CN108629309A - Foundation pit surrounding people's method for protecting - Google Patents
Foundation pit surrounding people's method for protecting Download PDFInfo
- Publication number
- CN108629309A CN108629309A CN201810404017.1A CN201810404017A CN108629309A CN 108629309 A CN108629309 A CN 108629309A CN 201810404017 A CN201810404017 A CN 201810404017A CN 108629309 A CN108629309 A CN 108629309A
- Authority
- CN
- China
- Prior art keywords
- foundation pit
- surrounding people
- region
- safety warning
- camera
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Multimedia (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Bioinformatics & Computational Biology (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Evolutionary Biology (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Human Computer Interaction (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Signal Processing (AREA)
- Alarm Systems (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses foundation pit surrounding people's method for protecting, include the following steps：S1：To foundation pit neighboring area setting safety warning region；S2：Camera is set up around foundation pit, ensures that the photo that camera takes can include foundation pit and warning region simultaneously；S3：All cameras set up around foundation pit are connect with remote control center, remote control center adjusts monitoring area according to foundation pit ambient conditions and delimit foundation pit region；S4：Remote control center utilizes the picture of camera acquisition, pass through pedestrian detection algorithm, detect foundation pit surrounding people, and obtain the relative position between the surrounding people detected and safety warning region, judge whether to send a warning message, the probability that few sample class is calculated in prototype network has been aggravated in pedestrian detection algorithm therein.The present invention improves accuracy and the speed of foundation pit surrounding people detection by above-mentioned principle, quickly timely reminds foundation pit surrounding people, guarantee foundation pit surrounding people's safety of high-reliability.
Description
Technical field
The present invention relates to personnel safety fields, and in particular to foundation pit surrounding people's method for protecting.
Background technology
It is ensureing that the accuracy and reliability of foundation pit surrounding people safety is influenced by detection technique in the prior art, is examining
The accuracy and speed of survey technology can all influence the accuracy of foundation pit surrounding people's safety, if detection technique can not accurately identify base
Hole surrounding people then will appear warning and the situation of mistake reminded to occur, and can not ensure the safety of foundation pit surrounding people.Existing
In object detection technology, there are two types of technologies.One is the positions for obtaining a large amount of object first and being likely to occur, and then input to one
A convolutional neural networks.It is another then be that voluminous object position that may be present is directly inputed to a grader.The first
Mode precision higher, but speed is slower.The second way, speed, but precision higher.In order to maintain one it is very fast
Speed, while not losing object detection precision, we mainly use second of object detection mode in the prior art.Using
When second of object is detected, there is also the situation that target detection is not allowed, the main original for causing target detection inaccurate at this stage
Because being：First, extremely unbalanced positive and negative sample proportion.When training grader, negative sample quantity is far longer than positive sample
This quantity.Second, entire neural network during training, by the negative sample of easy differentiation dominated by gradient.It was training
Cheng Zhong, although the loss very littles that the negative sample for being easy to distinguish is brought finally account for the exhausted of entire loss due to large number of
It is most of.So as to cause less than one ideal result of convergence.
Invention content
The technical problem to be solved by the present invention is to improve the accuracy of foundation pit surrounding people detection and speed, it is therefore intended that
Foundation pit surrounding people's method for protecting is provided, accuracy and the speed of foundation pit surrounding people detection is improved, quickly timely carries
Wake up foundation pit surrounding people, and foundation pit surrounding people is avoided to fall into foundation pit, guarantee foundation pit surrounding people's safety of high-reliability.
The present invention is achieved through the following technical solutions：
Foundation pit surrounding people's method for protecting, which is characterized in that include the following steps：
S1：To foundation pit neighboring area setting safety warning region；
S2：Camera is set up around foundation pit, ensures that the photo that camera takes can include foundation pit and warning simultaneously
Region；
S3：All cameras set up around foundation pit are connect with remote control center, remote control center is according to foundation pit
Ambient conditions adjust monitoring area and delimit foundation pit region；
S4：Remote control center detects foundation pit people around using the picture of camera acquisition by pedestrian detection algorithm
Member, and the relative position between the surrounding people detected and safety warning region is obtained, if detecting surrounding people and certain
There is coincidence in level-one safety warning region, then sends out corresponding warning message, model net has been aggravated in pedestrian detection algorithm therein
The probability of few sample class is calculated in network.
It is ensureing that the accuracy and reliability of foundation pit surrounding people safety is influenced by detection technique in the prior art, is examining
The accuracy and speed of survey technology can all influence the accuracy of foundation pit surrounding people's safety, if detection technique can not accurately identify base
Hole surrounding people then will appear warning and the situation of mistake reminded to occur, and can not ensure the safety of foundation pit surrounding people.Existing
In object detection technology, there are two types of technologies.One is the positions for obtaining a large amount of object first and being likely to occur, and then input to one
A convolutional neural networks.It is another then be that voluminous object position that may be present is directly inputed to a grader.The first
Mode precision higher, but speed is slower.The second way, speed, but precision higher.In order to maintain one it is very fast
Speed, while not losing object detection precision, we mainly use second of object detection mode in the prior art.Using
When second of object is detected, there is also the situation that target detection is not allowed, the main original for causing target detection inaccurate at this stage
Because being：First, extremely unbalanced positive and negative sample proportion.When training grader, negative sample quantity is far longer than positive sample
This quantity.Second, entire neural network during training, by the negative sample of easy differentiation dominated by gradient.It was training
Cheng Zhong, although the loss very littles that the negative sample for being easy to distinguish is brought finally account for the exhausted of entire loss due to large number of
It is most of.So as to cause less than one ideal result of convergence.
The present invention is overcome in the prior art by rationally designing Loss functions due to extremely unbalanced positive negative sample ratio
Caused by example when training grader, people caused by the case where negative sample quantity is far longer than positive sample quantity occurs
The not high problem of member's accuracy of detection, it is automatic to realize to base under the premise of improving the accuracy and speed that foundation pit surrounding people detects
Foundation pit surrounding people information, is then compared by the quick accurate detection for cheating surrounding people with safety warning region, if visited
It measures surrounding people to overlap with a certain level security warning zone, then sends out corresponding warning message, quickly timely remind base
Surrounding people is cheated, foundation pit surrounding people is avoided to fall into foundation pit, guarantee foundation pit surrounding people's safety of high-reliability.
Preferably, the data set that pedestrian detection algorithm is acquired using camera in step S4 is as training data and loss letter
Number, which is brought into together in prototype network framework, to be trained, and camera acquisition picture is finally brought directly to the network rack of training completion
The i.e. detectable foundation pit surrounding people's situation of structure, loss function therein is Loss (p)=- (1-p) alog (p), what P was referred to be by
Detect object be pedestrian probability, (1-p) a be regulatory factor, 2<a<3.
This programme uses a kind of more outstanding loss function definition mode Loss (p)=- (1-p) alog (p), and P refers to
Generation be detected object be pedestrian probability, (1-p) a be regulatory factor, 2<a<3, a value range is to pass through test of many times
The optimum range obtained afterwards, the loss function can increase the positive sample for being not easy to distinguish in loss due to the addition of regulatory factor
Weight in the middle, meanwhile, the leading role that the negative sample of numerous easy differentiations plays model training can be also eliminated, it is final to make
Model can converge to a preferably solution in the training process, to considerably enhance the accuracy and speed of object detection,
Effective guarantee is provided for the automatic accurate prompting foundation pit surrounding people in later stage.
Preferably, safety warning region is divided into three-level warning in step S1, is striding into danger zone i.e. from foundation pit
When within the scope of 4 meters to 5 meters, but from foundation pit also farther out when, be set as level-one safety warning region；From foundation pit closer to danger area
When domain is i.e. from foundation pit within the scope of 2 meters to 3 meters, it is set as secondary safety warning zone.Closer to foundation pit edge, all may at any time
When the region fallen is less than 2 meters, it is set as three-level safety warning region.
Compared with prior art, the present invention having the following advantages and advantages：
1, the present invention is overcome in the prior art by rationally designing Loss functions due to extremely unbalanced positive negative sample
Caused by ratio when training grader, caused by the case where negative sample quantity is far longer than positive sample quantity, occurs
The not high problem of personnel's accuracy of detection, under the premise of improving the accuracy and speed that foundation pit surrounding people detects, automatic realization pair
Foundation pit surrounding people information, is then compared by the quick accurate detection of foundation pit surrounding people with safety warning region, if
It detects surrounding people to overlap with a certain level security warning zone, then sends out corresponding warning message, quickly timely remind
Foundation pit surrounding people avoids foundation pit surrounding people from falling into foundation pit, guarantee foundation pit surrounding people's safety of high-reliability.
2, present invention employs a kind of more outstanding loss function definition mode Loss (p)=- (1-p) alog (p), should
Loss function can increase weight of the sample for being not easy to distinguish in loss due to the addition of regulatory factor (1-p) a, meanwhile,
Also the leading role that the negative sample of numerous easy differentiations plays model training can be eliminated, finally makes model in the training process
A preferably solution can be converged to, is the automatic accurate of later stage to considerably enhance the accuracy and speed of object detection
Foundation pit surrounding people is reminded to provide effective guarantee.
Description of the drawings
Attached drawing described herein is used for providing further understanding the embodiment of the present invention, constitutes one of the application
Point, do not constitute the restriction to the embodiment of the present invention.In the accompanying drawings：
Fig. 1 is the flow chart of the present invention.
Specific implementation mode
To make the objectives, technical solutions, and advantages of the present invention clearer, with reference to embodiment and attached drawing, to this
Invention is described in further detail, and exemplary embodiment of the invention and its explanation are only used for explaining the present invention, do not make
For limitation of the invention.
Embodiment 1：
As shown in Figure 1, the present invention includes foundation pit surrounding people's method for protecting, include the following steps：
S1：To foundation pit neighboring area setting safety warning region；
S2：Camera is set up around foundation pit, ensures that the photo that camera takes can include foundation pit and warning simultaneously
Region；
S3：All cameras set up around foundation pit are connect with remote control center, remote control center is according to foundation pit
Ambient conditions adjust monitoring area and delimit foundation pit region；
S4：Remote control center detects foundation pit people around using the picture of camera acquisition by pedestrian detection algorithm
Member, and the relative position between the surrounding people detected and safety warning region is obtained, if detecting surrounding people and certain
There is coincidence in level-one safety warning region, then sends out corresponding warning message, model net has been aggravated in pedestrian detection algorithm therein
The probability of few sample class is calculated in network.
It is ensureing that the accuracy and reliability of foundation pit surrounding people safety is influenced by detection technique in the prior art, is examining
The accuracy and speed of survey technology can all influence the accuracy of foundation pit surrounding people's safety, if detection technique can not accurately identify base
Hole surrounding people then will appear warning and the situation of mistake reminded to occur, and can not ensure the safety of foundation pit surrounding people.Existing
In object detection technology, there are two types of technologies.One is the positions for obtaining a large amount of object first and being likely to occur, and then input to one
A convolutional neural networks.It is another then be that voluminous object position that may be present is directly inputed to a grader.The first
Mode precision higher, but speed is slower.The second way, speed, but precision higher.In order to maintain one it is very fast
Speed, while not losing object detection precision, we mainly use second of object detection mode in the prior art.Using
When second of object is detected, there is also the situation that target detection is not allowed, the main original for causing target detection inaccurate at this stage
Because being：First, extremely unbalanced positive and negative sample proportion.When training grader, negative sample quantity is far longer than positive sample
This quantity.Second, entire neural network during training, by the negative sample of easy differentiation dominated by gradient.It was training
Cheng Zhong, although the loss very littles that the negative sample for being easy to distinguish is brought finally account for the exhausted of entire loss due to large number of
It is most of.So as to cause less than one ideal result of convergence.
The present invention is overcome in the prior art by rationally designing Loss functions due to extremely unbalanced positive negative sample ratio
Caused by example when training grader, people caused by the case where negative sample quantity is far longer than positive sample quantity occurs
The not high problem of member's accuracy of detection, it is automatic to realize to base under the premise of improving the accuracy and speed that foundation pit surrounding people detects
Foundation pit surrounding people information, is then compared by the quick accurate detection for cheating surrounding people with safety warning region, if visited
It measures surrounding people to overlap with a certain level security warning zone, then sends out corresponding warning message, quickly timely remind base
Surrounding people is cheated, foundation pit surrounding people is avoided to fall into foundation pit, guarantee foundation pit surrounding people's safety of high-reliability.
Embodiment 2：
The present embodiment is preferably as follows on the basis of embodiment 1：Pedestrian detection algorithm is acquired using camera in step S4
Picture obtain data set and bring into together in prototype network framework as training data and loss function to be trained, will finally take the photograph
As head acquisition picture is brought directly to the i.e. detectable foundation pit surrounding people's situation of the network architecture of training completion, loss function therein
For Loss (p)=- (1-p)aThat log (p), P are referred to is the probability that detected object is pedestrian, (1-p)aFor regulatory factor, 2<a<
3.Data set is to obtain the set of detection target from the target in interception image in complicated everyday scenes.Loss function is to use
Deviation between the measurement model detection object probability value provided and actual value.The penalty values that model is provided according to loss function
Network parameter is optimized.Simultaneously, it is contemplated that in training sample, the extreme of positive negative sample is uneven, in initialization god
When through network, the probability that few sample class is calculated in model has been aggravated so that training process is more stablized, and obtains most
Network architecture is completed in training eventually.The network architecture overcomes the unbalanced positive and negative sample proportion of the extreme occurred in the prior art and leads
The entire neural network caused during training, by the negative sample of easy differentiation dominated by gradient, and the precision occurred is not high
The problem of.
Detected object how is accurately obtained in pedestrian detection algorithm and is very important part, is examined in existing object
In survey technology, there are two types of technologies.One is the positions for obtaining a large amount of object first and being likely to occur, and then input to a convolution
Neural network.It is another then be that voluminous object position that may be present is directly inputed to a grader.First way essence
Higher is spent, but speed is slower.The second way, speed, but precision higher.
In order to maintain a faster speed, while object detection precision is not lost, we mainly adopt in the prior art
With second of object detection mode.When being detected using second of object, there is also the situation that target detection is not allowed, existing rank
The main reason for Duan Zaocheng target detections are inaccurate be：First, extremely unbalanced positive and negative sample proportion.Training grader when
It waits, negative sample quantity is far longer than positive sample quantity.Second, for entire neural network during training, gradient is by easy area
The negative sample divided is dominated.In the training process, although the loss very littles that the negative sample for being easy to distinguish is brought, due to quantity
It is numerous, finally account for the overwhelming majority of entire loss.So as to cause less than one ideal result of convergence.
In the conventional technology, loss function Loss (p)=- log (p) used for two classification problems is standard
Cross entropy, P represent detected object as the probability of pedestrian.For the sample (p for being easy to distinguish>>0.5), they bring
Loss it is also not small.When there is a large amount of sample for being easy to distinguish in training sample, it is not easy the sample distinguished on a small quantity, then when
After a large amount of sample loss summations for being easy to distinguish, these samples for being easy to distinguish will hide the sample for being not easy to distinguish
This.So that model for be not easy distinguish sample discrimination it is not high, eventually lead to neural network training when not
A more satisfactory result can be converged to.
This programme use a kind of more outstanding loss function definition mode Loss (p)=- (1-p)aLog (p), P refer to
Generation be detected object be pedestrian probability, (1-p)aFor regulatory factor, 2<a<3, a value range is to pass through test of many times
The optimum range obtained afterwards, the loss function can increase the sample for being not easy to distinguish and work as in loss due to the addition of regulatory factor
In weight, meanwhile, the leading role that the negative sample of numerous easy differentiations plays model training can be also eliminated, finally so that mould
Type can converge to a preferably solution in the training process, to considerably enhance the accuracy and speed of object detection, be
The automatic accurate prompting foundation pit surrounding people in later stage provides effective guarantee.
Safety warning region is divided into three-level warning in step S1, is striding into danger zone i.e. from foundation pit at 4 meters to 5
Rice range in when, but from foundation pit also farther out when, be set as level-one safety warning region；From foundation pit closer to danger zone i.e. from base
When hole is within the scope of 2 meters to 3 meters, it is set as secondary safety warning zone.Closer to foundation pit edge, may all fall at any time
Region i.e. be less than 2 meters when, be set as three-level safety warning region.It is alerted in the regional extent and has been fully able to meet protection
Foundation pit surrounding people's security needs.
Above-described specific implementation mode has carried out further the purpose of the present invention, technical solution and advantageous effect
It is described in detail, it should be understood that the foregoing is merely the specific implementation mode of the present invention, is not intended to limit the present invention
Protection domain, all within the spirits and principles of the present invention, any modification, equivalent substitution, improvement and etc. done should all include
Within protection scope of the present invention.
Claims (3)
1. foundation pit surrounding people's method for protecting, which is characterized in that include the following steps：
S1：To foundation pit neighboring area setting safety warning region；
S2：Camera is set up around foundation pit, ensures that the photo that camera takes can include foundation pit and warning region simultaneously；
S3：All cameras set up around foundation pit are connect with remote control center, remote control center according to foundation pit around
Situation adjusts monitoring area and delimit foundation pit region；
S4：Remote control center detects foundation pit surrounding people using the picture of camera acquisition by pedestrian detection algorithm,
And the relative position between the surrounding people detected and safety warning region is obtained, if detecting surrounding people and certain level-one
There is coincidence in safety warning region, then sends out corresponding warning message, prototype network meter has been aggravated in pedestrian detection algorithm therein
It calculates and obtains the probability of few sample class.
2. foundation pit surrounding people method for protecting according to claim 1, which is characterized in that pedestrian detection in step S4
The data set that algorithm is acquired using camera is brought into as training data and loss function in prototype network framework and is instructed together
Practice, camera acquisition picture be finally brought directly to the i.e. detectable foundation pit surrounding people's situation of the network architecture of training completion,
In loss function be Loss (p)=- (1-p)aThat log (p), P are referred to is the probability that detected object is pedestrian, (1-p)aFor
Regulatory factor, 2<a<3.
3. foundation pit surrounding people method for protecting according to claim 1, which is characterized in that safety warning in step S1
Region be divided into three-level warning, when having striden into danger zone i.e. from foundation pit within the scope of 4 meters to 5 meters, but from foundation pit also compared with
When remote, it is set as level-one safety warning region；From foundation pit closer to danger zone i.e. from foundation pit within the scope of 2 meters to 3 meters when, if
For secondary safety warning zone.When the region that closer to foundation pit edge, may all fall at any time is less than 2 meters, it is set as
Three-level safety warning region.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810404017.1A CN108629309A (en) | 2018-04-28 | 2018-04-28 | Foundation pit surrounding people's method for protecting |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810404017.1A CN108629309A (en) | 2018-04-28 | 2018-04-28 | Foundation pit surrounding people's method for protecting |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108629309A true CN108629309A (en) | 2018-10-09 |
Family
ID=63695098
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810404017.1A Pending CN108629309A (en) | 2018-04-28 | 2018-04-28 | Foundation pit surrounding people's method for protecting |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108629309A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109657716A (en) * | 2018-12-12 | 2019-04-19 | 天津卡达克数据有限公司 | A kind of vehicle appearance damnification recognition method based on deep learning |
| CN113256926A (en) * | 2021-05-11 | 2021-08-13 | 仲永东 | Active fence system based on construction safety protection |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101016053A (en) * | 2007-01-25 | 2007-08-15 | 吉林大学 | Warning method and system for preventing collision for vehicle on high standard highway |
| US20080260239A1 (en) * | 2007-04-17 | 2008-10-23 | Han Chin-Chuan | Object image detection method |
| CN102387345A (en) * | 2011-09-09 | 2012-03-21 | 浙江工业大学 | Safety monitoring system based on omnidirectional vision for old people living alone |
| CN103310202A (en) * | 2013-06-27 | 2013-09-18 | 西安电子科技大学 | System and method for guaranteeing driving safety |
| CN103975342A (en) * | 2012-01-12 | 2014-08-06 | 柯法克斯公司 | Systems and methods for mobile image capture and processing |
| CN105373783A (en) * | 2015-11-17 | 2016-03-02 | 高新兴科技集团股份有限公司 | Seat belt not-wearing detection method based on mixed multi-scale deformable component model |
| CN106156725A (en) * | 2016-06-16 | 2016-11-23 | 江苏大学 | A kind of method of work of the identification early warning system of pedestrian based on vehicle front and cyclist |
| CN106595537A (en) * | 2016-12-30 | 2017-04-26 | 浙大正呈科技有限公司 | Building safety state monitoring device based on BeiDou satellite and monitoring method thereof |
| CN106778590A (en) * | 2016-12-09 | 2017-05-31 | 厦门大学 | It is a kind of that video detecting method is feared based on convolutional neural networks model cruelly |
| CN106845387A (en) * | 2017-01-18 | 2017-06-13 | 合肥师范学院 | Pedestrian detection method based on self study |
| CN107609483A (en) * | 2017-08-15 | 2018-01-19 | 中国科学院自动化研究所 | Risk object detection method, device towards drive assist system |
| CN107862287A (en) * | 2017-11-08 | 2018-03-30 | 吉林大学 | A kind of front zonule object identification and vehicle early warning method |
-
2018
- 2018-04-28 CN CN201810404017.1A patent/CN108629309A/en active Pending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101016053A (en) * | 2007-01-25 | 2007-08-15 | 吉林大学 | Warning method and system for preventing collision for vehicle on high standard highway |
| US20080260239A1 (en) * | 2007-04-17 | 2008-10-23 | Han Chin-Chuan | Object image detection method |
| CN102387345A (en) * | 2011-09-09 | 2012-03-21 | 浙江工业大学 | Safety monitoring system based on omnidirectional vision for old people living alone |
| CN103975342A (en) * | 2012-01-12 | 2014-08-06 | 柯法克斯公司 | Systems and methods for mobile image capture and processing |
| CN103310202A (en) * | 2013-06-27 | 2013-09-18 | 西安电子科技大学 | System and method for guaranteeing driving safety |
| CN105373783A (en) * | 2015-11-17 | 2016-03-02 | 高新兴科技集团股份有限公司 | Seat belt not-wearing detection method based on mixed multi-scale deformable component model |
| CN106156725A (en) * | 2016-06-16 | 2016-11-23 | 江苏大学 | A kind of method of work of the identification early warning system of pedestrian based on vehicle front and cyclist |
| CN106778590A (en) * | 2016-12-09 | 2017-05-31 | 厦门大学 | It is a kind of that video detecting method is feared based on convolutional neural networks model cruelly |
| CN106595537A (en) * | 2016-12-30 | 2017-04-26 | 浙大正呈科技有限公司 | Building safety state monitoring device based on BeiDou satellite and monitoring method thereof |
| CN106845387A (en) * | 2017-01-18 | 2017-06-13 | 合肥师范学院 | Pedestrian detection method based on self study |
| CN107609483A (en) * | 2017-08-15 | 2018-01-19 | 中国科学院自动化研究所 | Risk object detection method, device towards drive assist system |
| CN107862287A (en) * | 2017-11-08 | 2018-03-30 | 吉林大学 | A kind of front zonule object identification and vehicle early warning method |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109657716A (en) * | 2018-12-12 | 2019-04-19 | 天津卡达克数据有限公司 | A kind of vehicle appearance damnification recognition method based on deep learning |
| CN113256926A (en) * | 2021-05-11 | 2021-08-13 | 仲永东 | Active fence system based on construction safety protection |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106372662A (en) | Helmet wearing detection method and device, camera, and server | |
| CN106778737B (en) | A kind of license plate antidote, device and a kind of video acquisition device | |
| CN108319926A (en) | A kind of the safety cap wearing detecting system and detection method of building-site | |
| CN109711370A (en) | A kind of data anastomosing algorithm based on WIFI detection and face cluster | |
| CN107845090A (en) | A kind of silicon chip detection method and silicon chip detection means | |
| CN108932510A (en) | A kind of rubbish detection method and device | |
| CN106412573A (en) | Method and device for detecting lens stain | |
| CN104240264B (en) | The height detection method and device of a kind of moving object | |
| CN109815787B (en) | Target identification method and device, storage medium and electronic equipment | |
| CN106355188A (en) | Image detection method and device | |
| CN107948166A (en) | Traffic anomaly detection method and device based on deep learning | |
| CN110390229B (en) | Face picture screening method and device, electronic equipment and storage medium | |
| CN103973697B (en) | A kind of thing network sensing layer intrusion detection method | |
| CN109740676A (en) | Object detection moving method based on similar purpose | |
| CN108009477A (en) | Stream of people's quantity detection method, device, storage medium and the electronic equipment of image | |
| CN105548216A (en) | Visual detection method for appearance of semi-finished battery | |
| CN105241811B (en) | Multi-level focus adopts drawing method and system automatically | |
| CN108629309A (en) | Foundation pit surrounding people's method for protecting | |
| CN109697441A (en) | A kind of object detection method, device and computer equipment | |
| CN108960145A (en) | Facial image detection method, device, storage medium and electronic equipment | |
| WO2019201065A1 (en) | Geographic iconography-based high-pedestrian volume area determination method and apparatus | |
| CN104121850A (en) | Canopy density measurement method and device | |
| CN110321944A (en) | A kind of construction method of the deep neural network model based on contact net image quality evaluation | |
| CN109657580A (en) | A kind of urban track traffic gate passing control method | |
| CN110135420A (en) | Dial plate state identification method and device, readable storage medium storing program for executing and electronic equipment |
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 |
Application publication date: 20181009 |
|
| RJ01 | Rejection of invention patent application after publication |