CN109409315A - A kind of ATM machine panel zone remnant object detection method and system - Google Patents
A kind of ATM machine panel zone remnant object detection method and system Download PDFInfo
- Publication number
- CN109409315A CN109409315A CN201811319006.XA CN201811319006A CN109409315A CN 109409315 A CN109409315 A CN 109409315A CN 201811319006 A CN201811319006 A CN 201811319006A CN 109409315 A CN109409315 A CN 109409315A
- Authority
- CN
- China
- Prior art keywords
- image
- pedestrian
- leaves
- video flowing
- atm machine
- 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
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
- G06V20/42—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items of sport video content
-
- 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
-
- 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
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/254—Analysis of motion involving subtraction of images
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
- Closed-Circuit Television Systems (AREA)
Abstract
The invention discloses a kind of ATM machine panel zone remnant object detection method and systems, the described method includes: step S1, the video flowing for obtaining ATM machine region, is handled video flowing obtained using background modeling method, is detected the entrance of pedestrian in video flowing or is left；Step S2 respectively intercepts a frame image when detecting that pedestrian enters and pedestrian leaves respectively, enters image to pedestrian and pedestrian leaves image and carries out difference processing, extract the grey scale change region of image；Pedestrian is left convolutional neural networks model of the image after a training and is split, extracted and doubtful leave object area by step S3；Step S4, the grey scale change region that step S2 is extracted and the doubtful object area of leaving that step S3 is extracted are subjected to Overlapping Calculation, to determine that panel with the presence or absence of residue, can remain to the residue of accurate detection panel zone in the case where environmental change such as illumination condition interferes through the invention.
Description
Technical field
The present invention relates to image procossings and video brainpower watch and control technical field, more particularly to one kind based on depth convolution mind
ATM machine panel zone remnant object detection method and system through network.
Background technique
Currently, self-help teller machine (Automatic teller machine, ATM) have become people life in it is indispensable
Few a part, each bank are equipped with large number of ATM.However ATM is for people's lives while offer convenience, also band
Some puzzlements, such as articles forgetting etc..Camera can be installed additional above ATM machine in usual bank, in the same of monitoring financial crime
When also carry out the monitoring of residue.Therefore, panel residue test and analyze have become ATM security system important component it
One, its main function is judged by a series of algorithm, is detected after user leaves ATM machine, the panel zone of ATM machine
It whether there is the object of illegal stickup or installation with the presence or absence of the panel zone for leaving article and ATM machine of user.
Currently, existing atm machine side plate residue detection algorithm is realized based on traditional Digital image analysis technique,
It is mainly by passing in and out the intensity contrast of the panel image of front and back, the grey scale change of the image by judging panel zone to pedestrian
Degree, which judges whether there is, leaves object, and this method is simply direct, but due to this method for the variation of image-region and
It is sensitive, so there are the following problems: 1, can be because environmental change or other factors lead to the fault judged, such as indoor outer ring
Border, the intensity of illumination and the variation in direction, the irregular disengaging of client, being likely to can mistake initiation residue alarm；2, when
It leaves object and alarms from when panel zone " disappearance ", can also cause residue.
Therefore, it is really necessary to propose a kind of technological means, to solve present in above-mentioned ATM machine panel residue detection technique
The problem of.
Summary of the invention
In order to overcome the deficiencies of the above existing technologies, purpose of the present invention is to provide a kind of ATM machine panel zone something lost
Object detecting method and system are stayed, to remain to accurate detection panel zone in the case where environmental change such as illumination condition interferes
Residue.
In view of the above and other objects, the present invention proposes a kind of ATM machine panel zone remnant object detection method, including as follows
Step:
Step S1 is obtained the video flowing in ATM machine region, is handled using background modeling method video flowing obtained,
It detects the entrance of pedestrian in video flowing or leaves；
Step S2 respectively intercepts a frame image when detecting that pedestrian enters and pedestrian leaves respectively, enters image to pedestrian
Image is left with pedestrian and carries out difference processing, extracts the grey scale change region of image；
Pedestrian is left convolutional neural networks model of the image after a training and is split, extracted doubtful by step S3
Seemingly leave object area；
Step S4, by grey scale change region that step S2 is extracted and step S3 extract it is doubtful leave object area carry out it is be overlapped
Analytical calculation is spent, to determine panel with the presence or absence of residue.
Preferably, step S1 further comprises:
Step S100 obtains video flowing；
Step S101 is built using Gaussian Background using the video flowing obtained in step S100 and touches method building background model；
Step S102 successively obtains a frame image from video flowing, obtains fortune using the background model of step S101 building
Dynamic foreground image；
Step S103 according to the entrance for obtaining sport foreground image confirmation pedestrian or leaves.
Preferably, in step S103, judgement has pedestrian's entrance if detecting that the prospect of movement appears in pedestrian area,
Judge that pedestrian leaves if sport foreground disappears.
Preferably, step S2 further comprises:
Step S200, acquisition pedestrian enters the gray value of each pixel of image respectively and pedestrian leaves each pixel of image
The gray value of point；
Pedestrian is left the gray value of each pixel of image and pedestrian enters the ashes of image corresponding pixel points by step S201
It is poor that angle value is made, and takes absolute value to difference, obtains difference image.
Preferably, step S3 further comprises:
Step S300, builds convolutional neural networks model, and input is adjusted automatically after being manually labelled with the image for leaving object area
Mould preparation shape parameter, after being trained to model, generation can extract the mathematical model that panel leaves object area.
Pedestrian is left image and inputs the convolutional neural networks model, is split, mention to described image by step S301
It takes out and doubtful leaves object area.
Preferably, in step S4, if the grey scale change region and the doubtful object area degree of overlapping of leaving are greater than default
Threshold value, it is determined that there are residues for panel.
In order to achieve the above objectives, the present invention also provides a kind of ATM machine panel zone residue detection systems, comprising:
Background modeling unit flows into obtained video using background modeling method for obtaining the video flowing in ATM machine region
Row processing detects the entrance of pedestrian or leaves；
Image difference processing unit, for respectively intercepting a frame image when detecting that pedestrian enters and pedestrian leaves respectively,
Image is entered to pedestrian and pedestrian leaves image and carries out difference processing, extracts the grey scale change region of image；
Deep learning unit divides for pedestrian to be left convolutional neural networks model of the image after a training
It cuts, extracts and doubtful leave object area；
Degree of overlapping analytical unit, grey scale change region and deep learning for extracting described image difference processing unit
The doubtful object area of leaving that unit extracts carries out Overlapping Calculation, to determine panel with the presence or absence of residue.
Preferably, the background modeling unit includes:
Video flowing acquiring unit, for obtaining video flowing；
Background model construction unit, the video flowing for being obtained using video flowing acquiring unit build the side of touching using Gaussian Background
Method constructs background model；
Sport foreground acquiring unit is constructed for successively obtaining a frame image from video flowing using the background model
The background model of building unit obtains sport foreground image；
State detection unit, for confirming the entrance of pedestrian according to acquisition sport foreground image or leaving.
Preferably, described image difference processing unit further comprises:
Gray value obtains module, and for obtaining respectively, pedestrian enters the gray value of each pixel of image and pedestrian leaves figure
As the gray value of each pixel；
Differential processing module enters the corresponding picture of image for pedestrian to be left to the gray value of each pixel of image and pedestrian
It is poor that the gray value of vegetarian refreshments is made, and takes absolute value to difference, obtains difference image, that is, obtains corresponding grey scale change region.
Preferably, the deep learning unit further comprises:
Convolutional neural networks construction unit, for building convolutional neural networks model, input is manually labelled with residue area
Self-optimizing model parameter after the image in domain, after being trained to model, generation can extract the mathematics that panel leaves object area
Model；
Image segmentation unit inputs the convolutional neural networks model for pedestrian to be left image, to described image into
Row segmentation, extracts and doubtful leaves object area.
Compared with prior art, a kind of ATM machine panel zone remnant object detection method of the present invention and system pass through to combine and doubt
Like residue extraction and grey scale change regional correlation, accurate detection panel zone can be gone out in the case where illumination condition interferes
Residue.
Detailed description of the invention
Fig. 1 is a kind of step flow chart of ATM machine panel zone remnant object detection method of the present invention；
Fig. 2 is the structural schematic diagram of convolutional neural networks model in the specific embodiment of the invention；
Fig. 3 is a kind of system architecture diagram of ATM machine panel zone residue detection system of the present invention；
Fig. 4 is that the ATM machine panel zone of the specific embodiment of the invention leaves the process schematic of analyte detection.
Specific embodiment
Below by way of specific specific example and embodiments of the present invention are described with reference to the drawings, those skilled in the art can
Understand further advantage and effect of the invention easily by content disclosed in the present specification.The present invention can also pass through other differences
Specific example implemented or applied, details in this specification can also be based on different perspectives and applications, without departing substantially from
Various modifications and change are carried out under spirit of the invention.
Fig. 1 is a kind of step flow chart of ATM machine panel zone remnant object detection method of the present invention.As shown in Figure 1, this hair
A kind of bright ATM machine panel zone remnant object detection method, includes the following steps:
Step S1 is obtained the video flowing in ATM machine region, is handled using background modeling method obtained image, is detected
The entrance of pedestrian is left.
In the specific embodiment of the invention, in step S1, transported using image of the Gaussian Background modeling method to acquisition
The detection of moving-target, judgement has pedestrian's entrance if detecting that the prospect of movement appears in pedestrian area (settable), if prospect
Disappearance then judges that pedestrian leaves.In the specific embodiment of the invention, view is obtained by being set to the camera at the top of ATM machine region
Frequency flows, and specifically, step S1 further comprises:
Step S100 obtains video flowing；
Step S101 is built using Gaussian Background using the video flowing obtained in step S100 and touches method building background model；
Step S102 successively obtains a frame image from video flowing, before obtaining movement using the background model of S101 building
Scape image；
Step S103 according to the entrance for obtaining sport foreground image confirmation pedestrian or leaves.If detecting the prospect of movement
It appears in pedestrian area (settable) and then judges there is pedestrian's entrance, judge that pedestrian leaves if sport foreground disappears.
Step S2 respectively intercepts a frame image when detecting that pedestrian enters and pedestrian leaves respectively, enters image to pedestrian
Image is left with pedestrian and carries out difference processing, extracts the grey scale change region of image.That is, by pedestrian enter image with
The two field pictures that pedestrian leaves image are checked the mark, and image of checking the mark, i.e. grey scale change region are obtained.Specifically, step S2 is into one
Step includes:
Step S200, acquisition pedestrian enters the gray value of each pixel of image respectively and pedestrian leaves each pixel of image
The gray value of point；
Pedestrian is left the gray value of each pixel of image and pedestrian enters the ashes of image corresponding pixel points by step S201
It is poor that angle value is made, and takes absolute value to difference, obtains difference image, that is, obtains corresponding grey scale change region.
Pedestrian is left convolutional neural networks model of the image after a training and is split, extracted doubtful by step S3
Seemingly leave object area.
Specifically, step S3 further comprises:
Step S300, builds convolutional neural networks model, and input is adjusted automatically after being manually labelled with the image for leaving object area
Mould preparation shape parameter, after being trained to model, generation can extract the mathematical model that panel leaves object area.Fig. 2 is the present invention
The structural schematic diagram for the convolutional neural networks model built in specific embodiment.Original image is input to 14 layers of MobileNet
Afterwards output 1 × 1 × 512 characteristic pattern, after up-sampling (Upsampling) is restored to 1/4 size of original image, by its with
The 4x characteristic pattern and 8x characteristic pattern of MobileNet output in channel direction to being spliced, eventually by subsequent convolutional layer and
It up-samples after layer calculates and exports original image segmentation result.Since the training to convolutional neural networks is using the prior art,
It will not go into details for this.
Pedestrian is left image and inputs the convolutional neural networks model, is split, mention to described image by step S301
It takes out and doubtful leaves object area.
Step S4, by grey scale change region that step S2 is extracted and step S3 extract it is doubtful leave object area carry out it is be overlapped
Degree calculates, to determine panel with the presence or absence of residue, if the two degree of overlapping is larger (when being greater than preset threshold), it is determined that face
There are residues for plate, and push alarm.
Fig. 3 is a kind of system architecture diagram of ATM machine panel zone residue detection system of the present invention.As shown in figure 3, this hair
A kind of bright ATM machine panel zone remnant object detection method, includes the following steps:
Background modeling unit 201, for obtaining the video flowing in ATM machine region, using background modeling method to view obtained
Frequency stream is handled, and is detected the entrance of pedestrian or is left.
In the specific embodiment of the invention, background modeling unit 201 is using Gaussian Background modeling method to the video of acquisition
The image of stream carries out the detection of moving target, and judgement has row if detecting that the prospect of movement appears in pedestrian area (settable)
People enters, and judges that pedestrian leaves if prospect disappears.Specifically, pedestrian's state detection unit 201 further comprises:
Video flowing acquiring unit, for obtaining video flowing；
Background model construction unit, the video flowing for being obtained using video flowing acquiring unit build the side of touching using Gaussian Background
Method constructs background model；
Sport foreground acquiring unit is constructed for successively obtaining a frame image from video flowing using the background model
The background model of building unit obtains sport foreground image；
State detection unit, for confirming the entrance of pedestrian according to acquisition sport foreground image or leaving.If detecting fortune
Dynamic prospect appears in pedestrian area (settable) and then judges there is pedestrian's entrance, judges that pedestrian leaves if sport foreground disappears.
Image difference processing unit 202, for respectively intercepting a frame image when pedestrian enters and pedestrian leaves respectively, to row
People enters image and pedestrian leaves image and carries out difference processing, extracts the grey scale change region of image.That is, by pedestrian
The two field pictures for leaving image with pedestrian into image are checked the mark, and image of checking the mark, i.e. grey scale change region are obtained.Specifically,
Image difference processing unit 202 further comprises:
Gray value obtains module, and for obtaining respectively, pedestrian enters the gray value of each pixel of image and pedestrian leaves figure
As the gray value of each pixel；
Differential processing module enters the corresponding picture of image for pedestrian to be left to the gray value of each pixel of image and pedestrian
It is poor that the gray value of vegetarian refreshments is made, and takes absolute value to difference, obtains difference image, that is, obtains corresponding grey scale change region.
Deep learning unit 203 is carried out for pedestrian to be left convolutional neural networks model of the image after a training
Segmentation extracts and doubtful leaves object area.
Specifically, deep learning unit 203 further comprises:
Convolutional neural networks construction unit, for building convolutional neural networks model, input is manually labelled with residue area
Self-optimizing model parameter after the image in domain, after being trained to model, generation can extract the mathematics that panel leaves object area
Model.In the specific embodiment of the invention, convolutional neural networks model is as shown in Fig. 2, it will not be described here.
Image segmentation unit inputs the convolutional neural networks model for pedestrian to be left image, to described image into
Row segmentation, extracts and doubtful leaves object area.
Degree of overlapping analytical unit 204, grey scale change region and depth for extracting image difference processing unit 202
It practises the doubtful object area of leaving that unit 203 extracts and carries out Overlapping Calculation, to determine panel with the presence or absence of residue, if the two
Degree of overlapping is larger, it is determined that there are residues for panel, and push alarm.
Fig. 4 is that the ATM machine panel zone of the specific embodiment of the invention leaves the process schematic of analyte detection.
1, deep neural network model is constructed first, and the ATM machine for collecting live tape label (artificial mark) leaves object image
Sample is trained the deep neural network model, and the deep neural network model is allow to tell the face in image
Plate region, light area and leaves object area；
2, after the image for obtaining ATM panel camera, entering and leaving for pedestrian is detected by background modeling method, and divide
It does not intercept an image when pedestrian enters and when pedestrian leaves respectively, is sent into image difference processing unit and is handled, is i.e. in Fig. 3
Solid line process, image (pedestrian enter with pedestrian leaves) opened to front and back two carry out difference and extract variation of image grayscale region,
It is considered being likely to be leaving object area；Image after pedestrian is left simultaneously passes through the deep neural network model after training
It is split, i.e. dotted line process in Fig. 3, extracts and doubtful leave object area.
3, by leaving object area and the progress degree of overlapping analysis of grey scale change region to panel is doubtful, if the two degree of overlapping
It is larger, it is determined that there are residues for panel, and push alarm.
In conclusion a kind of ATM machine panel zone remnant object detection method of the present invention and system, which pass through, combines doubtful leave
Object extracts and grey scale change regional correlation, accurate detection can go out leaving for panel zone in the case where illumination condition interferes
Object.
The above-described embodiments merely illustrate the principles and effects of the present invention, and is not intended to limit the present invention.Any
Without departing from the spirit and scope of the present invention, modifications and changes are made to the above embodiments by field technical staff.Therefore,
The scope of the present invention, should be as listed in the claims.
Claims (10)
1. a kind of ATM machine panel zone remnant object detection method, includes the following steps:
Step S1 is obtained the video flowing in ATM machine region, is handled using background modeling method video flowing obtained, is detected
In video flowing pedestrian entrance or leave；
Step S2 respectively intercepts a frame image when detecting that pedestrian enters and pedestrian leaves respectively, enters image and row to pedestrian
People leaves image and carries out difference processing, extracts the grey scale change region of image；
Pedestrian is left convolutional neural networks model of the image after a training and is split, extracts doubtful something lost by step S3
Stay object area；
The grey scale change region that step S2 is extracted and the doubtful object area of leaving that step S3 is extracted are carried out degree of overlapping meter by step S4
It calculates, to determine panel with the presence or absence of residue.
2. a kind of ATM machine panel zone remnant object detection method as described in claim 1, which is characterized in that step S1 is into one
Step includes:
Step S100 obtains video flowing；
Step S101 is built using Gaussian Background using the video flowing obtained in step S100 and touches method building background model；
Step S102 successively obtains a frame image from video flowing, before obtaining movement using the background model of step S101 building
Scape image；
Step S103 according to the entrance for obtaining sport foreground image confirmation pedestrian or leaves.
3. a kind of ATM machine panel zone remnant object detection method as claimed in claim 2, it is characterised in that: in step S103
In, judgement has pedestrian's entrance if detecting that the prospect of movement appears in pedestrian area, if sport foreground disappearance judges pedestrian
It leaves.
4. a kind of ATM machine panel zone remnant object detection method as described in claim 1, which is characterized in that step S2 is into one
Step includes:
Step S200, acquisition pedestrian enters the gray value of each pixel of image respectively and pedestrian leaves each pixel of image
Gray value；
Pedestrian is left the gray value of each pixel of image and pedestrian enters the gray values of image corresponding pixel points by step S201
It is poor to make, and takes absolute value to difference, obtains difference image.
5. a kind of ATM machine panel zone remnant object detection method as described in claim 1, which is characterized in that step S3 is into one
Step includes:
Step S300, builds convolutional neural networks model, and input is manually labelled with adjust automatically mould after the image for leaving object area
Shape parameter, after being trained to model, generation can extract the mathematical model that panel leaves object area.
Pedestrian is left image and inputs the convolutional neural networks model, is split, extract to described image by step S301
It is doubtful to leave object area.
6. a kind of ATM machine panel zone remnant object detection method as described in claim 1, it is characterised in that: in step S4, if
The grey scale change region and the doubtful object area degree of overlapping of leaving are greater than preset threshold, it is determined that there are residues for panel.
7. a kind of ATM machine panel zone residue detection system, comprising:
Background modeling unit, for obtaining the video flowing in ATM machine region, using background modeling method to obtained video flowing at
Reason detects the entrance of pedestrian or leaves；
Image difference processing unit, for respectively intercepting a frame image when detecting that pedestrian enters and pedestrian leaves respectively, to row
People enters image and pedestrian leaves image and carries out difference processing, extracts the grey scale change region of image；
Deep learning unit is split for pedestrian to be left convolutional neural networks model of the image after a training, mentions
It takes out and doubtful leaves object area；
Degree of overlapping analytical unit, grey scale change region and deep learning unit for extracting described image difference processing unit
The doubtful object area of leaving extracted carries out Overlapping Calculation, to determine panel with the presence or absence of residue.
8. a kind of ATM machine panel zone residue detection system as claimed in claim 7, which is characterized in that the background is built
Form unit includes:
Video flowing acquiring unit, for obtaining video flowing；
Background model construction unit, the video flowing for being obtained using video flowing acquiring unit is built using Gaussian Background touches method structure
Build background model；
Sport foreground acquiring unit utilizes the background model construction unit for successively obtaining a frame image from video flowing
The background model of building obtains sport foreground image；
State detection unit, for confirming the entrance of pedestrian according to acquisition sport foreground image or leaving.
9. a kind of ATM machine panel zone residue detection system as claimed in claim 7, which is characterized in that described image is poor
Point processing unit further comprises:
Gray value obtains module, enters the gray value of each pixel of image for obtaining pedestrian respectively with pedestrian to leave image each
The gray value of a pixel；
Differential processing module, gray value and pedestrian for pedestrian to be left to each pixel of image enter image corresponding pixel points
Gray value make poor, and take absolute value to difference, obtain difference image, that is, obtain corresponding grey scale change region.
10. a kind of ATM machine panel zone residue detection system as claimed in claim 7, which is characterized in that the depth
Practising unit further comprises:
Convolutional neural networks construction unit, for building convolutional neural networks model, input is manually labelled with and leaves object area
Self-optimizing model parameter after image, after being trained to model, generation can extract the mathematical model that panel leaves object area；
Image segmentation unit inputs the convolutional neural networks model for pedestrian to be left image, divides described image
It cuts, extracts and doubtful leave object area.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811319006.XA CN109409315B (en) | 2018-11-07 | 2018-11-07 | Method and system for detecting remnants in panel area of ATM (automatic Teller machine) |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811319006.XA CN109409315B (en) | 2018-11-07 | 2018-11-07 | Method and system for detecting remnants in panel area of ATM (automatic Teller machine) |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109409315A true CN109409315A (en) | 2019-03-01 |
| CN109409315B CN109409315B (en) | 2022-01-11 |
Family
ID=65472088
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811319006.XA Active CN109409315B (en) | 2018-11-07 | 2018-11-07 | Method and system for detecting remnants in panel area of ATM (automatic Teller machine) |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109409315B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110399792A (en) * | 2019-06-19 | 2019-11-01 | 河南北斗卫星导航平台有限公司 | A kind of monitoring method, device and computer readable storage medium |
| CN110956102A (en) * | 2019-11-19 | 2020-04-03 | 上海眼控科技股份有限公司 | Bank counter monitoring method and device, computer equipment and storage medium |
| CN111091097A (en) * | 2019-12-20 | 2020-05-01 | ***通信集团江苏有限公司 | Method, device, equipment and storage medium for identifying remnants |
| CN111259955A (en) * | 2020-01-15 | 2020-06-09 | 国家测绘产品质量检验测试中心 | Method and system for reliable property detection of geographical national condition monitoring result |
| CN111932525A (en) * | 2020-08-19 | 2020-11-13 | 中国银行股份有限公司 | Method and device for detecting left physical objects at physical delivery port of bank equipment |
| CN112837326A (en) * | 2021-01-27 | 2021-05-25 | 南京中兴力维软件有限公司 | Remnant detection method, device and equipment |
| CN117036482A (en) * | 2023-08-22 | 2023-11-10 | 北京智芯微电子科技有限公司 | Target object positioning method, device, shooting equipment, chip, equipment and medium |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101426128A (en) * | 2007-10-30 | 2009-05-06 | 三星电子株式会社 | Detection system and method for stolen and lost packet |
| CN102034240A (en) * | 2010-12-23 | 2011-04-27 | 北京邮电大学 | Method for detecting and tracking static foreground |
| CN102413321A (en) * | 2011-12-26 | 2012-04-11 | 浙江省电力公司 | Automatic image-recording system and method |
| CN103679123A (en) * | 2012-09-17 | 2014-03-26 | 浙江大华技术股份有限公司 | Method and system for detecting remnant on ATM (Automatic Teller Machine) panel |
| CN104156942A (en) * | 2014-07-02 | 2014-11-19 | 华南理工大学 | Detection method for remnants in complex environment |
| CN105678803A (en) * | 2015-12-29 | 2016-06-15 | 南京理工大学 | Video monitoring target detection method based on W4 algorithm and frame difference |
| CN106296677A (en) * | 2016-08-03 | 2017-01-04 | 浙江理工大学 | A kind of remnant object detection method of double mask context updates based on double-background model |
| CN106372576A (en) * | 2016-08-23 | 2017-02-01 | 南京邮电大学 | Deep learning-based intelligent indoor intrusion detection method and system |
| CN106469311A (en) * | 2015-08-19 | 2017-03-01 | 南京新索奇科技有限公司 | Object detection method and device |
-
2018
- 2018-11-07 CN CN201811319006.XA patent/CN109409315B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101426128A (en) * | 2007-10-30 | 2009-05-06 | 三星电子株式会社 | Detection system and method for stolen and lost packet |
| CN102034240A (en) * | 2010-12-23 | 2011-04-27 | 北京邮电大学 | Method for detecting and tracking static foreground |
| CN102413321A (en) * | 2011-12-26 | 2012-04-11 | 浙江省电力公司 | Automatic image-recording system and method |
| CN103679123A (en) * | 2012-09-17 | 2014-03-26 | 浙江大华技术股份有限公司 | Method and system for detecting remnant on ATM (Automatic Teller Machine) panel |
| CN104156942A (en) * | 2014-07-02 | 2014-11-19 | 华南理工大学 | Detection method for remnants in complex environment |
| CN106469311A (en) * | 2015-08-19 | 2017-03-01 | 南京新索奇科技有限公司 | Object detection method and device |
| CN105678803A (en) * | 2015-12-29 | 2016-06-15 | 南京理工大学 | Video monitoring target detection method based on W4 algorithm and frame difference |
| CN106296677A (en) * | 2016-08-03 | 2017-01-04 | 浙江理工大学 | A kind of remnant object detection method of double mask context updates based on double-background model |
| CN106372576A (en) * | 2016-08-23 | 2017-02-01 | 南京邮电大学 | Deep learning-based intelligent indoor intrusion detection method and system |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110399792A (en) * | 2019-06-19 | 2019-11-01 | 河南北斗卫星导航平台有限公司 | A kind of monitoring method, device and computer readable storage medium |
| CN110956102A (en) * | 2019-11-19 | 2020-04-03 | 上海眼控科技股份有限公司 | Bank counter monitoring method and device, computer equipment and storage medium |
| CN111091097A (en) * | 2019-12-20 | 2020-05-01 | ***通信集团江苏有限公司 | Method, device, equipment and storage medium for identifying remnants |
| CN111259955A (en) * | 2020-01-15 | 2020-06-09 | 国家测绘产品质量检验测试中心 | Method and system for reliable property detection of geographical national condition monitoring result |
| CN111259955B (en) * | 2020-01-15 | 2023-12-08 | 国家测绘产品质量检验测试中心 | Reliable quality inspection method and system for geographical national condition monitoring result |
| CN111932525A (en) * | 2020-08-19 | 2020-11-13 | 中国银行股份有限公司 | Method and device for detecting left physical objects at physical delivery port of bank equipment |
| CN111932525B (en) * | 2020-08-19 | 2024-02-23 | 中国银行股份有限公司 | Method and device for detecting left-over real object of real object delivery port of bank equipment |
| CN112837326A (en) * | 2021-01-27 | 2021-05-25 | 南京中兴力维软件有限公司 | Remnant detection method, device and equipment |
| CN112837326B (en) * | 2021-01-27 | 2024-04-09 | 南京中兴力维软件有限公司 | Method, device and equipment for detecting carryover |
| CN117036482A (en) * | 2023-08-22 | 2023-11-10 | 北京智芯微电子科技有限公司 | Target object positioning method, device, shooting equipment, chip, equipment and medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109409315B (en) | 2022-01-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109409315A (en) | A kind of ATM machine panel zone remnant object detection method and system | |
| CN110689054B (en) | Worker violation monitoring method | |
| CN105788142A (en) | Video image processing-based fire detection system and detection method | |
| CN106897720A (en) | A kind of firework detecting method and device based on video analysis | |
| WO2015085811A1 (en) | Method and device for banknote identification based on thickness signal identification | |
| CN105930822A (en) | Human face snapshot method and system | |
| CN108804987B (en) | Door opening and closing state detection method and device and people flow detection system | |
| CN111126293A (en) | Flame and smoke abnormal condition detection method and system | |
| CN109035226B (en) | Mura defect detection method based on LSTM model | |
| CN110189355A (en) | Safe escape channel occupies detection method, device, electronic equipment and storage medium | |
| CN110084987A (en) | A kind of foreign matter inspecting system and method towards rail traffic | |
| CN112184773A (en) | Helmet wearing detection method and system based on deep learning | |
| CN105976398A (en) | Daylight fire disaster video detection method | |
| US20220230519A1 (en) | Method and device for detecting smoke | |
| CN110298410A (en) | Weak target detection method and device in soft image based on deep learning | |
| CN109274945B (en) | Method and system for self-adaptively performing true color restoration on image | |
| CN105447863A (en) | Residue detection method based on improved VIBE | |
| CN114772208A (en) | Non-contact belt tearing detection system and method based on image segmentation | |
| Pan et al. | Key algorithms of video target detection and recognition in intelligent transportation systems | |
| CN108563997A (en) | It is a kind of establish Face datection model, recognition of face method and apparatus | |
| CN113052055A (en) | Smoke detection method based on optical flow improvement and Yolov3 | |
| CN116206155A (en) | Waste steel classification and identification method based on YOLOv5 network | |
| CN110349178A (en) | A kind of human body unusual checking and identifying system and method | |
| CN110858392A (en) | Monitoring target positioning method based on fusion background model | |
| CN109472223A (en) | A kind of face identification method and device |
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 |