multiplied by M, wherein the optimized matrix of M multiplied by M is an indoor track grid image IBP. The smoothing process is as follows:
wherein,
m is the number of pixel squares, xijThe original value of the pixel grid.
It should be noted that, in the above smoothing process using gaussian filtering, the selection of values has a great influence on the rasterization effect of the image, if the selection is too small, the weight of the non-central pixel point is very small, the effect of gaussian filtering is very small, if the selection is too large, the details of the indoor rasterized image are easily lost, and the conventional gaussian filtering is fixed. Therefore, the Gaussian template can be adaptively selected according to the variance size by adopting the formula. Namely: and calculating the size of the variance ID in the neighborhood, and automatically selecting a value according to the size of the variance ID, wherein the larger the variance ID is, the smaller the sigma value is, and the smaller the variance ID is, the larger the sigma value is.
Fig. 7 is a schematic diagram showing the indoor trajectory data of a certain floor (jth floor) being rasterized.
As shown in FIG. 7, the pixel value N of the pixel (i, j) at the set raster image resolutioniiIs 0, and 3 of its 8 neighborhood pixels have values other than 0, so (i, j) is a topologically disconnected pixel, then NijAnd the average value of 8 adjacent pixels is assigned again. And if the pixel value of only one pixel (m-1, n) in the 8 neighborhood pixels of the pixel (m, n) is not 0, the pixel is determined as a noise pixel and is not processed.
When the resolution of the image element is small or the area of the topological disconnected region is generally large, the 3 × 3 window can be expanded into 5 × 5 windows. If more noise is represented as continuous noise pixels, the number of pixels which are not 0 in the neighborhood pixels can be increased, and the probability of processing the noise pixels is reduced. The rasterization method can weaken the influence of the selection of the resolution of the raster image on the road network extraction result, can effectively avoid the problem of navigation network topology disconnection caused by the difference of the corridor track density, and can better overcome the influence of noise.
Step S103, performing binarization processing on the indoor track grid image to obtain a binarization indoor track grid image of the layer. The embodiment of the invention also provides an algorithm for the binarization processing of the indoor track grid image.
The method comprises the following steps of carrying out binarization processing on an indoor track grid image so as to convert a gray scale map so as to extract a skeleton map, and further carrying out binarization processing on the indoor track grid image:
firstly, a binarization threshold th is set for an indoor track grid imagevalue。
Secondly, judging whether the pixel value of the pixel (i, j) in each grid in the indoor track raster image is larger than the binarization threshold thvalue。
Then, if the pixel value of the pixel (i, j) is greater than or equal to the binarization threshold thvalueThe pel value of this pel (i, j) is reassigned to 1. If the pixel value of the pixel (i, j) is less than the binary threshold thvalueThe pixel value of this pixel (i, j) is assigned 0.
And obtaining a binary indoor track grid image by performing the judgment processing on each grid.
Note that the binary threshold thvalueThe larger the setting is, the greater the reliability of the obtained image element as a gallery road is, but the road surface image element is easy to be too sparse. Thus, the binary threshold thvalueAnd the decision needs to be made according to the pixel value after the self-adaptive rasterization reassignment.
And S104, selecting a series of structural elements with preset shapes to perform morphological transformation on the binarization indoor trajectory raster image to obtain an indoor two-dimensional navigation network.
Firstly, performing morphological filtering operation on the obtained binary indoor track grid image to perform denoising processing on the binary indoor track grid image.
Specifically, a morphological erosion operation is used to remove patches belonging to roads in the image.
It should be noted that, in the binarized indoor trajectory grid image, there may be a phenomenon that the trajectory is adhered to the road, and at this time, the size of the structural element is adjusted according to the size of the maximum noise pattern spot by using morphological erosion operation, so that the noise pattern spot in the binarized indoor trajectory grid image can be completely removed as much as possible, and the details of the road in the binarized indoor trajectory grid image are maintained as much as possible.
The 'unevenness' of the edge of the indoor corridor can cause more fine branches to appear on the central line of the binaryzation indoor track grid image, at this time, an open operation method is adopted to process the binaryzation indoor track grid image, then a closed operation is adopted to the image, and the zigzag position of the edge of the road in the binaryzation indoor track grid image is processed, so that the zigzag of the edge of the road in the image is further smooth.
Fig. 8(a) schematically shows a grid diagram of a binarized indoor trajectory at a certain floor in the first embodiment of the present invention. As shown in fig. 8(a), in the binarized indoor trajectory grid map, white represents a road, and black represents a position where a pedestrian cannot move.
Fig. 8(b) is a schematic diagram showing the result of the morphological etching process applied to the binarized indoor trajectory grid map shown in fig. 8 (a).
And (4) processing the adhesion part of the binarized indoor track grid image and the road in the black frame in the step (a) of fig. 8 by a morphological erosion operation, wherein the processing result can be seen in fig. 8 (b). The formula in which the morphological erosion operation is performed can be expressed as:
wherein, B is the processed indoor binary grid image, and A is the structural element.
Fig. 8(c) is a diagram showing the result of the on operation applied to fig. 8 (b).
As shown in fig. 8(b), the open operation is performed on the black frame portion, and the closed operation is performed on the graph after the open operation, thereby further smoothing the jaggy of the road edge, as shown in fig. 8 (c). As can be seen in fig. 8(c), the road connectivity of the indoor track image after the optimization process is further enhanced, the road edge jaggy is also smoothed, and the noise is basically removed, wherein the noise with larger individual area is mistaken for the road due to the size of the structural element being exceeded and is not filtered.
The formula for this open operation can be expressed as:
wherein, B is the processed indoor binary grid image, and A is the structural element.
And finally, selecting a series of structural elements with preset shapes to perform morphological transformation on the filtered binary indoor track grid image to obtain an indoor two-dimensional navigation network.
Specifically, in this step, a series of structural elements with predetermined shapes are selected to transform the binarized indoor trajectory raster image, iteration is performed continuously, pixels meeting the requirement of hitting the conversion are deleted in each iteration process, and when the binarized indoor trajectory raster image is not changed, an indoor two-dimensional navigation network is obtained, which can be seen in fig. 9 (b). Wherein a series of structural elements of a predetermined shape can be referred to fig. 10.
The formula for the conversion of the hit miss can be referred to as:
wherein, B is the processed indoor binary grid image, and A is the structural element.
Specifically, the filtered binarized indoor trajectory grid map is represented by X, and the skeleton refined by mathematical morphology is s (X):
wherein S isn(X) an nth subset of skeletons that is X; n is to satisfy
And
n, i.e. X is eroded at least (N +1) times by B to become a null image;
represents the expansion of n B; collection
It should be noted that the structural elements are selected to ensure structural connectivity of the target image during each iteration of the refinement algorithm and to ensure that the structure of the entire image is unchanged.
The sequence D ═ D is determined using the structural element pair shown in fig. 101，D2，D3，D4D ═ E1，E2，E3，E4Refine the indoor trajectory image, where structure pair sequence D is used to eliminate points in the north-west, north-east, south-east and south-west 4 directions, and structure pair sequence E is used to eliminate points in the north, east, south and west 4 directions. "1" represents a point on the target image, "0" represents a point on the background image, and "", may be a point on the target image or a point on the background image. Morphological refinement is carried out on the indoor track image, and the nth iteration process can be represented as:
in the formula,
specifically, a binary neighborhood coding mode is preset in the invention, and assuming that each element of a matrix has eight adjacent grids, for an actual matrix element, there are only 3 or 5 grids actually, and then the binary value corresponding to the actual grid is 1, and the binary value corresponding to the actual non-grid is 0. A binary number is obtained in the order of the P indices of table 1 below, for example a binary value of 00011011 for table 2 below, corresponding to a decimal value of 27. And then binary values of 8 neighborhoods corresponding to each element in the M multiplied by M matrix can be calculated.
TABLE 1
| P9 | P2 | P3 |
| P8 | P1 | P4 |
| P7 | P6 | P5 |
TABLE 2
| P9 | P8 | P7 | P6 | P5 | P4 | P3 | P2 |
| 1 | 1 | 0 | 1 | 1 | 0 | 0 | 0 |
The invention divides the type of the neighborhood into 2 types, the first type is 8 neighborhoods, the value of two pixel points is 1, when the binary coding meets the following conditions, the type is determined.
This type of binary encoding is: IZB (P)0) ∈ {5, 20, 80, 65} (see D1-D4 in FIG. 10).
The second type is 8 neighborhoods, the values of three pixel points are 1, and the binary coding of the type meets the following conditions: IZB (P)0) ∈ {13, 22, 52, 67, 88, 97, 133, 208} (see E1-E4 in fig. 10).
If the binary values of 8 neighborhoods corresponding to a certain grid in the M x M matrix satisfy the two types, the grid is extracted, a plurality of grids are further obtained, each grid corresponds to a road where pedestrians can walk, and all the extracted grids are the 'skeleton lines' of the layer. Namely, the skeleton line corresponds to the position where the pedestrian can walk on the whole floor.
The formula for extracting the skeleton line is as follows:
1)2≤INZ(P0)≤6
2)P2×P4×P6＝0
3)P4×P6×P8＝0
wherein, INZ (P)0) Represents P0The number of peripheral non-zero points.
Finally, converting the indoor track image thinning result into vector data to obtain the indoor navigation network of the floor
The graph shown in fig. 8(c) is subjected to a mathematical morphology refinement algorithm, and the extracted grid skeleton line of the indoor navigation network is used to obtain fig. 9 (a). As can be seen in fig. 9(a), the road skeleton lines substantially outline the overall geometry of the indoor corridor, with fewer road discontinuities and fewer road fine-breaking branches.
Vectorizing the binarized indoor trajectory grid map shown in fig. 9(a) to obtain a vectorized indoor navigation network map 9(b), wherein the road fine-crushing branches and the noise are filtered in the vectorization process. After the navigation network is overlapped with a market plan, the navigation network extracted by the invention basically covers all roads where the test area data are located, and the accuracy is higher.
The indoor navigation network extraction method provided by the embodiment of the invention can be used for obtaining the two-dimensional navigation network, and the method is used for rasterizing the indoor track points to extract the indoor navigation network, so that the problem of large positioning error of indoor tracks can be effectively solved, the method is better suitable for the situation of track data with large positioning error and sparse sampling, and can be used for more rapidly providing indoor path planning.
Fig. 11 is a method for extracting an indoor navigation network according to a second embodiment of the present invention, where the method includes steps S201 to S203:
step S201, the method provided by the first embodiment is adopted to extract two-dimensional navigation networks of two adjacent floors.
Specifically, all indoor track data sets Intraj of the mth floor are extractedfmIndoor rail of nth floorTrace dataset IntrajfnAnd performing indoor three-dimensional navigation network extraction. Wherein the mth floor and the nth floor are adjacent floors.
And step S202, identifying a connected point set of the two adjacent floors from the two-dimensional navigation network of the two adjacent floors.
Specifically, a floor attribute f in each original trajectory is identifiediInconsistency and positioning time difference is less than time threshold TthrA plurality of pairs of trace points; wherein the pairs of trace points can be represented by the following formula:
{…，Pi，Pj，…|tj-tj＜Tthr，fi≠fj}，Pi＝(ID，x，y，t，f)，i∈[1，n]。
if the position difference of a certain pair of track points is within the set distance threshold disthrWithin the range of, i.e.
The pair of trace points is stored in a data set of a floor communication area
Wherein identifying a set of floor connectivity points from the set of floor connectivity zone data comprises:
acquiring an indoor track data point set IPT (IPT) of each layer1，ipt2，...，iptNAnd calculating the distance id between any two track pointsijAnd finally obtaining a distance set id.
Arranging the distances in the obtained track distance set according to an ascending order to obtain
id＝{(id1，id2，...，idM)|idi≤idi+1}
Extracting a value id of which the track distance frequency value is more than 20%c。
idc＝f(id)
Traversing the indoor track data point set, and calculating the local density rho of each pointi，
Wherein,
when idij＜idcThen, χ (id, it) ═ 1. When idij≥idcWhen x (id, it) is 0.
Arranging all the local densities in a descending order to obtain a local density set rho:
then calculating the distance valuei。
By (ρ)i，i) A decision graph is constructed, which can be seen in fig. 15.
According to the distance gammaiDetermining cluster center, distance gammaiThe larger the value, the more likely it is a cluster center.
γi＝ρi i
Wherein, γiThe arrangement is shown in figure 16.
After the clustering center is obtained, traversing the remaining non-clustering center points, and judging the category attribute of the non-clustering center.
And finally, clustering indoor track points, and finding a clustering center of the clustering cluster set to serve as a topological communication point set between two indoor floors.
Fig. 12(a) is a schematic view of a floor connected zone acquired in the second embodiment of the present invention. Fig. 12(b) is a schematic center view of the connected component extracted from fig. 12 (a).
As shown in fig. 12(a), which is a schematic diagram of a three-dimensional topological connected region, four dense points are shown on the diagram, and these points can represent the region where floors are connected. Fig. 12(b) is a view obtained by extracting fig. 12(a) and obtaining the center of the floor connected region.
Fig. 13 is a schematic diagram illustrating identification of a floor connection point in a connected region according to the second embodiment of the present invention.
As shown in fig. 13, each point in the floor connected region is at a cutoff distance dcThere are corresponding local densities ρ and distances γ, and it can be seen that the sample points O1, O2 have both larger values of ρ and p, i.e., a larger value of ρ. Therefore O1, O2 are determined as data sets
Two floor connection points. Identifying from the identified floor connected region
Middle floor connection point
And step S203, connecting lines based on the two-dimensional navigation networks of the two adjacent floors and the connected point sets of the two adjacent floors to obtain two adjacent layers of indoor three-dimensional navigation networks.
In one embodiment, the method further comprises step S204:
two-dimensional navigation network NG for extracting all indoor floors2D；
Circularly traversing all adjacent two floors in the room to obtain a communication point set Cpt of all adjacent two floors;
two-dimensional navigation network NG based on all floors2DAnd obtaining the indoor three-dimensional navigation network by the connected point sets Cpt of all the adjacent two floors.
Fig. 14 is a schematic diagram of an indoor three-dimensional navigation network. Therefore, the indoor navigation network extracted by the invention can better update the navigation network structure of the mall.
In order to verify the effectiveness of the patented method, the experiment adopts two indexes of accuracy and recall rate:
wherein P represents accuracy, R represents recall,
denotes a set of navigation networks extracted from the text, phi ═ L1，L2，…，LmPhi represents the original navigation network set, sigma represents the navigation network set in the buffer area, and len (L) represents the navigation network length.
The invention discloses an algorithm for extracting a road network central line by a 2D histogram-based method, which is proposed by Daves and the like in a comparison experiment.
The following table 2 shows the results of the experiment,
TABLE 2 evaluation of the results
As can be seen from table 2 above, when the radius of the buffer is 0.2m, 0.5m, and 0.7m, the accuracy of the navigation network of the method of the present invention respectively reaches 32.9%, 68.6%, and 76.6%, which are respectively increased by 0.8%, 2.2%, and 2.5% compared with the comparative method, and the recall of the navigation network of the method of the present invention respectively reaches 30.7%, 64.1%, and 70.3%, which are respectively increased by 0.8%, 1.9%, and 2.2%, and the accuracy of the topology is increased by 13.7%. Therefore, the method can effectively improve the geometric accuracy and the topological accuracy of the indoor navigation network updating.
According to the indoor navigation network extraction method provided by the second embodiment of the invention, an indoor three-dimensional navigation network is obtained, and by the method, the indoor navigation network generation efficiency can be effectively improved, the navigation network generation cost is reduced, the problems of high cost and low efficiency of the conventional manual surveying and mapping are solved, data support is provided for real-time navigation of pedestrians in an indoor environment, indoor position service is promoted to be further popularized and applied, and support is provided for the fields of mobile position service, urban traffic, mobile internet technology and the like.
The third embodiment of the present invention further provides a storage medium, wherein the storage medium stores a computer program, and the program, when executed by a processor, implements the steps of the indoor navigation network extraction method according to the first embodiment.
The fourth embodiment of the present invention further provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and when the processor executes the computer program, the steps of the indoor navigation network extraction method according to the first embodiment are implemented.
It is to be understood that the above-described embodiments of the present invention are merely illustrative of or explaining the principles of the invention and are not to be construed as limiting the invention. Therefore, any modification, equivalent replacement, improvement and the like made without departing from the spirit and scope of the present invention shall be included in the protection scope of the present invention. Further, it is intended that the appended claims cover all such variations and modifications as fall within the scope and boundaries of the appended claims or the equivalents of such scope and boundaries.
Claims (10)
1. An indoor navigation network extraction method is characterized by comprising the following steps:
extracting any layer of indoor track data from the indoor track data set;
rasterizing the indoor track data of any layer to obtain an indoor track raster image;
carrying out binarization processing on the indoor track grid image to obtain a binarization indoor track grid image of the layer;
and selecting a series of structural elements with preset shapes to perform morphological transformation on the binary indoor track grid image to obtain an indoor two-dimensional navigation network.
2. The indoor navigation network extraction method of claim 1, wherein the step of extracting the indoor trajectory data of any layer from the indoor trajectory data set further comprises:
acquiring original track data of an indoor moving object,
performing exploratory analysis on the original track data, and deleting tracks with points smaller than a first preset value to obtain track data of indoor moving objects;
and layering the indoor moving object trajectory data according to floors to obtain an indoor trajectory data set containing each floor.
3. The method of claim 1, further comprising, after extracting any layer of indoor trajectory data from the indoor trajectory data set and before obtaining the binarized indoor trajectory raster image for that layer:
a time threshold Etm and a distance threshold Eps are set.
Acquiring a space-time neighborhood set STN (ipt) of each indoor pedestrian trajectory point ipt, wherein the space-time neighborhood set is acquired by the following method:
STN(ipt)＝{iqt∈IS|dists(ipt，iqt)≤Eps，distt(ipt，iqt)≤Etm}
combining two space-time neighborhood sets with n common repeated track points to obtain a new space-time neighborhood set;
repeatedly executing the steps until the number of the repeated track points in any two space-time neighborhood sets is less than n, and further obtaining a clustered set of clustered high-density track points;
clustering a cluster set aiming at each high-density track point, and extracting a centroid point;
deleting high-density track point clusters outside the track point centroid; and/or
And deleting the trace point with the speed of 0 in continuous time.
4. The method of claim 1, wherein rasterizing the arbitrary layer of indoor trajectory data comprises:
taking each floor as a matrix;
traversing squares of each matrix, and taking the number of trace points in each square as the pixel value of the square to obtain an indoor trace matrix filled with the pixel value;
smoothing each square by adopting Gaussian filtering to obtain an optimized matrix, wherein the optimized matrix is an indoor track grid image; the smoothing process is as follows:
5. The method of claim 4, wherein the variance ID is adaptively selected by the following method when performing the smoothing process by Gaussian filtering on each square:
wherein,
m is the number of pixel squares, xijThe original value of the pixel grid.
6. The method as claimed in claim 1, wherein the step of binarizing the indoor-trajectory raster image to obtain the binarized indoor-trajectory raster image of the layer comprises:
setting a binarization threshold th for an indoor track grid imagevalue；
Judge the indoorWhether the pixel value of the pixel (i, j) in each grid in the track grid image is larger than the binarization threshold thvalue。；
If the pixel value of the pixel (i, j) is larger than the binary threshold thvalueThe pel value of this pel (i, j) is reassigned to 1. If the pixel value of the pixel (i, j) is less than the binary threshold thvalueThe pixel value of this pixel (i, j) is assigned 0.
7. The method as claimed in claim 1, further comprising, after obtaining the binarized indoor trajectory raster image of the layer, performing a morphological filtering operation on the binarized indoor trajectory raster image, which includes the steps of:
removing the pattern spots belonging to the road in the image by adopting a morphological corrosion operation;
processing the jagged road edges by adopting an open operation method to smooth the jagged road edges;
and selecting a series of structural elements with preset shapes to perform morphological transformation on the filtered binary indoor track grid image to obtain an indoor two-dimensional navigation network.
8. An indoor navigation network extraction method is characterized by comprising the following steps:
extracting a two-dimensional navigation network of two adjacent floors by using the method of any one of claims 1-7;
identifying a connected point set of the two adjacent floors from the two-dimensional navigation network of the two adjacent floors;
and connecting lines based on the two-dimensional navigation networks of the two adjacent floors and the connected point sets of the two adjacent floors to obtain indoor three-dimensional navigation networks of the two adjacent layers.
9. The method of claim 8, wherein identifying a set of floor connectivity points from the two-dimensional navigation network of the two adjacent floors comprises:
identifying a plurality of pairs of track points with inconsistent floor attributes in each original track and positioning time difference smaller than a time threshold;
if the position difference of a certain pair of track points is within a set distance threshold range, storing the pair of track points into a floor communication area data set;
a set of floor connectivity points is identified from the set of floor connectivity region data.
10. The method of claim 8 or 9, further comprising:
extracting two-dimensional navigation networks of all indoor floors;
circularly traversing all adjacent two floors in the room to obtain a connected point set of all adjacent two floors;
and obtaining the indoor three-dimensional navigation network based on the two-dimensional navigation network of all floors and the connected point sets of all adjacent two floors.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010223992.XA CN111536973A (en) | 2020-03-26 | 2020-03-26 | Indoor navigation network extraction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010223992.XA CN111536973A (en) | 2020-03-26 | 2020-03-26 | Indoor navigation network extraction method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111536973A true CN111536973A (en) | 2020-08-14 |
Family
ID=71973117
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010223992.XA Pending CN111536973A (en) | 2020-03-26 | 2020-03-26 | Indoor navigation network extraction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111536973A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112148829A (en) * | 2020-09-30 | 2020-12-29 | 重庆市规划设计研究院 | GIS algorithm optimization method applied to elimination of broken pattern spots |
| CN113686339A (en) * | 2021-08-16 | 2021-11-23 | 东南大学 | Indoor navigation road network extraction method based on crowdsourcing data of mobile terminal |
| CN114786199A (en) * | 2022-04-21 | 2022-07-22 | 中国联合网络通信集团有限公司 | Method, device, equipment and storage medium for determining network problem point |
| CN118031975A (en) * | 2024-04-15 | 2024-05-14 | 山东省科霖检测有限公司 | Large-scale environmental humidity monitoring method and system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106353724A (en) * | 2016-08-19 | 2017-01-25 | 北京安易康科技有限公司 | Method for accurate indoor positioning based on wireless sensor network |
| CN108882173A (en) * | 2018-07-04 | 2018-11-23 | 中国科学院地理科学与资源研究所 | A kind of pretreatment of interior Wifi location data and trajectory reconstruction method |
| CN110001637A (en) * | 2019-04-10 | 2019-07-12 | 吉林大学 | A kind of pilotless automobile path following control device and control method based on multiple spot tracking |
-
2020
- 2020-03-26 CN CN202010223992.XA patent/CN111536973A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106353724A (en) * | 2016-08-19 | 2017-01-25 | 北京安易康科技有限公司 | Method for accurate indoor positioning based on wireless sensor network |
| CN108882173A (en) * | 2018-07-04 | 2018-11-23 | 中国科学院地理科学与资源研究所 | A kind of pretreatment of interior Wifi location data and trajectory reconstruction method |
| CN110001637A (en) * | 2019-04-10 | 2019-07-12 | 吉林大学 | A kind of pilotless automobile path following control device and control method based on multiple spot tracking |
Non-Patent Citations (6)
| Title |
|---|
| 傅梦颖: ""基于移动对象轨迹的室内导航网络构建方法"", 《地球信息科学》 * |
| 李晓峰: "《中国地图学年鉴 1991》", 30 November 1992 * |
| 牟少敏: ""一种改进的快速并行细化算法"", 《微电子学与计算机》 * |
| 王培晓: ""ST-CFSFDP：快速搜索密度峰值的时空聚类算法"", 《测绘学报》 * |
| 王海菊: ""自适应高斯滤波图像去噪算法"", 《福建电脑》 * |
| 蔡林: ""基于分级栅格化和改进细化算法的轨迹数据路网生成研究"", 《数字制造科学》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112148829A (en) * | 2020-09-30 | 2020-12-29 | 重庆市规划设计研究院 | GIS algorithm optimization method applied to elimination of broken pattern spots |
| CN112148829B (en) * | 2020-09-30 | 2023-05-16 | 重庆市规划设计研究院 | GIS algorithm optimization method applied to broken pattern spot elimination |
| CN113686339A (en) * | 2021-08-16 | 2021-11-23 | 东南大学 | Indoor navigation road network extraction method based on crowdsourcing data of mobile terminal |
| CN113686339B (en) * | 2021-08-16 | 2023-11-28 | 东南大学 | Indoor navigation road network extraction method based on crowdsourcing data of mobile terminal |
| CN114786199A (en) * | 2022-04-21 | 2022-07-22 | 中国联合网络通信集团有限公司 | Method, device, equipment and storage medium for determining network problem point |
| CN118031975A (en) * | 2024-04-15 | 2024-05-14 | 山东省科霖检测有限公司 | Large-scale environmental humidity monitoring method and system |
| CN118031975B (en) * | 2024-04-15 | 2024-06-11 | 山东省科霖检测有限公司 | Large-scale environmental humidity monitoring method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111536973A (en) | Indoor navigation network extraction method | |
| Yadav et al. | Extraction of road surface from mobile LiDAR data of complex road environment | |
| Wu et al. | Automated extraction of ground surface along urban roads from mobile laser scanning point clouds | |
| Wu et al. | An extended minimum spanning tree method for characterizing local urban patterns | |
| CN103703490B (en) | For generation of the equipment of three-dimensional feature data and the method for generation of three-dimensional feature data | |
| CN108320323B (en) | Building three-dimensional modeling method and device | |
| Sohn et al. | An implicit regularization for 3D building rooftop modeling using airborne lidar data | |
| CN101751449A (en) | Spatial overlap analysis method and system used in geographic information system | |
| CN107182036A (en) | The adaptive location fingerprint positioning method merged based on multidimensional characteristic | |
| CN115512216A (en) | City functional area fine recognition method coupling block space-time characteristics and ensemble learning | |
| Galvanin et al. | Extraction of building roof contours from LiDAR data using a Markov-random-field-based approach | |
| Sun et al. | Roads and Intersections Extraction from High‐Resolution Remote Sensing Imagery Based on Tensor Voting under Big Data Environment | |
| CN111429698A (en) | Geological disaster early warning system | |
| Chen et al. | 3D model-based terrestrial laser scanning (TLS) observation network planning for large-scale building facades | |
| CN116662468A (en) | Urban functional area identification method and system based on geographic object space mode characteristics | |
| CN113724279A (en) | System, method, equipment and storage medium for automatically dividing traffic cells into road networks | |
| CN106844642B (en) | Method for calculating population density in road network grid based on GIS | |
| Dey et al. | Machine learning-based segmentation of aerial LiDAR point cloud data on building roof | |
| Kong et al. | A graph-based neural network approach to integrate multi-source data for urban building function classification | |
| CN115019163A (en) | City factor identification method based on multi-source big data | |
| Wu et al. | A non-rigid hierarchical discrete grid structure and its application to UAVs conflict detection and path planning | |
| Stal et al. | Classification of airborne laser scanning point clouds based on binomial logistic regression analysis | |
| CN112070787A (en) | Aviation three-dimensional point cloud plane segmentation method based on opponent reasoning theory | |
| Ruiz et al. | Automatic extraction of road intersections from images based on texture characterisation | |
| CN111080080A (en) | Method and system for estimating risk of geological disaster of villages and small towns |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200814 |