CN114241424A - Unmanned vehicle driving route planning system and method for surveying and mapping - Google Patents

Unmanned vehicle driving route planning system and method for surveying and mapping Download PDF

Info

Publication number
CN114241424A
CN114241424A CN202210145851.XA CN202210145851A CN114241424A CN 114241424 A CN114241424 A CN 114241424A CN 202210145851 A CN202210145851 A CN 202210145851A CN 114241424 A CN114241424 A CN 114241424A
Authority
CN
China
Prior art keywords
module
data
mapping
submodule
positioning
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202210145851.XA
Other languages
Chinese (zh)
Other versions
CN114241424B (en
Inventor
计宏
时英理
郭绪
谢丹
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jiangsu Wisdom Automobile Research Institute Co ltd
Original Assignee
Jiangsu Wisdom Automobile Research Institute Co ltd
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
Application filed by Jiangsu Wisdom Automobile Research Institute Co ltd filed Critical Jiangsu Wisdom Automobile Research Institute Co ltd
Priority to CN202210145851.XA priority Critical patent/CN114241424B/en
Publication of CN114241424A publication Critical patent/CN114241424A/en
Application granted granted Critical
Publication of CN114241424B publication Critical patent/CN114241424B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting

Abstract

The invention discloses a system and a method for planning the driving route of an unmanned vehicle for surveying and mapping, belonging to the technical field of unmanned vehicles for surveying and mapping. The system comprises a first data set building module, a curve fitting model building module, a random acquisition module, a real-time positioning module and a route planning module; the output end of the first data set building module is connected with the input end of the curve fitting model building module; the output end of the curve fitting model building module is connected with the input end of the random acquisition module; the output end of the random acquisition module is connected with the input end of the real-time positioning module; the output end of the real-time positioning module is connected with the input end of the route planning module. The invention can solve the problems of personnel shortage, low inspection precision and poor security measures in the field of inspection in scenic spots, and greatly improves the intellectualization and scientific and technological level in the field of inspection.

Description

Unmanned vehicle driving route planning system and method for surveying and mapping
Technical Field
The invention relates to the technical field of surveying and mapping inspection unmanned vehicles, in particular to a system and a method for planning a driving route of an unmanned vehicle for surveying and mapping inspection.
Background
At abroad, products similar to the unmanned patrol cars are only used in military patrol scenes at present, products in the civil field are not available, and domestic unmanned products with ground-air integration function combination also belong to the blank field. The unmanned patrol industry is not developed enough, according to statistics, the number of nationwide parks, campuses and cells is nearly 100 thousands, the number of closed scenic spots is nearly 10 thousands, and the number of wharfs and closed industrial parks is nearly 20 thousands, but most unmanned vehicles are not applied to patrol.
Surveying and mapping is generally understood as measuring and mapping, and in surveying and mapping inspection unmanned vehicles, a computer technology, a photoelectric technology, a network communication technology, space science and information science are taken as the basis, a global navigation satellite positioning system (GNSS), Remote Sensing (RS) and a Geographic Information System (GIS) are taken as the technical core, existing characteristic points and boundary lines on the ground are selected, and the graph and position information reflecting the current situation of the ground are obtained through a measuring means, so that the surveying and mapping inspection unmanned vehicles are used for engineering construction, planning and design and administrative management.
In a large number of regions in the country, monitoring points are arranged, but most of the regions are not provided with unmanned vehicle inspection, but only manual inspection is utilized, but with the disappearance of population dividends, the security industry can face a loss rate as high as 55%, many units are more and more difficult to bring qualified security, and more than 60% of the existing employees are expected to be over 55 years old. This kind of operational environment is boring, the work post that repeatability is high will hardly attract outstanding young strength to add in the future, consequently, it is the trade development trend to utilize the unmanned car to patrol and examine the security protection, to the garden that the unmanned car patrols and examines that a small part is equipped with, the control position coverage area that it can provide is little, pixel precision is low, there is the angle blind area, for example crowd gathers in the garden and smokes in the corner, the back shadow or side can often only be shot to the control position, unable discernment crowd's specific action, and the cigarette end that drops after the smoking very easily arouses the conflagration, cause unnecessary loss.
Disclosure of Invention
The invention aims to provide a system and a method for planning the driving route of an unmanned vehicle for surveying and mapping, so as to solve the problems in the background technology.
In order to solve the technical problems, the invention provides the following technical scheme:
an unmanned vehicle driving route planning system for surveying and mapping inspection comprises a first data set building module, a curve fitting model building module, a random acquisition module, a real-time positioning module and a route planning module;
the first data set construction module is used for acquiring action acquisition data and interval time data of historical people who lift arms and put down the arms to form a first data set; the curve fitting model building module is used for building a curve fitting model according to the data of the first data set, analyzing the arm action interval time of the person in the smoking process and outputting the curve fitting model; the random acquisition module is used for setting monitoring points in the park, acquiring arm action data and time data of people at all places in the park and judging whether smoking people exist in the people at all places in the park; the real-time positioning module is used for acquiring the positioning of the smoking crowd in the garden and transmitting the positioning to the route planning module; the route planning module is used for constructing a mapping and inspecting unmanned vehicle route model, generating a driving route of the mapping and inspecting unmanned vehicle, and commanding the mapping and inspecting unmanned vehicle to drive and plan the driving route;
the output end of the first data set building module is connected with the input end of the curve fitting model building module; the output end of the curve fitting model building module is connected with the input end of the random acquisition module; the output end of the random acquisition module is connected with the input end of the real-time positioning module; the output end of the real-time positioning module is connected with the input end of the route planning module.
According to the technical scheme, the first data set construction module comprises a historical crowd data acquisition submodule and a training set construction submodule;
the historical crowd data acquisition submodule is used for acquiring action acquisition data and interval time data of the historical crowd after the arms are lifted and put down; the training set construction submodule is used for constructing a first data set according to the data acquired by the historical crowd data acquisition submodule and taking the first data set as a training set;
the output end of the historical crowd data acquisition submodule is connected with the input end of the training set construction submodule; and the output end of the training set construction submodule is connected with the input end of the curve fitting model construction module.
According to the technical scheme, the curve fitting model building module comprises a curve fitting model building submodule and a model output submodule;
the curve fitting model building submodule is used for obtaining data in the first data set building module and building a curve fitting model; the model output submodule is used for outputting a model training result according to the curve fitting model;
the output end of the curve fitting model building submodule is connected with the input end of the model output submodule; and the output end of the model output submodule is connected with the input end of the random acquisition module.
According to the technical scheme, the random acquisition module comprises a monitoring point position setting submodule and a random acquisition submodule;
the monitoring point location setting submodule is used for setting a monitoring point location in the park; the random acquisition submodule is used for randomly acquiring arm action data and time data of people at all places of the park according to the monitoring point positions and judging whether smoking people exist in all places of the park;
the output end of the monitoring point location setting submodule is connected with the input end of the random acquisition submodule; and the output end of the random acquisition submodule is connected with the input end of the real-time positioning module.
According to the technical scheme, the real-time positioning module comprises a real-time positioning sub-module and a transmission sub-module;
the real-time positioning sub-module is used for positioning the smoking people in the garden in real time to obtain positioning information; the transmission sub-module is used for transmitting the positioning information to the route planning module;
the output end of the real-time positioning sub-module is connected with the input end of the transmission sub-module; and the output end of the transmission submodule is connected with the input end of the route planning module.
According to the technical scheme, the route planning module comprises a mapping sub-module, a positioning frequency recording sub-module and a route planning sub-module;
the surveying and mapping submodule is used for surveying and mapping the garden to generate a garden survey map when the surveying and mapping inspection unmanned vehicle reaches the garden for the first time; the positioning frequency recording submodule is used for acquiring the positioning frequency of the same person in the community and recording the positioning frequency; the route planning submodule is used for generating a driving route according to the positioning occurrence times of the same person in the garden crowd and commanding, surveying and mapping and patrolling unmanned vehicles to drive;
the output end of the mapping sub-module is connected with the input end of the positioning frequency recording sub-module; and the output end of the positioning frequency recording submodule is connected with the input end of the route planning submodule.
In above-mentioned technical scheme, survey and drawing the sub-module can be launched when unmanned car first arrived a garden district in survey and drawing, utilizes the survey and drawing sub-module to survey and draw all routes in the garden, generates a garden survey and drawing, and all subsequent patrol routes all generate according to the data in the garden survey and drawing.
A method for planning the driving route of an unmanned vehicle for surveying and mapping inspection comprises the following steps:
s1, constructing a first data set, wherein the first data set is collected data and interval time data of actions of historical people for lifting and lowering arms;
s2, constructing a curve fitting model by using the first data set, analyzing the arm action interval time of the person in the smoking process, and outputting the curve fitting model;
s3, constructing a detection module, monitoring the garden, acquiring arm action data and time data of people at all places of the garden, judging whether people at all places of the garden have smoking people, and positioning the smoking people at the garden in real time;
s4, obtaining real-time positioning of smoking people in the garden, inputting the positioning into the mapping and inspection unmanned vehicle route model, generating a driving route of the mapping and inspection unmanned vehicle, and commanding the mapping and inspection unmanned vehicle to drive.
According to the technical scheme, in step S1, the first data set includes collected data of actions of historical people of raising and lowering arms and interval time data, a collection frequency is set to be M, and a person stops until the collection frequency meets M, and the data is recorded in the first data set.
According to the above technical solution, in step S2, the curve fitting model further includes:
s9-1, randomly collecting historical crowd actions as a first data set, obtaining M times of arm lifting and lowering actions, recording the interval time of every two adjacent actions, obtaining the average value of the interval time, and recording the average value as
Figure 100002_DEST_PATH_IMAGE001
Wherein M is a settable constant;
s9-2, taking the average value of the interval time as the value of X, taking Y as the result output, and constructing a training set, wherein Y is smoking or non-smoking, wherein smoking is represented by 1, and non-smoking is represented by-1; the training set is
Figure 18686DEST_PATH_IMAGE002
Wherein
Figure 100002_DEST_PATH_IMAGE003
、
Figure 512115DEST_PATH_IMAGE004
、
Figure 100002_DEST_PATH_IMAGE005
Is a concrete expression after the value of X is taken,
Figure 657926DEST_PATH_IMAGE006
、
Figure 100002_DEST_PATH_IMAGE007
、
Figure 236806DEST_PATH_IMAGE008
is a concrete expression after the value of Y is taken,
Figure 100002_DEST_PATH_IMAGE009
；
s9-3, initializing weight distribution of training data in the training set, endowing each training data with the same weight, and recording as:
Figure 100002_DEST_PATH_IMAGE011
wherein, is
Figure 939314DEST_PATH_IMAGE012
The weight value of the ith training data; the ratio of i =1, 2,
Figure 100002_DEST_PATH_IMAGE013
；
the initial training set weight distribution is noted as:
Figure 100002_DEST_PATH_IMAGE015
s9-4, performing an iteration, t =1, 2,
Figure 818408DEST_PATH_IMAGE016
(ii) a Wherein T represents the T-th iteration, and T represents the highest iteration time;
s9-4-1, selecting a weak classifier with the lowest current error rate
Figure 100002_DEST_PATH_IMAGE017
As the t-th basic classifier
Figure 553363DEST_PATH_IMAGE018
And calculating the weak classifier
Figure 619539DEST_PATH_IMAGE017
Error of (2):
Figure 100002_DEST_PATH_IMAGE019
wherein the content of the first and second substances,
Figure 719213DEST_PATH_IMAGE020
is represented by
Figure 249552DEST_PATH_IMAGE018
The sum of the sample data weights for the misclassification,
Figure 100002_DEST_PATH_IMAGE021
representing the weight value of the ith training data in the t iteration;
Figure 737165DEST_PATH_IMAGE022
is represented in
Figure 100002_DEST_PATH_IMAGE023
Basic classifier
Figure 290637DEST_PATH_IMAGE018
The classification result is obtained;
s9-4-2, calculating weak classifier
Figure 990740DEST_PATH_IMAGE017
Weight in the final classifier:
Figure 100002_DEST_PATH_IMAGE025
wherein the content of the first and second substances,
Figure 250951DEST_PATH_IMAGE026
is a weak classifier
Figure 909466DEST_PATH_IMAGE017
Weight is taken up in the final classifier;
s9-4-3, updating weight distribution of training data:
Figure 746972DEST_PATH_IMAGE028
wherein the content of the first and second substances,
Figure 100002_DEST_PATH_IMAGE029
for the training set weight distribution at the t-th iteration,
Figure 781924DEST_PATH_IMAGE030
the weight distribution of the training set after the t iteration, namely the weight distribution of the training set during the t +1 iteration;
Figure 100002_DEST_PATH_IMAGE031
is a normalization constant;
Figure 100002_DEST_PATH_IMAGE033
s9-5, combining according to the weight of each weak classifier in the final classifier to generate a strong classifier:
Figure 100002_DEST_PATH_IMAGE035
wherein the content of the first and second substances,
Figure 772008DEST_PATH_IMAGE036
representing the weight of each weak classifier in the final classifier;
s9-6, and
Figure 100002_DEST_PATH_IMAGE037
and (3) analyzing the processes of M times of arm lifting and arm lowering actions, classifying the average value of interval time, judging whether the current person smokes, if the output is 1, indicating that the person smokes, and if the output is-1, indicating that the person does not smoke.
In the technical scheme, by utilizing the ensemble learning technology of reinforcement learning, a weak learner with the prediction precision slightly higher than the random guess degree can be reinforced into a strong learner with the high prediction precision, the problem that the strong learner is difficult to construct directly can be solved, in the smoking judgment process, the weight of a sample which is wrongly classified by the previous basic classifier is increased, the weight of a sample which is correctly classified is reduced, and the previous basic classifier is used for training the next basic classifier again. At the same time, a new weak classifier is added in each iteration, and the final strong classifier is not determined until a predetermined error rate is reached or a pre-specified maximum number of iterations is reached. The accuracy of the reinforcing judgement that utilizes such mode can be very big satisfies the needs of high accuracy, can effectively avoid the spurious triggering of some similar actions, for example push away the confusion of actions such as glasses, cover mouth with smoking action, realizes high-tech standard.
According to the above technical solution, in the steps S3-S4, the method further comprises:
surveying and mapping the garden to generate a garden survey map;
a detection module is constructed to monitor the park and acquire arm action data and time data of people at all places of the park;
substituting the curve fitting model with the input data, outputting a result, wherein if the output is 1, the output represents smoking, and the output is-1, and represents no smoking;
acquiring real-time positioning of smoking people in a park;
inputting the data into a mapping and inspection unmanned vehicle route model, and sequencing the mapping and inspection unmanned vehicle route model according to the positioning occurrence times of the same person in the garden crowd;
generating a driving route of the surveying and mapping inspection unmanned vehicle;
and commanding the surveying and mapping inspection unmanned vehicle to run according to the running route of the surveying and mapping inspection unmanned vehicle.
Smoking is a process, and the smoking is finished firstly and then finished, so that the positioning occurrence times of the same person in the population in the garden are sequenced, the positioning occurrence times are more, which indicates that the person is monitored for a period of time, and the person must be in the state of being finished firstly, and the place where the person with the most positioning occurrence times is selected as a first arrival place; and the rest are analogized in the same way to generate a driving route of the unmanned vehicle.
Compared with the prior art, the invention has the following beneficial effects:
1. the invention can solve the practical problem of population loss faced by the traditional security form, improve the creativity and technological sense of the security patrol working environment, attract excellent young force to be added, simultaneously, one unmanned patrol vehicle can realize 24-hour all-weather operation, one patrol vehicle can complete the patrol working tasks of about 10 security personnel, and the efficiency can be greatly improved;
2. the invention can avoid the defect that manual patrol can not carry out a large amount of accurate identification and control on target groups under a complex scene with more people flow, can effectively realize environment datamation by utilizing a big data intelligent AI technology, is beneficial to effectively and rapidly identifying danger factors in the complex environment by security personnel through screening and managing data, locks security targets in time and improves the safety guarantee capability.
3. The invention can solve the problem that occupational pain points with more blind areas and low pixels exist in the monitoring point positions in the current scene, accurately judges the smoking crowd in the garden according to the arm action frequency, realizes route planning and meets the fire protection requirement;
4. the invention can promote the growth of regional GDP, promote the development of industrial chain and the employment, and promote new business model and service by fusing with 5G technology.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention. In the drawings:
FIG. 1 is a schematic flow diagram of an unmanned vehicle driving route planning system for surveying and mapping inspection according to the present invention;
fig. 2 is a schematic step diagram of the unmanned vehicle driving route planning method for surveying and mapping.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1-2, the present invention provides a technical solution:
an unmanned vehicle driving route planning system for surveying and mapping inspection comprises a first data set building module, a curve fitting model building module, a random acquisition module, a real-time positioning module and a route planning module;
the first data set construction module is used for acquiring action acquisition data and interval time data of historical people who lift arms and put down the arms to form a first data set; the curve fitting model building module is used for building a curve fitting model according to the data of the first data set, analyzing the arm action interval time of the person in the smoking process and outputting the curve fitting model; the random acquisition module is used for setting monitoring points in the park, acquiring arm action data and time data of people at all places in the park and judging whether smoking people exist in the people at all places in the park; the real-time positioning module is used for acquiring the positioning of the smoking crowd in the garden and transmitting the positioning to the route planning module; the route planning module is used for constructing a mapping and inspecting unmanned vehicle route model, generating a driving route of the mapping and inspecting unmanned vehicle, and commanding the mapping and inspecting unmanned vehicle to drive and plan the driving route;
the output end of the first data set building module is connected with the input end of the curve fitting model building module; the output end of the curve fitting model building module is connected with the input end of the random acquisition module; the output end of the random acquisition module is connected with the input end of the real-time positioning module; the output end of the real-time positioning module is connected with the input end of the route planning module.
The first data set construction module comprises a historical crowd data acquisition submodule and a training set construction submodule;
the historical crowd data acquisition submodule is used for acquiring action acquisition data and interval time data of the historical crowd after the arms are lifted and put down; the training set construction submodule is used for constructing a first data set according to the data acquired by the historical crowd data acquisition submodule and taking the first data set as a training set;
the output end of the historical crowd data acquisition submodule is connected with the input end of the training set construction submodule; and the output end of the training set construction submodule is connected with the input end of the curve fitting model construction module.
The curve fitting model building module comprises a curve fitting model building submodule and a model output submodule;
the curve fitting model building submodule is used for obtaining data in the first data set building module and building a curve fitting model; the model output submodule is used for outputting a model training result according to the curve fitting model;
the output end of the curve fitting model building submodule is connected with the input end of the model output submodule; and the output end of the model output submodule is connected with the input end of the random acquisition module.
The random acquisition module comprises a monitoring point position setting submodule and a random acquisition submodule;
the monitoring point location setting submodule is used for setting a monitoring point location in the park; the random acquisition submodule is used for randomly acquiring arm action data and time data of people at all places of the park according to the monitoring point positions and judging whether smoking people exist in all places of the park;
the output end of the monitoring point location setting submodule is connected with the input end of the random acquisition submodule; and the output end of the random acquisition submodule is connected with the input end of the real-time positioning module.
The real-time positioning module comprises a real-time positioning sub-module and a transmission sub-module;
the real-time positioning sub-module is used for positioning the smoking people in the garden in real time to obtain positioning information; the transmission sub-module is used for transmitting the positioning information to the route planning module;
the output end of the real-time positioning sub-module is connected with the input end of the transmission sub-module; and the output end of the transmission submodule is connected with the input end of the route planning module.
The route planning module comprises a mapping sub-module, a positioning frequency recording sub-module and a route planning sub-module;
the surveying and mapping submodule is used for surveying and mapping the garden to generate a garden survey map when the surveying and mapping inspection unmanned vehicle reaches the garden for the first time; the positioning frequency recording submodule is used for acquiring the positioning frequency of the same person in the community and recording the positioning frequency; the route planning submodule is used for generating a driving route according to the positioning occurrence times of the same person in the garden crowd and commanding, surveying and mapping and patrolling unmanned vehicles to drive;
the output end of the mapping sub-module is connected with the input end of the positioning frequency recording sub-module; and the output end of the positioning frequency recording submodule is connected with the input end of the route planning submodule.
A method for planning the driving route of an unmanned vehicle for surveying and mapping inspection comprises the following steps:
s1, constructing a first data set, wherein the first data set is collected data and interval time data of actions of historical people for lifting and lowering arms;
s2, constructing a curve fitting model by using the first data set, analyzing the arm action interval time of the person in the smoking process, and outputting the curve fitting model;
s3, constructing a detection module, monitoring the garden, acquiring arm action data and time data of people at all places of the garden, judging whether people at all places of the garden have smoking people, and positioning the smoking people at the garden in real time;
s4, obtaining real-time positioning of smoking people in the garden, inputting the positioning into the mapping and inspection unmanned vehicle route model, generating a driving route of the mapping and inspection unmanned vehicle, and commanding the mapping and inspection unmanned vehicle to drive.
In step S1, the first data set includes collected data of actions of historical people when their arms are lifted and their arms are lowered, and interval time data, where a collection time is set to M, and the collection is stopped until the collection time meets M, and the data is recorded in the first data set.
In step S2, the curve fitting model further includes:
s9-1, randomly collecting historical crowd actions as a first data set, obtaining M times of arm lifting and lowering actions, recording the interval time of every two adjacent actions, obtaining the average value of the interval time, and recording the average value as
Figure 539106DEST_PATH_IMAGE001
Wherein M is a settable constant;
s9-2, taking the average value of the interval time as the value of X, taking Y as the result output, and constructing a training set, wherein Y is smoking or non-smoking, wherein smoking is represented by 1, and non-smoking is represented by-1; the training set is
Figure 191805DEST_PATH_IMAGE002
Wherein
Figure 499289DEST_PATH_IMAGE003
、
Figure 593147DEST_PATH_IMAGE004
、
Figure 327885DEST_PATH_IMAGE005
Is a concrete expression after the value of X is taken,
Figure 77666DEST_PATH_IMAGE006
、
Figure 188842DEST_PATH_IMAGE007
、
Figure 402785DEST_PATH_IMAGE008
is a concrete expression after the value of Y is taken,
Figure 308425DEST_PATH_IMAGE009
；
s9-3, initializing weight distribution of training data in the training set, endowing each training data with the same weight, and recording as:
Figure 670136DEST_PATH_IMAGE038
wherein, is
Figure 319423DEST_PATH_IMAGE012
The weight value of the ith training data; the ratio of i =1, 2,
Figure 387873DEST_PATH_IMAGE013
；
the initial training set weight distribution is noted as:
Figure DEST_PATH_IMAGE039
s9-4, performing an iteration, t =1, 2,
Figure 605359DEST_PATH_IMAGE016
(ii) a Wherein T represents the T-th iteration, and T represents the highest iteration time;
s9-4-1, selecting a weak classifier with the lowest current error rate
Figure 380331DEST_PATH_IMAGE017
As the t-th basic classifier
Figure 567729DEST_PATH_IMAGE018
And calculating the weak classifier
Figure 818582DEST_PATH_IMAGE017
Error of (2):
Figure 331603DEST_PATH_IMAGE019
wherein the content of the first and second substances,
Figure 340010DEST_PATH_IMAGE020
is represented by
Figure 331100DEST_PATH_IMAGE018
The sum of the sample data weights for the misclassification,
Figure 311826DEST_PATH_IMAGE021
representing the weight value of the ith training data in the t iteration;
Figure 730169DEST_PATH_IMAGE022
is represented in
Figure 225872DEST_PATH_IMAGE023
Basic classifier
Figure 348549DEST_PATH_IMAGE018
The classification result is obtained;
s9-4-2, calculating weak classifier
Figure 980519DEST_PATH_IMAGE017
Weight in the final classifier:
Figure 835342DEST_PATH_IMAGE040
wherein the content of the first and second substances,
Figure 818342DEST_PATH_IMAGE026
is a weak classifier
Figure 354496DEST_PATH_IMAGE017
Weight is taken up in the final classifier;
s9-4-3, updating weight distribution of training data:
Figure DEST_PATH_IMAGE041
wherein the content of the first and second substances,
Figure 44235DEST_PATH_IMAGE029
for the training set weight distribution at the t-th iteration,
Figure 69960DEST_PATH_IMAGE030
the weight distribution of the training set after the t iteration, namely the weight distribution of the training set during the t +1 iteration;
Figure 602572DEST_PATH_IMAGE031
is a normalization constant;
Figure 739155DEST_PATH_IMAGE033
s9-5, combining according to the weight of each weak classifier in the final classifier to generate a strong classifier:
Figure 345717DEST_PATH_IMAGE035
wherein the content of the first and second substances,
Figure 745606DEST_PATH_IMAGE036
representing the weight of each weak classifier in the final classifier;
s9-6, and
Figure 437618DEST_PATH_IMAGE037
and (3) analyzing the processes of M times of arm lifting and arm lowering actions, classifying the average value of interval time, judging whether the current person smokes, if the output is 1, indicating that the person smokes, and if the output is-1, indicating that the person does not smoke.
In steps S3-S4, the method further includes:
mapping the park to generate a park map
A detection module is constructed to monitor the park and acquire arm action data and time data of people at all places of the park;
substituting the curve fitting model with the input data, outputting a result, wherein if the output is 1, the output represents smoking, and the output is-1, and represents no smoking;
acquiring real-time positioning of smoking people in a park;
inputting the data into a mapping and inspection unmanned vehicle route model, and sequencing the mapping and inspection unmanned vehicle route model according to the positioning occurrence times of the same person in the garden crowd;
selecting the place of the person with the largest positioning occurrence frequency as a first arrival place;
generating a driving route of the surveying and mapping inspection unmanned vehicle;
and commanding the surveying and mapping inspection unmanned vehicle to run according to the running route of the surveying and mapping inspection unmanned vehicle.
In this embodiment:
surveying and mapping the garden to generate a garden survey map;
randomly collecting historical crowd actions as a first data set;
obtaining M times of arm lifting and lowering actions, recording the interval time of every two adjacent actions, obtaining the average value of the interval time, and recording the average value as
Figure 377892DEST_PATH_IMAGE001
Where M is a settable constant, setting M = 6;
taking the average value of the interval time as the value of X (taking the second as the unit), taking Y as the result output, obtaining 10 groups of data to construct a training set,
1 2 3 4 5 6 7 8 9 10
X 16 18 22 6 11 20 64 26 28 19
Y 1 1 1 -1 -1 1 -1 1 1 -1
y is smoking or non-smoking, wherein smoking is represented by 1, and non-smoking is represented by-1; the training set is
Figure 166857DEST_PATH_IMAGE002
Wherein
Figure 534384DEST_PATH_IMAGE003
、
Figure 713693DEST_PATH_IMAGE004
、
Figure 192079DEST_PATH_IMAGE005
Is a concrete expression after the value of X is taken,
Figure 710916DEST_PATH_IMAGE006
、
Figure 249344DEST_PATH_IMAGE007
、
Figure 915949DEST_PATH_IMAGE008
is a concrete expression after the value of Y is taken,
Figure 525922DEST_PATH_IMAGE009
；
s9-3, initializing weight distribution of training data in the training set, endowing each training data with the same weight, and recording as:
Figure 961583DEST_PATH_IMAGE011
wherein, is
Figure 670913DEST_PATH_IMAGE012
The weight value of the ith training data; the ratio of i =1, 2,
Figure 28076DEST_PATH_IMAGE013
；
each weight is 0.1;
the initial training set weight distribution is noted as:
Figure DEST_PATH_IMAGE043
an iteration is performed, t =1, 2,
Figure 51527DEST_PATH_IMAGE016
(ii) a Wherein T represents the T-th iteration, and T represents the highest iteration time;
selecting a weak classifier with the lowest current error rate
Figure 341694DEST_PATH_IMAGE017
As the t-th basic classifier
Figure 284242DEST_PATH_IMAGE018
And calculating the weak classifier
Figure 484078DEST_PATH_IMAGE017
Error of (2):
Figure 373537DEST_PATH_IMAGE019
wherein the content of the first and second substances,
Figure 721473DEST_PATH_IMAGE020
is represented by
Figure 772605DEST_PATH_IMAGE018
The sum of the sample data weights for the misclassification,
Figure 635519DEST_PATH_IMAGE021
representing the weight value of the ith training data in the t iteration;
Figure 797510DEST_PATH_IMAGE022
is represented in
Figure 62269DEST_PATH_IMAGE023
Basic classifier
Figure 549883DEST_PATH_IMAGE018
The classification result is obtained;
the weak classifier
Figure 962409DEST_PATH_IMAGE017
The selection is as follows:
Figure DEST_PATH_IMAGE045
then there are 7 and 10 data that are erroneous, so:
Figure DEST_PATH_IMAGE047
computing weak classifier
Figure 69037DEST_PATH_IMAGE017
Weight in the final classifier:
Figure DEST_PATH_IMAGE049
wherein the content of the first and second substances,
Figure 329248DEST_PATH_IMAGE026
is a weak classifier
Figure 987762DEST_PATH_IMAGE017
Weight is taken up in the final classifier;
updating the weight distribution of the training data for the next iteration:
Figure 825268DEST_PATH_IMAGE028
wherein the content of the first and second substances,
Figure 656958DEST_PATH_IMAGE029
for the training set weight distribution at the t-th iteration,
Figure 630730DEST_PATH_IMAGE030
the weight distribution of the training set after the t iteration, namely the weight distribution of the training set during the t +1 iteration;
Figure 460146DEST_PATH_IMAGE031
is a normalization constant;
Figure 519369DEST_PATH_IMAGE033
the weight updates for correctly classified training sample data 1, 2, 3, 4, 5, 6, 8, 9 (total of 8) are:
Figure DEST_PATH_IMAGE051
as can be seen, the weight of the correctly classified samples is reduced from 0.1 to
Figure 967799DEST_PATH_IMAGE052
；
The weights of the erroneous samples 7 and 10 are updated, so that:
Figure 61657DEST_PATH_IMAGE054
it can be seen that the weight of misclassified samples increases from 0.1 to
Figure DEST_PATH_IMAGE055
；
And so on;
and combining according to the weight occupied by each weak classifier in the final classifier to generate a strong classifier:
Figure 327553DEST_PATH_IMAGE035
wherein the content of the first and second substances,
Figure 874072DEST_PATH_IMAGE036
representing the weight of each weak classifier in the final classifier;
the method specifically comprises the following steps:
Figure DEST_PATH_IMAGE057
wherein the content of the first and second substances,
Figure 922931DEST_PATH_IMAGE058
to
Figure DEST_PATH_IMAGE059
Calculated in the same way, direct fitting can be done with software, examples not specifically shown;
to be provided with
Figure 277820DEST_PATH_IMAGE037
And (3) analyzing the processes of 6 arm lifting and lowering actions, classifying the average value of interval time, judging whether the current person smokes, if the output is 1, indicating that the person smokes, and if the output is-1, indicating that the person does not smoke.
Therefore, judgment can be carried out when the monitoring point positions face back to back or the direction is opposite to the crowd, and the accuracy degree is higher.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
Finally, it should be noted that: although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that changes may be made in the embodiments and/or equivalents thereof without departing from the spirit and scope of the invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (10)

1. The utility model provides a survey and drawing is patrolled and examined and is used unmanned vehicles route planning system that traveles which characterized in that: the system comprises a first data set building module, a curve fitting model building module, a random acquisition module, a real-time positioning module and a route planning module;
the first data set construction module is used for acquiring action acquisition data and interval time data of historical people who lift arms and put down the arms to form a first data set; the curve fitting model building module is used for building a curve fitting model according to the data of the first data set, analyzing the arm action interval time of the person in the smoking process and outputting the curve fitting model; the random acquisition module is used for setting monitoring points in the park, acquiring arm action data and time data of people at all places in the park and judging whether smoking people exist in the people at all places in the park; the real-time positioning module is used for acquiring the positioning of the smoking crowd in the garden and transmitting the positioning to the route planning module; the route planning module is used for constructing a mapping and inspecting unmanned vehicle route model, generating a driving route of the mapping and inspecting unmanned vehicle, and commanding the mapping and inspecting unmanned vehicle to drive and plan the driving route;
the output end of the first data set building module is connected with the input end of the curve fitting model building module; the output end of the curve fitting model building module is connected with the input end of the random acquisition module; the output end of the random acquisition module is connected with the input end of the real-time positioning module; the output end of the real-time positioning module is connected with the input end of the route planning module.
2. The unmanned vehicle driving route planning system for surveying and mapping inspection according to claim 1, wherein: the first data set construction module comprises a historical crowd data acquisition submodule and a training set construction submodule;
the historical crowd data acquisition submodule is used for acquiring action acquisition data and interval time data of the historical crowd after the arms are lifted and put down; the training set construction submodule is used for constructing a first data set according to the data acquired by the historical crowd data acquisition submodule and taking the first data set as a training set;
the output end of the historical crowd data acquisition submodule is connected with the input end of the training set construction submodule; and the output end of the training set construction submodule is connected with the input end of the curve fitting model construction module.
3. The unmanned vehicle driving route planning system for surveying and mapping inspection according to claim 1, wherein: the curve fitting model building module comprises a curve fitting model building submodule and a model output submodule;
the curve fitting model building submodule is used for obtaining data in the first data set building module and building a curve fitting model; the model output submodule is used for outputting a model training result according to the curve fitting model;
the output end of the curve fitting model building submodule is connected with the input end of the model output submodule; and the output end of the model output submodule is connected with the input end of the random acquisition module.
4. The unmanned vehicle driving route planning system for surveying and mapping inspection according to claim 1, wherein: the random acquisition module comprises a monitoring point position setting submodule and a random acquisition submodule;
the monitoring point location setting submodule is used for setting a monitoring point location in the park; the random acquisition submodule is used for randomly acquiring arm action data and time data of people at all places of the park according to the monitoring point positions and judging whether smoking people exist in all places of the park;
the output end of the monitoring point location setting submodule is connected with the input end of the random acquisition submodule; and the output end of the random acquisition submodule is connected with the input end of the real-time positioning module.
5. The unmanned vehicle driving route planning system for surveying and mapping inspection according to claim 1, wherein: the real-time positioning module comprises a real-time positioning sub-module and a transmission sub-module;
the real-time positioning sub-module is used for positioning the smoking people in the garden in real time to obtain positioning information; the transmission sub-module is used for transmitting the positioning information to the route planning module;
the output end of the real-time positioning sub-module is connected with the input end of the transmission sub-module; and the output end of the transmission submodule is connected with the input end of the route planning module.
6. The unmanned vehicle driving route planning system for surveying and mapping inspection according to claim 1, wherein: the route planning module comprises a mapping sub-module, a positioning frequency recording sub-module and a route planning sub-module;
the surveying and mapping submodule is used for surveying and mapping the garden to generate a garden survey map when the surveying and mapping inspection unmanned vehicle reaches the garden for the first time; the positioning frequency recording submodule is used for acquiring the positioning frequency of the same person in the community and recording the positioning frequency; the route planning submodule is used for generating a driving route according to the positioning occurrence times of the same person in the garden crowd and commanding, surveying and mapping and patrolling unmanned vehicles to drive;
the output end of the mapping sub-module is connected with the input end of the positioning frequency recording sub-module; and the output end of the positioning frequency recording submodule is connected with the input end of the route planning submodule.
7. The utility model provides a survey and drawing is patrolled and examined and is used unmanned vehicles route planning method which characterized in that: the method comprises the following steps:
s1, constructing a first data set, wherein the first data set is collected data and interval time data of actions of historical people for lifting and lowering arms;
s2, constructing a curve fitting model by using the first data set, analyzing the arm action interval time of the person in the smoking process, and outputting the curve fitting model;
s3, constructing a detection module, monitoring the garden, acquiring arm action data and time data of people at all places of the garden, judging whether people at all places of the garden have smoking people, and positioning the smoking people at the garden in real time;
s4, obtaining real-time positioning of smoking people in the garden, inputting the positioning into the mapping and inspection unmanned vehicle route model, generating a driving route of the mapping and inspection unmanned vehicle, and commanding the mapping and inspection unmanned vehicle to drive.
8. The unmanned vehicle driving route planning method for surveying and mapping inspection according to claim 7, wherein: in step S1, the first data set includes collected data of actions of historical people when their arms are lifted and their arms are lowered, and interval time data, where a collection time is set to M, and the collection is stopped until the collection time meets M, and the data is recorded in the first data set.
9. The unmanned vehicle driving route planning method for surveying and mapping inspection according to claim 8, wherein: in step S2, the curve fitting model further includes:
s9-1, randomly collecting historical crowd actions as a first data set, obtaining M times of arm lifting and lowering actions, recording the interval time of every two adjacent actions, obtaining the average value of the interval time, and recording the average value as
Figure DEST_PATH_IMAGE001
Wherein M is a settable constant;
s9-2, taking the average value of the interval time as the value of X, taking Y as the result output, and constructing a training set, wherein Y is smoking or non-smoking, wherein smoking is represented by 1, and non-smoking is represented by-1; the training set is
Figure 559244DEST_PATH_IMAGE002
Wherein
Figure DEST_PATH_IMAGE003
、
Figure 967223DEST_PATH_IMAGE004
、
Figure DEST_PATH_IMAGE005
Is a concrete expression after the value of X is taken,
Figure 244751DEST_PATH_IMAGE006
、
Figure DEST_PATH_IMAGE007
、
Figure 788996DEST_PATH_IMAGE008
is a concrete expression after the value of Y is taken,
Figure DEST_PATH_IMAGE009
；
s9-3, initializing weight distribution of training data in the training set, endowing each training data with the same weight, and recording as:
Figure DEST_PATH_IMAGE011
wherein, is
Figure 798672DEST_PATH_IMAGE012
The weight value of the ith training data; the ratio of i =1, 2,
Figure DEST_PATH_IMAGE013
；
the initial training set weight distribution is noted as:
Figure DEST_PATH_IMAGE015
s9-4, performing an iteration, t =1, 2,
Figure 710258DEST_PATH_IMAGE016
(ii) a Wherein T represents the T-th iteration, and T represents the highest iteration time;
s9-4-1, selecting a weak classifier with the lowest current error rate
Figure DEST_PATH_IMAGE017
As the t-th basic classifier
Figure 405858DEST_PATH_IMAGE018
And calculating the weak classifier
Figure 866926DEST_PATH_IMAGE017
Error of (2):
Figure DEST_PATH_IMAGE019
wherein the content of the first and second substances,
Figure 172137DEST_PATH_IMAGE020
is represented by
Figure 85866DEST_PATH_IMAGE018
The sum of the sample data weights for the misclassification,
Figure DEST_PATH_IMAGE021
representing the weight value of the ith training data in the t iteration;
Figure 829831DEST_PATH_IMAGE022
is represented in
Figure DEST_PATH_IMAGE023
Basic classifier
Figure 614248DEST_PATH_IMAGE018
The classification result is obtained;
s9-4-2, calculating weak classifier
Figure 887097DEST_PATH_IMAGE017
Weight in the final classifier:
Figure DEST_PATH_IMAGE025
wherein the content of the first and second substances,
Figure 694647DEST_PATH_IMAGE026
is a weak classifier
Figure 976724DEST_PATH_IMAGE017
Weight is taken up in the final classifier;
s9-4-3, updating weight distribution of training data:
Figure 412384DEST_PATH_IMAGE028
wherein the content of the first and second substances,
Figure DEST_PATH_IMAGE029
for the training set weight distribution at the t-th iteration,
Figure 731501DEST_PATH_IMAGE030
the weight distribution of the training set after the t iteration, namely the weight distribution of the training set during the t +1 iteration;
Figure DEST_PATH_IMAGE031
is a normalization constant;
Figure DEST_PATH_IMAGE033
s9-5, combining according to the weight of each weak classifier in the final classifier to generate a strong classifier:
Figure DEST_PATH_IMAGE035
wherein the content of the first and second substances,
Figure 901714DEST_PATH_IMAGE036
representing the weight of each weak classifier in the final classifier;
s9-6, and
Figure DEST_PATH_IMAGE037
and (3) analyzing the processes of M times of arm lifting and arm lowering actions, classifying the average value of interval time, judging whether the current person smokes, if the output is 1, indicating that the person smokes, and if the output is-1, indicating that the person does not smoke.
10. The unmanned vehicle driving route planning method for surveying and mapping inspection according to claim 9, wherein: in steps S3-S4, the method further includes:
surveying and mapping the garden to generate a garden survey map;
a detection module is constructed to monitor the park and acquire arm action data and time data of people at all places of the park;
substituting the curve fitting model with the input data, outputting a result, wherein if the output is 1, the output represents smoking, and the output is-1, and represents no smoking;
acquiring real-time positioning of smoking people in a park;
inputting the data into a mapping and inspection unmanned vehicle route model, and sequencing the mapping and inspection unmanned vehicle route model according to the positioning occurrence times of the same person in the garden crowd;
selecting the place of the person with the largest positioning occurrence frequency as a first arrival place;
generating a driving route of the surveying and mapping inspection unmanned vehicle;
and commanding the surveying and mapping inspection unmanned vehicle to run according to the running route of the surveying and mapping inspection unmanned vehicle.
CN202210145851.XA 2022-02-17 2022-02-17 Unmanned vehicle driving route planning system and method for surveying and mapping inspection Active CN114241424B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210145851.XA CN114241424B (en) 2022-02-17 2022-02-17 Unmanned vehicle driving route planning system and method for surveying and mapping inspection

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210145851.XA CN114241424B (en) 2022-02-17 2022-02-17 Unmanned vehicle driving route planning system and method for surveying and mapping inspection

Publications (2)

Publication Number Publication Date
CN114241424A true CN114241424A (en) 2022-03-25
CN114241424B CN114241424B (en) 2022-05-31

Family

ID=80747560

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210145851.XA Active CN114241424B (en) 2022-02-17 2022-02-17 Unmanned vehicle driving route planning system and method for surveying and mapping inspection

Country Status (1)

Country Link
CN (1) CN114241424B (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101554499B1 (en) * 2014-11-28 2015-09-21 금호마린테크 (주) System for planning optimized vessel seaway using visual interactive modeling
CN111914667A (en) * 2020-07-08 2020-11-10 浙江大华技术股份有限公司 Smoking detection method and device
CN112238452A (en) * 2019-07-19 2021-01-19 顺丰科技有限公司 Mechanical arm path planning method and device, industrial control equipment and storage medium
CN113239754A (en) * 2021-04-23 2021-08-10 泰山学院 Dangerous driving behavior detection and positioning method and system applied to Internet of vehicles
CN113987727A (en) * 2021-11-25 2022-01-28 浙江省送变电工程有限公司 Big data based cable laying planning system and method in slope environment

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101554499B1 (en) * 2014-11-28 2015-09-21 금호마린테크 (주) System for planning optimized vessel seaway using visual interactive modeling
CN112238452A (en) * 2019-07-19 2021-01-19 顺丰科技有限公司 Mechanical arm path planning method and device, industrial control equipment and storage medium
CN111914667A (en) * 2020-07-08 2020-11-10 浙江大华技术股份有限公司 Smoking detection method and device
CN113239754A (en) * 2021-04-23 2021-08-10 泰山学院 Dangerous driving behavior detection and positioning method and system applied to Internet of vehicles
CN113987727A (en) * 2021-11-25 2022-01-28 浙江省送变电工程有限公司 Big data based cable laying planning system and method in slope environment

Also Published As

Publication number Publication date
CN114241424B (en) 2022-05-31

Similar Documents

Publication Publication Date Title
CN106650825B (en) Motor vehicle exhaust emission data fusion system
CN110346517B (en) Smart city industrial atmosphere pollution visual early warning method and system
CN105206057B (en) Detection method and system based on Floating Car resident trip hot spot region
CN108446470A (en) Medical facilities analysis method of reachability based on track of vehicle data and population distribution
CN113240249B (en) Urban engineering quality intelligent evaluation method and system based on unmanned aerial vehicle augmented reality
CN115310361B (en) Underground coal mine dust concentration prediction method and system based on WGAN-CNN
CN110555551A (en) air quality big data management method and system for smart city
CN115775085A (en) Smart city management method and system based on digital twin
CN115691058A (en) Holographic three-dimensional networking landslide intelligent early warning method based on multiple monitoring occasions
CN116486289A (en) Gas pipeline high-consequence area identification method driven by multi-source data and knowledge
CN116227362A (en) Municipal drainage pipe network state prediction method based on graph and deep learning
CN113450573A (en) Traffic monitoring method and traffic monitoring system based on unmanned aerial vehicle image recognition
CN114662774A (en) City block vitality prediction method, storage medium and terminal
CN116415756B (en) Urban virtual scene experience management system based on VR technology
CN114241424B (en) Unmanned vehicle driving route planning system and method for surveying and mapping inspection
CN111400877B (en) Intelligent city simulation system and method based on GIS data
CN113902259A (en) Wet avalanche disaster risk assessment method and device
CN112097743A (en) Building earthquake-resistant characteristic analysis method based on unmanned aerial vehicle technology
CN112215059A (en) Urban village identification and population estimation method and system based on deep learning and computer readable storage medium
CN110765900A (en) DSSD-based automatic illegal building detection method and system
CN113033443B (en) Unmanned aerial vehicle-based automatic pedestrian crossing facility whole road network checking method
CN115840908A (en) Method for constructing PM2.5 three-dimensional dynamic monitoring field by microwave link based on LSTM model
CN114913447A (en) Police intelligent command room system and method based on scene recognition
CN114139773A (en) Public traffic flow prediction method based on space-time graph convolutional network
Shi et al. An innovative approach to determine building window-to-wall ratios for urban energy simulation

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant