CN110580455A - image recognition-based illegal off-duty detection method and device for personnel - Google Patents
image recognition-based illegal off-duty detection method and device for personnel Download PDFInfo
- Publication number
- CN110580455A CN110580455A CN201910776281.2A CN201910776281A CN110580455A CN 110580455 A CN110580455 A CN 110580455A CN 201910776281 A CN201910776281 A CN 201910776281A CN 110580455 A CN110580455 A CN 110580455A
- Authority
- CN
- China
- Prior art keywords
- image
- duty
- features
- personnel
- people
- 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
- 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
-
- 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
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
- Burglar Alarm Systems (AREA)
Abstract
the invention discloses a method and a device for detecting illegal personnel off duty based on image recognition, wherein the method comprises the following steps: performing image extraction on the monitoring video according to a preset time interval; carrying out feature extraction on the extracted image by using the trained convolutional neural network to obtain an image feature result; performing target scanning on the extracted image features by using a regional suggestion network in a sliding window mode to obtain regional suggestion image features; framing image features according to the suggested region to perform ROI processing, and calculating the number of on-duty personnel according to the calculated foreground scores of all target frames in the image; and when the number of the people on duty is less than the preset target number of the people on duty, judging that the people are illegally off duty. According to the invention, by combining the regional neural network and the convolutional neural network, whether the monitored personnel are illegally off duty can be accurately and effectively judged, so that security holes caused by illegal off duty of the personnel can be effectively avoided.
Description
Technical Field
the invention relates to the technical field of image recognition, in particular to a person illegal off-duty detection method and device based on image recognition.
Background
In some scenes, such as video monitoring rooms and other places, the problem that people leave the post without authorization and accidents in a video monitoring area are not found in time, so that the problem cannot be timely and effectively handled, and a large loss or safety problem is caused.
in contrast, the conventional method adopts a mode of manually watching images for monitoring, but the method for manually monitoring the video has low efficiency, consumes manpower and is difficult to ensure accuracy. In the field of intelligent monitoring and identification, the prior art judges that the mode that people are not on duty is greatly influenced by the environment through the area change of an image area, and the accuracy is not high. Therefore, a technology capable of accurately judging whether the person is off duty or not in monitoring is urgently needed.
Disclosure of Invention
the technical problem to be solved by the embodiments of the present invention is to provide a method and a device for detecting illegal off-duty of a person based on image recognition, which can accurately judge whether the person under monitoring is off-duty.
in order to solve the technical problem, the invention provides a person illegal off-duty detection method based on image recognition, which comprises the following steps:
Performing image extraction on the monitoring video according to a preset time interval;
carrying out feature extraction on the extracted image by using the trained convolutional neural network to obtain an image feature result;
performing target scanning on the extracted image features by using a regional suggestion network in a sliding window mode to obtain regional suggestion image features;
framing image features according to the suggested region to perform ROI processing, and calculating the number of on-duty personnel according to the calculated foreground scores of all target frames in the image;
And when the number of the people on duty is less than the preset target number of the people on duty, judging that the people are illegally off duty.
further, the feature extraction is performed on the extracted image by using the trained convolutional neural network to obtain an image feature result, which specifically comprises:
after the extracted image is preprocessed, performing convolution on the preprocessed image by using a preset convolution kernel to obtain a plurality of characteristic results;
After the plurality of characteristic results are subjected to normalization processing and Relu activation, extracting N layers of pooling layers through a resnet101 network model to form a characteristic pyramid;
And after the feature extraction is carried out on the N layers of features of the feature pyramid by using the same convolutional network, carrying out shape classification through the first convolutional network to obtain a background/foreground classification result, and simultaneously carrying out shape regression through the second convolutional network to obtain an offset regression result.
further, the target scanning is performed on the extracted image features by using the regional suggestion network in a sliding window manner to obtain regional suggestion image features, and the method specifically comprises the following steps:
performing target scanning on the extracted image features by using a regional suggestion network in a sliding window mode, extracting i foreground features of which the scores are higher than a preset threshold value from the image features, and obtaining an initial regional suggestion image;
NMS processing is carried out on the shifted suggested areas, j target frames in the suggested images of the initial areas are screened out, and the suggested images of the areas are obtained; wherein j < i.
further, the image features are framed according to the suggested region to perform ROI processing, and the number of on-duty people is calculated according to the calculated foreground score of each target frame in the image, which specifically includes:
combining the region suggested image and the N layers of features of the feature pyramid, and performing region feature aggregation by using a ROIAlign layer to obtain a final classification result;
and calculating the foreground score of each target frame in the image according to the final classification result, and acquiring the number of the final target frames according to a preset classification credibility threshold value to obtain the number of the on-duty personnel.
further, before the image extraction is performed on the monitoring video at the preset time interval, the method further includes:
And setting the preset target post arrival number according to a setting instruction input by a user.
further, the method further comprises:
and when judging that the personnel are illegally off duty, sending an alarm instruction to control the alarm device to remind the manager.
In order to solve the same technical problem, the invention also provides a device for detecting illegal people leaving behind based on image recognition, which comprises an image extraction module, an image feature extraction module, an area suggestion image generation module, an on-duty personnel number calculation module and a personnel leaving behind judgment module; wherein the content of the first and second substances,
the image extraction module is used for extracting images of the monitoring video according to a preset time interval;
the image feature extraction module is used for extracting features of the extracted image by using the trained convolutional neural network to obtain an image feature result;
the region suggestion image generation module is used for performing target scanning on the extracted image features in a sliding window mode by using a region suggestion network to obtain region suggestion image features;
the on-duty personnel number calculating module is used for framing the image characteristics according to the suggested region to perform ROI processing, and calculating the on-duty personnel number according to the calculated foreground scores of all the target frames in the image;
And the personnel off-duty judging module is used for judging that the personnel illegally leave the duty when the number of the personnel on duty is less than the preset target number of the people arriving on duty.
further, the image feature extraction module is specifically configured to:
After the extracted image is preprocessed, performing convolution on the preprocessed image by using a preset convolution kernel to obtain a plurality of characteristic results;
After the plurality of characteristic results are subjected to normalization processing and Relu activation, extracting N layers of pooling layers through a resnet101 network model to form a characteristic pyramid;
and after the feature extraction is carried out on the N layers of features of the feature pyramid by using the same convolutional network, carrying out shape classification through the first convolutional network to obtain a background/foreground classification result, and simultaneously carrying out shape regression through the second convolutional network to obtain an offset regression result.
further, the image feature extraction module is specifically configured to:
performing target scanning on the extracted image features by using a regional suggestion network in a sliding window mode, extracting i foreground features of which the scores are higher than a preset threshold value from the image features, and obtaining an initial regional suggestion image;
NMS processing is carried out on the shifted suggested areas, j target frames in the suggested images of the initial areas are screened out, and the suggested images of the areas are obtained; wherein j < i.
further, the image feature extraction module is specifically configured to:
Combining the region suggested image and the N layers of features of the feature pyramid, and performing region feature aggregation by using a ROIAlign layer to obtain a final classification result;
and calculating the foreground score of each target frame in the image according to the final classification result, and acquiring the number of the final target frames according to a preset classification credibility threshold value to obtain the number of the on-duty personnel.
Compared with the prior art, the invention has the following beneficial effects:
The invention discloses a method and a device for detecting illegal personnel off duty based on image recognition, wherein the method comprises the following steps: performing image extraction on the monitoring video according to a preset time interval; carrying out feature extraction on the extracted image by using the trained convolutional neural network to obtain an image feature result; performing target scanning on the extracted image features by using a regional suggestion network in a sliding window mode to obtain regional suggestion image features; framing image features according to the suggested region to perform ROI processing, and calculating the number of on-duty personnel according to the calculated foreground scores of all target frames in the image; and when the number of the people on duty is less than the preset target number of the people on duty, judging that the people are illegally off duty. According to the invention, by combining the regional neural network and the convolutional neural network, whether the monitored personnel are illegally off duty can be accurately and effectively judged, so that security holes caused by illegal off duty of the personnel can be effectively avoided.
drawings
fig. 1 is a schematic flow chart of a method for detecting illegal people leaving post based on image recognition according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of a device for detecting illegal people leaving post based on image recognition according to an embodiment of the present invention.
Detailed Description
the technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the accompanying 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, an embodiment of the present invention provides a method for detecting illegal people leaving post based on image recognition, including:
step S1, extracting images of the monitoring video according to a preset time interval;
Step S2, extracting the features of the extracted image by using the trained convolutional neural network to obtain an image feature result;
Step S3, performing target scanning on the extracted image features by using a regional suggestion network in a sliding window mode to obtain regional suggestion image features;
Step S4, framing image features according to the suggested region to perform ROI processing, and calculating the number of on-duty personnel according to the calculated foreground scores of all target frames in the image;
and step S5, when the number of the people on duty is less than the preset target number of the people on duty, judging that the people are illegally off duty.
in the embodiment of the present invention, step S2 specifically includes:
after the extracted image is preprocessed, performing convolution on the preprocessed image by using a preset convolution kernel to obtain a plurality of characteristic results;
After the plurality of characteristic results are subjected to normalization processing and Relu activation, extracting N layers of pooling layers through a resnet101 network model to form a characteristic pyramid;
and after the feature extraction is carried out on the N layers of features of the feature pyramid by using the same convolutional network, carrying out shape classification through the first convolutional network to obtain a background/foreground classification result, and simultaneously carrying out shape regression through the second convolutional network to obtain an offset regression result.
in the embodiment of the present invention, step S3 specifically includes:
Performing target scanning on the extracted image features by using a regional suggestion network in a sliding window mode, extracting i foreground features of which the scores are higher than a preset threshold value from the image features, and obtaining an initial regional suggestion image;
NMS processing is carried out on the shifted suggested areas, j target frames in the suggested images of the initial areas are screened out, and the suggested images of the areas are obtained; wherein j < i.
in the embodiment of the present invention, further, step S4 specifically includes:
Combining the region suggested image and the N layers of features of the feature pyramid, and performing region feature aggregation by using a ROIAlign layer to obtain a final classification result;
And calculating the foreground score of each target frame in the image according to the final classification result, and acquiring the number of the final target frames according to a preset classification credibility threshold value to obtain the number of the on-duty personnel.
In the embodiment of the invention, a classification neural network is established, optionally, a 3-layer convolution + 2-layer full-connection neural network is adopted to train a human picture and extract features. Then, using the region suggestion network (RPN), the image is scanned in a sliding window manner to find a region where the target exists. Using the prediction of the RPN, the anchor that best contains the target can be selected and its position and size fine-tuned. In the case where there are multiple anchors overlapping each other, the anchor with the highest foreground score will be retained and the remainder discarded. The final regional proposal is then obtained and passed on to the next stage. The anchors are then analyzed by the ROI to identify if the target has human features, since the classifier does not handle many input sizes well, usually only fixed input sizes. However, due to the bounding box refinement step in the RPN, the ROI box can have different sizes, and therefore pooling is used to solve this problem. And finally, returning the number of the target areas with the scores higher than the rated value, namely acquiring the number of the people. The invention judges the number of people in the area from the characteristics of people and judges the number of people from various limbs and postures of people, thereby effectively reducing misjudgment.
the process flow of the present invention is illustrated below with reference to specific parameters.
Firstly, extracting frames in a monitoring video, setting timing extraction every X seconds, preprocessing an extracted image, for example, adjusting the size of a picture frame to an equal-ratio picture with a width of 1280 pixels, then filling 3X 3 around the picture, then performing convolution through a convolution kernel with a size of 7X 7 and a step length of 2 (when the convolution kernel is image processing, an input image is given, each pixel in an output image is a weighted average of pixels in a small area in the input image, wherein a weight value is defined by a function, and the function is called convolution kernel), acquiring 64 feature results with a size of 64X 64, then performing normalization processing on the 64 feature results, activating by Relu to prevent overfitting, then transmitting the feature results to a resnet101 network, extracting 5 layers of pooling layers from the feature results to form a feature pyramid, and then transmitting the feature pyramid to an RPN.
And after further extraction of 5 layers of features of the feature pyramid through a 3 × 3 same convolution network, performing background/foreground binary classification on 3 shapes of each feature point through 1 × 1 convolution, and performing regression on the target area of the 3 shapes through another 1 × 1 convolution network to adjust the central x offset, the central y offset, the width and the high scaling. Finally, a background/foreground classification result ([ batch, anchors,2]) and an offset regression result ([ batch, anchors,4]) are generated, wherein the anchor is in a height width value of 3 anchor ratios.
and combining the two generated results with 5-layer features of the feature pyramid to perform ROIAlign (regional feature clustering). Specifically, 6000 foreground features with the highest score are extracted, and a frame with high contact ratio is removed through NMS (non-maximum suppression). The first 2000 boxes with targets are finally output. And finally, performing bilinear interpolation scaling on each feature point by combining the target frame with 5 layers of features of the feature pyramid through ROIAlign to adjust the feature points into a uniform size, and performing convolution with 1 x 1 through full convolution to obtain a final classification result. And then, taking the target frames with higher scores in combination with the classification credibility, and returning the number of the corresponding target frames, namely the number of the on-duty personnel in the whole monitoring picture. And when the number of the people on duty is less than the preset number of the people on duty, judging that the condition of illegal off duty exists, or when the number of the people on duty acquired for multiple times in a period of time is less than the preset number, indicating that the people are off duty.
it should be noted that ROI Align is a region feature aggregation method, and well solves the problem of region mismatch (mis-alignment) caused by two quantization operations in the ROI Pooling operation.
Further, before the image extraction is performed on the monitoring video at the preset time interval, the method further includes:
and setting the preset target post arrival number according to a setting instruction input by a user.
Further, the method further comprises:
and when judging that the personnel are illegally off duty, sending an alarm instruction to control the alarm device to remind the manager.
It should be noted that the above method or flow embodiment is described as a series of acts or combinations for simplicity, but those skilled in the art should understand that the present invention is not limited by the described acts or sequences, as some steps may be performed in other sequences or simultaneously according to the present invention. Further, those skilled in the art will appreciate that the embodiments described in the specification are exemplary embodiments and that no single embodiment is necessarily required by the inventive embodiments.
referring to fig. 2, in order to solve the same technical problem, the present invention further provides a device for detecting illegal people leaving behind based on image recognition, which includes an image extraction module, an image feature extraction module, an area suggestion image generation module, an on-duty personnel number calculation module and a personnel leaving behind judgment module; wherein the content of the first and second substances,
the image extraction module is used for extracting images of the monitoring video according to a preset time interval;
the image feature extraction module is used for extracting features of the extracted image by using the trained convolutional neural network to obtain an image feature result;
the region suggestion image generation module is used for performing target scanning on the extracted image features in a sliding window mode by using a region suggestion network to obtain region suggestion image features;
the on-duty personnel number calculating module is used for framing the image characteristics according to the suggested region to perform ROI processing, and calculating the on-duty personnel number according to the calculated foreground scores of all the target frames in the image;
And the personnel off-duty judging module is used for judging that the personnel illegally leave the duty when the number of the personnel on duty is less than the preset target number of the people arriving on duty.
Further, the image feature extraction module is specifically configured to:
After the extracted image is preprocessed, performing convolution on the preprocessed image by using a preset convolution kernel to obtain a plurality of characteristic results;
After the plurality of characteristic results are subjected to normalization processing and Relu activation, extracting N layers of pooling layers through a resnet101 network model to form a characteristic pyramid;
And after the feature extraction is carried out on the N layers of features of the feature pyramid by using the same convolutional network, carrying out shape classification through the first convolutional network to obtain a background/foreground classification result, and simultaneously carrying out shape regression through the second convolutional network to obtain an offset regression result.
Further, the image feature extraction module is specifically configured to:
performing target scanning on the extracted image features by using a regional suggestion network in a sliding window mode, extracting i foreground features of which the scores are higher than a preset threshold value from the image features, and obtaining an initial regional suggestion image;
NMS processing is carried out on the shifted suggested areas, j target frames in the suggested images of the initial areas are screened out, and the suggested images of the areas are obtained; wherein j < i.
Further, the image feature extraction module is specifically configured to:
Combining the region suggested image and the N layers of features of the feature pyramid, and performing region feature aggregation by using a ROIAlign layer to obtain a final classification result;
And calculating the foreground score of each target frame in the image according to the final classification result, and acquiring the number of the final target frames according to a preset classification credibility threshold value to obtain the number of the on-duty personnel.
it can be understood that the above system item embodiment corresponds to the method item embodiment of the present invention, and the device for detecting illegal people leaving off post based on image recognition provided by the present invention can implement the method for detecting illegal people leaving off post based on image recognition provided by any method item embodiment of the present invention.
Compared with the prior art, the invention has the following beneficial effects:
the invention discloses a method and a device for detecting illegal personnel off duty based on image recognition, wherein the method comprises the following steps: performing image extraction on the monitoring video according to a preset time interval; carrying out feature extraction on the extracted image by using the trained convolutional neural network to obtain an image feature result; performing target scanning on the extracted image features by using a regional suggestion network in a sliding window mode to obtain regional suggestion image features; framing image features according to the suggested region to perform ROI processing, and calculating the number of on-duty personnel according to the calculated foreground scores of all target frames in the image; and when the number of the people on duty is less than the preset target number of the people on duty, judging that the people are illegally off duty. According to the invention, by combining the regional neural network and the convolutional neural network, whether the monitored personnel are illegally off duty can be accurately and effectively judged, so that security holes caused by illegal off duty of the personnel can be effectively avoided.
While the foregoing is directed to the preferred embodiment of the present invention, it will be understood by those skilled in the art that various changes and modifications may be made without departing from the spirit and scope of the invention.
Claims (10)
1. A person illegal off-duty detection method based on image recognition is characterized by comprising the following steps:
performing image extraction on the monitoring video according to a preset time interval;
carrying out feature extraction on the extracted image by using the trained convolutional neural network to obtain an image feature result;
Performing target scanning on the extracted image features by using a regional suggestion network in a sliding window mode to obtain regional suggestion image features;
framing image features according to the suggested region to perform ROI processing, and calculating the number of on-duty personnel according to the calculated foreground scores of all target frames in the image;
and when the number of the people on duty is less than the preset target number of the people on duty, judging that the people are illegally off duty.
2. The image recognition-based illegal people off duty detection method according to claim 1, wherein the trained convolutional neural network is used for extracting features of the extracted image to obtain an image feature result, and specifically comprises the following steps:
after the extracted image is preprocessed, performing convolution on the preprocessed image by using a preset convolution kernel to obtain a plurality of characteristic results;
after the plurality of characteristic results are subjected to normalization processing and Relu activation, extracting N layers of pooling layers through a resnet101 network model to form a characteristic pyramid;
and after the feature extraction is carried out on the N layers of features of the feature pyramid by using the same convolutional network, carrying out shape classification through the first convolutional network to obtain a background/foreground classification result, and simultaneously carrying out shape regression through the second convolutional network to obtain an offset regression result.
3. The illegal people off duty detection method based on image recognition according to claim 2, characterized in that the target scanning is performed on the extracted image features by using the regional suggestion network in a sliding window manner to obtain regional suggestion image features, specifically:
performing target scanning on the extracted image features by using a regional suggestion network in a sliding window mode, extracting i foreground features of which the scores are higher than a preset threshold value from the image features, and obtaining an initial regional suggestion image;
NMS processing is carried out on the shifted suggested areas, j target frames in the suggested images of the initial areas are screened out, and the suggested images of the areas are obtained; wherein j < i.
4. the image-recognition-based illegal people off duty detection method according to claim 3, wherein the image features are framed according to the suggested region for ROI processing, and the number of people on duty is calculated according to the foreground score of each target frame in the calculated image, specifically:
Combining the region suggested image and the N layers of features of the feature pyramid, and performing region feature aggregation by using a ROIAlign layer to obtain a final classification result;
and calculating the foreground score of each target frame in the image according to the final classification result, and acquiring the number of the final target frames according to a preset classification credibility threshold value to obtain the number of the on-duty personnel.
5. the illegal people off duty detection method based on image recognition according to claim 1, characterized in that before the image extraction of the surveillance video at the preset time interval, the method further comprises:
And setting the preset target post arrival number according to a setting instruction input by a user.
6. The illegal people off duty detection method based on image recognition according to claim 1, characterized by further comprising:
and when judging that the personnel are illegally off duty, sending an alarm instruction to control the alarm device to remind the manager.
7. A detection device for illegal people leaving behind post based on image recognition is characterized by comprising an image extraction module, an image feature extraction module, an area suggestion image generation module, an on-post personnel number calculation module and a personnel leaving behind post judgment module; wherein the content of the first and second substances,
the image extraction module is used for extracting images of the monitoring video according to a preset time interval;
the image feature extraction module is used for extracting features of the extracted image by using the trained convolutional neural network to obtain an image feature result;
The region suggestion image generation module is used for performing target scanning on the extracted image features in a sliding window mode by using a region suggestion network to obtain region suggestion image features;
The on-duty personnel number calculating module is used for framing the image characteristics according to the suggested region to perform ROI processing, and calculating the on-duty personnel number according to the calculated foreground scores of all the target frames in the image;
and the personnel off-duty judging module is used for judging that the personnel illegally leave the duty when the number of the personnel on duty is less than the preset target number of the people arriving on duty.
8. the device for detecting illegal people leaving post based on image recognition according to claim 7, wherein the image feature extraction module is specifically configured to:
after the extracted image is preprocessed, performing convolution on the preprocessed image by using a preset convolution kernel to obtain a plurality of characteristic results;
After the plurality of characteristic results are subjected to normalization processing and Relu activation, extracting N layers of pooling layers through a resnet101 network model to form a characteristic pyramid;
And after the feature extraction is carried out on the N layers of features of the feature pyramid by using the same convolutional network, carrying out shape classification through the first convolutional network to obtain a background/foreground classification result, and simultaneously carrying out shape regression through the second convolutional network to obtain an offset regression result.
9. The device for detecting illegal people leaving post based on image recognition according to claim 8, wherein the image feature extraction module is specifically configured to:
performing target scanning on the extracted image features by using a regional suggestion network in a sliding window mode, extracting i foreground features of which the scores are higher than a preset threshold value from the image features, and obtaining an initial regional suggestion image;
NMS processing is carried out on the shifted suggested areas, j target frames in the suggested images of the initial areas are screened out, and the suggested images of the areas are obtained; wherein j < i.
10. the device for detecting illegal people leaving post based on image recognition according to claim 9, wherein the image feature extraction module is specifically configured to:
Combining the region suggested image and the N layers of features of the feature pyramid, and performing region feature aggregation by using a ROIAlign layer to obtain a final classification result;
and calculating the foreground score of each target frame in the image according to the final classification result, and acquiring the number of the final target frames according to a preset classification credibility threshold value to obtain the number of the on-duty personnel.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910776281.2A CN110580455A (en) | 2019-08-21 | 2019-08-21 | image recognition-based illegal off-duty detection method and device for personnel |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910776281.2A CN110580455A (en) | 2019-08-21 | 2019-08-21 | image recognition-based illegal off-duty detection method and device for personnel |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110580455A true CN110580455A (en) | 2019-12-17 |
Family
ID=68811901
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910776281.2A Pending CN110580455A (en) | 2019-08-21 | 2019-08-21 | image recognition-based illegal off-duty detection method and device for personnel |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110580455A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111931691A (en) * | 2020-08-31 | 2020-11-13 | 四川骏逸富顿科技有限公司 | On-duty monitoring method and monitoring system thereof |
| CN113052049A (en) * | 2021-03-18 | 2021-06-29 | 国网内蒙古东部电力有限公司 | Off-duty detection method and device based on artificial intelligence tool identification |
| CN117253176A (en) * | 2023-11-15 | 2023-12-19 | 江苏海内软件科技有限公司 | Safe production Al intelligent detection method based on video analysis and computer vision |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106570467A (en) * | 2016-10-25 | 2017-04-19 | 南京南瑞集团公司 | Convolutional neutral network-based worker absence-from-post detection method |
| CN108416250A (en) * | 2017-02-10 | 2018-08-17 | 浙江宇视科技有限公司 | Demographic method and device |
| CN109711320A (en) * | 2018-12-24 | 2019-05-03 | 兴唐通信科技有限公司 | A kind of operator on duty's unlawful practice detection method and system |
-
2019
- 2019-08-21 CN CN201910776281.2A patent/CN110580455A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106570467A (en) * | 2016-10-25 | 2017-04-19 | 南京南瑞集团公司 | Convolutional neutral network-based worker absence-from-post detection method |
| CN108416250A (en) * | 2017-02-10 | 2018-08-17 | 浙江宇视科技有限公司 | Demographic method and device |
| CN109711320A (en) * | 2018-12-24 | 2019-05-03 | 兴唐通信科技有限公司 | A kind of operator on duty's unlawful practice detection method and system |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111931691A (en) * | 2020-08-31 | 2020-11-13 | 四川骏逸富顿科技有限公司 | On-duty monitoring method and monitoring system thereof |
| CN111931691B (en) * | 2020-08-31 | 2021-01-05 | 四川骏逸富顿科技有限公司 | On-duty monitoring method and monitoring system thereof |
| CN113052049A (en) * | 2021-03-18 | 2021-06-29 | 国网内蒙古东部电力有限公司 | Off-duty detection method and device based on artificial intelligence tool identification |
| CN113052049B (en) * | 2021-03-18 | 2023-12-19 | 国网内蒙古东部电力有限公司 | Off-duty detection method and device based on artificial intelligent tool identification |
| CN117253176A (en) * | 2023-11-15 | 2023-12-19 | 江苏海内软件科技有限公司 | Safe production Al intelligent detection method based on video analysis and computer vision |
| CN117253176B (en) * | 2023-11-15 | 2024-01-26 | 江苏海内软件科技有限公司 | Safe production Al intelligent detection method based on video analysis and computer vision |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106600888B (en) | Automatic forest fire detection method and system | |
| CN110580455A (en) | image recognition-based illegal off-duty detection method and device for personnel | |
| CN107679471B (en) | Indoor personnel air post detection method based on video monitoring platform | |
| CN101141608B (en) | Video instant communication system and method | |
| CN108052900A (en) | A kind of method by monitor video automatic decision dressing specification | |
| CN106339657B (en) | Crop straw burning monitoring method based on monitor video, device | |
| CN106372576A (en) | Deep learning-based intelligent indoor intrusion detection method and system | |
| CN111091098B (en) | Training method of detection model, detection method and related device | |
| CN109344688A (en) | The automatic identifying method of people in a kind of monitor video based on convolutional neural networks | |
| CN112733690A (en) | High-altitude parabolic detection method and device and electronic equipment | |
| US11048973B1 (en) | Action classification using aggregated background subtraction images | |
| CN107330414A (en) | Act of violence monitoring method | |
| CN110096945B (en) | Indoor monitoring video key frame real-time extraction method based on machine learning | |
| CN112257669A (en) | Pedestrian re-identification method and device and electronic equipment | |
| CN112153373A (en) | Fault identification method and device for bright kitchen range equipment and storage medium | |
| CN108446690A (en) | A kind of human face in-vivo detection method based on various visual angles behavioral characteristics | |
| Zhang et al. | A novel camera-based drowning detection algorithm | |
| CN110222647B (en) | Face in-vivo detection method based on convolutional neural network | |
| CN115953719A (en) | Multi-target recognition computer image processing system | |
| CN107317992A (en) | video communication control device and control method | |
| CN110516538B (en) | Prison double off-duty violation assessment method based on deep learning target detection | |
| KR101600617B1 (en) | Method for detecting human in image frame | |
| Li et al. | Research on forest fire image recognition algorithm based on color feature statistics | |
| CN114332983A (en) | Face image definition detection method, face image definition detection device, electronic equipment and medium | |
| CN111814590B (en) | Personnel safety state monitoring method, equipment and computer readable storage medium |
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 | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: 510000 room 414, room 413, No. 662, Huangpu Avenue middle, Tianhe District, Guangzhou City, Guangdong Province (office only) Applicant after: GUANGZHOU HONGSEN TECHNOLOGY Co.,Ltd. Address before: 510000 Room 501 and 502, 233 Tianhe North Road, Tianhe District, Guangzhou City, Guangdong Province Applicant before: GUANGZHOU HONGSEN TECHNOLOGY Co.,Ltd. |