CN110807753A - Radioactive source monitoring method and device and electronic equipment - Google Patents
Radioactive source monitoring method and device and electronic equipment Download PDFInfo
- Publication number
- CN110807753A CN110807753A CN201810803761.9A CN201810803761A CN110807753A CN 110807753 A CN110807753 A CN 110807753A CN 201810803761 A CN201810803761 A CN 201810803761A CN 110807753 A CN110807753 A CN 110807753A
- Authority
- CN
- China
- Prior art keywords
- source
- image
- radioactive source
- video
- feature
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30241—Trajectory
Abstract
The invention discloses a method and a device for monitoring a radioactive source and electronic equipment, wherein the method comprises the steps of acquiring a video to be detected; extracting a video image in a video to be detected; performing feature extraction on the video image by using a radioactive source classifier; matching the extracted features with features in a radioactive source feature database; the radioactive source feature database is a feature database of a radioactive source obtained by performing feature training by using an image of the radioactive source; and when the extracted features are matched and consistent with the features in the radioactive source feature database, determining that the video image corresponding to the extracted features is a source image containing the radioactive source. The method can be used for detecting the radioactive source in real time in video monitoring of video images in the video stream, the image involved in the video is determined through feature extraction and feature matching, and after machine learning is carried out on the image of the radioactive source, the image involved in the video is determined through feature matching, so that high radioactive source detection can be guaranteed, and the detection precision is improved.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a method and a device for monitoring a radioactive source and electronic equipment.
Background
Radioactive sources are a generic term for small compact sources of radiation made from radioactive materials, the basic feature of which is the ability to provide useful radiation continuously. With the development of market economy and the further opening of various fields, radioactive sources have been widely applied to various fields such as industry, medicine, environmental and scientific research as a high-tech product.
However, the radiation source emits radiation having an energy that destroys cell tissue, causing harm to the human body. The international atomic energy agency classifies the radioactive sources into 5 classes (class i-class v) according to the degree of possible damage to the human body by the radioactive sources. According to survey data, the stolen or out-of-control radioactive sources in China mostly belong to IV or V types.
In the prior art, the detection of IV-type or V-type radioactive sources mainly depends on various radiation detectors, the core component of the radiation detector is a sensor, and the task of the radiation detector is to convert various physical, chemical and other variable information needing to be detected into measurable electric signals and then transmit the measurable electric signals to a chip for calculation to obtain a result. However, the radiation intensity of the class v radioactive source is low, so that the detection sensitivity of the radiation detector for the class v radioactive source is not sufficient.
Disclosure of Invention
In view of this, the embodiment of the present invention provides a method for monitoring a radioactive source, so as to solve the problem in the prior art that the detection sensitivity of the radioactive source is relatively low.
According to a first aspect, an embodiment of the present invention provides a method for monitoring a radiation source, including:
acquiring a video to be detected;
extracting a video image in the video to be detected;
performing feature extraction on the video image by using a radioactive source classifier;
matching the extracted features with features in a radioactive source feature database; the radioactive source feature database is obtained by performing feature training on images of radioactive sources;
and when the extracted features are matched and consistent with the features in the radioactive source feature database, determining that the video image corresponding to the extracted features is a source image containing the radioactive source.
The method for monitoring the radioactive source provided by the embodiment of the invention can be used for detecting the radioactive source in real time on the video image in the video stream while monitoring the video, and particularly determines the source image containing the radioactive source through feature extraction and feature matching.
With reference to the first aspect, in a first implementation manner of the first aspect, the radiation source feature database stores feature values corresponding to features of a plurality of radiation sources;
wherein, the matching of the extracted features with the features in the radiation source feature database comprises:
calculating a feature value of the extracted feature;
comparing the calculated characteristic values with the similarity of the characteristic values in the radioactive source characteristic database;
and when the similarity is within a preset range, determining that the extracted features are matched with features in the radioactive source feature database.
According to the monitoring method of the radioactive source provided by the embodiment of the invention, the characteristic value corresponding to the extracted characteristic is calculated and compared with the characteristic value in the radioactive source characteristic database, so that the matching step is simplified, the source-related image can be determined only through the characteristic value, and the efficiency of radioactive source monitoring is improved.
With reference to the first aspect, in a second implementation manner of the first aspect, after the step of determining that the extracted feature is that the corresponding video image is a source image including a radiation source, the method further includes:
extracting a source image containing the radioactive source in the video to be detected;
detecting source-related personnel in the source-related image by utilizing a personnel classifier;
tracking the radiation source based on the image of the source-related person.
According to the monitoring method of the radioactive source provided by the embodiment of the invention, source-related personnel in the source-related image are combined with the radioactive source, namely the radioactive source is positioned to the source-related personnel, and the position of the radioactive source can be determined by tracking the source-related personnel, so that the radioactive source can be accurately and efficiently tracked.
With reference to the second implementation manner of the first aspect, in a third implementation manner of the first aspect, the detecting, by using a face classifier, a source-related person in the source-related image includes:
extracting a person image in the source-related image;
extracting the features of the personnel images by utilizing the personnel classifier;
matching the extracted features with features in a personnel feature database; the personnel feature database is a feature database of a preset personnel obtained after feature training is carried out by utilizing an image of the preset personnel;
and when the extracted features are matched with the features in the feature database of the preset personnel in a consistent manner, determining the personnel image corresponding to the extracted features as the image of the source-related personnel.
According to the method for monitoring the radioactive source, provided by the embodiment of the invention, the source related personnel in the source related image are determined through feature extraction and feature matching.
With reference to the second embodiment of the first aspect, in a fourth embodiment of the first aspect, the tracking the radiation source based on the image of the source related person includes:
sequentially identifying the source related personnel in each frame of video image in the video to be detected;
recording the position of the source-related personnel;
combining all recorded positions of the source personnel to form a motion track of the source personnel.
According to the monitoring method of the radioactive source provided by the embodiment of the invention, the radioactive source can be conveniently tracked through the motion trail of source-related personnel.
With reference to the first aspect, in a fifth implementation manner of the first aspect, before the step of acquiring the video to be detected, the method further includes:
acquiring images of a plurality of preset radioactive sources;
performing feature training based on the image of the radioactive source to form a radioactive source feature database;
and/or the presence of a gas in the gas,
before the step of obtaining the video to be detected, the method further comprises the following steps:
acquiring images of a plurality of preset persons;
and performing feature training based on the images of the preset personnel to form a personnel database.
According to the monitoring method of the radioactive source provided by the embodiment of the invention, machine learning is carried out by utilizing the image of the preset radioactive source and the image of the preset personnel so as to respectively form the radioactive source characteristic database and the personnel database, namely, the tracking of the radioactive source and the personnel involved in the radioactive source is carried out by utilizing the characteristics of the preset radioactive source and the preset personnel, so that conditions are provided for further subsequent characteristic matching.
According to a second aspect, the present invention also provides a monitoring device for a radioactive source, comprising:
the acquisition module is used for acquiring a video to be detected;
the first detection module is used for detecting the radioactive source in the video to be detected by utilizing the radioactive source classifier;
the image extraction module is used for extracting a video image in the video to be detected;
the characteristic extraction module is used for extracting the characteristics of the video image by using the radioactive source classifier;
the characteristic matching module is used for matching the extracted characteristics with the characteristics in the radioactive source characteristic database; the radioactive source feature database is obtained by performing feature training on images of radioactive sources;
and the determining module is used for determining that the video image corresponding to the extracted features is a source image containing the radioactive source when the extracted features are matched and consistent with the features in the radioactive source feature database.
The monitoring device for the radioactive source provided by the embodiment of the invention can be used for carrying out real-time radioactive source detection on video images in a video stream while carrying out video monitoring, and particularly determining a source image containing the radioactive source through feature extraction and feature matching.
With reference to the second aspect, in the first embodiment of the second aspect, the radiation source feature database stores feature values corresponding to features of a plurality of radiation sources;
the image extraction module comprises:
a calculation unit configured to calculate a feature value of the extracted feature;
the comparison unit is used for comparing the similarity between the calculated characteristic value and the characteristic value in the radioactive source characteristic database;
and the determining unit is used for determining that the extracted features are matched with the features in the radioactive source feature database when the similarity is within a preset range.
According to a third aspect, an embodiment of the present invention further provides an electronic device, including:
a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing therein computer instructions, and the processor executing the computer instructions to perform the method for monitoring a radiation source according to the first aspect of the present invention or any embodiment of the first aspect.
According to a fourth aspect, the present invention further provides a computer-readable storage medium, which stores computer instructions for causing a computer to execute the method for monitoring a radiation source according to the first aspect of the present invention or any implementation manner of the first aspect.
Drawings
The features and advantages of the present invention will be more clearly understood by reference to the accompanying drawings, which are illustrative and not to be construed as limiting the invention in any way, and in which:
FIG. 1 illustrates a flow chart of a particular schematic of a method of monitoring a radiation source in an embodiment of the present invention;
FIG. 2 illustrates another detailed schematic flow chart of a method of monitoring a radiation source in an embodiment of the present invention;
FIG. 3 illustrates another detailed schematic flow chart of a method of monitoring a radiation source in an embodiment of the present invention;
FIG. 4 illustrates another detailed schematic flow chart of a method of monitoring a radiation source in an embodiment of the present invention;
FIG. 5 illustrates another detailed schematic flow chart of a method for monitoring a radiation source in an embodiment of the present invention;
FIG. 6 is a diagram illustrating a structural model of a convolutional neural network in an embodiment of the present invention;
FIG. 7 is a block diagram showing a specific schematic of a monitoring apparatus for a radiation source in an embodiment of the present invention;
FIG. 8 is a block diagram showing another specific illustration of a monitoring device for a radiation source in an embodiment of the present invention;
FIG. 9 is a block diagram showing another specific illustration of a monitoring device for a radiation source in an embodiment of the present invention;
fig. 10 shows a specific schematic structural diagram of an electronic device in the embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. 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.
The embodiment of the invention provides a monitoring method of a radioactive source, which can be applied to electronic equipment and is used for monitoring the radioactive source. As shown in fig. 1, the method includes:
and S11, acquiring the video to be detected.
The video to be detected can be a real-time monitoring video stream, or a video to be detected provided to the electronic device from the outside, or a video to be detected provided in other modes, and only the electronic device needs to be ensured to be capable of acquiring the video to be detected.
And S12, extracting the video image in the video to be detected.
The electronic device extracts a video image in a video to be detected, specifically, the obtained video to be detected may be decoded and frame-split to obtain a video image of each frame, or the video image in the video to be detected may be extracted in other manners.
And S13, performing feature extraction on the video image by using the radioactive source classifier.
The radioactive source classifier is obtained by performing machine learning on images of radioactive sources in advance. The electronic equipment can extract the characteristics of the video image by using the radioactive source classifier, namely, the entity in the video image is extracted.
And S14, matching the extracted features with features in the radioactive source feature database.
The radioactive source feature database is a feature database of a radioactive source obtained by performing feature training by using an image of the radioactive source.
The electronic equipment stores a radioactive source feature database, the database stores the features of a plurality of radioactive sources, and the features extracted by the radioactive source classifier are matched with the features in the radioactive source feature database, so that whether the extracted features are the features of the radioactive sources or not can be determined.
Specifically, the matching judgment can be performed by calculating the similarity between two matched features, or by calculating the feature value of the extracted feature and combining the feature value corresponding to the feature in the radiation source feature database, and the like.
And S15, when the extracted features are matched and consistent with the features in the radioactive source feature database, determining that the video image corresponding to the extracted features is a source-involved image containing a radioactive source.
When the features extracted by the radioactive source classifier are matched with the features in the radioactive source feature database in a consistent manner, the electronic equipment can determine that the features extracted by the radioactive source classifier are the features of the radioactive source at the moment, and then the video images corresponding to the extracted features are determined as the source image containing the radioactive source.
The method for monitoring the radioactive source provided by the embodiment of the invention can be used for detecting the radioactive source in real time on the video image in the video stream while monitoring the video, and particularly determines the source image containing the radioactive source through feature extraction and feature matching.
An embodiment of the present invention provides a method for monitoring a radiation source, as shown in fig. 2, the method includes:
and S21, acquiring images of a plurality of preset radioactive sources.
The image of the preset radioactive source can be an image of a radioactive source related in a monitoring unit, and the image is specifically an appearance image of the preset radioactive source during storage. The images of the involved radiation sources are provided to an electronic device, or other device, for subsequent formation of a radiation source characteristic database based on the images of the radiation sources.
In addition, the image of the preset radioactive source can also be an image of other types of radioactive sources, and is not limited to the radioactive source related to the monitoring unit.
And S22, performing feature training based on the image of the radioactive source to form a radioactive source feature database.
After the images of the radiation source are acquired, the images of the radiation source are feature trained using a deep learning algorithm, such as the CNN algorithm, or other algorithms to form a radiation source feature database.
And S23, acquiring the video to be detected. Please refer to S11 in fig. 1, which is not described herein again.
And S24, extracting the video image in the video to be detected. Please refer to S12 in fig. 1, which is not described herein again.
And S25, performing feature extraction on the video image by using the radioactive source classifier. Please refer to S12 in fig. 1, which is not described herein again.
And S26, matching the extracted features with features in the radioactive source feature database.
The radioactive source feature database is a feature database of a radioactive source obtained by performing feature training by using an image of the radioactive source. In addition, the characteristic value corresponding to the characteristics of a plurality of radioactive sources is stored in the radioactive source characteristic database.
Specifically, the steps include:
s261, calculates a feature value of the extracted feature.
The algorithm for calculating the characteristic value corresponding to the characteristic in the radioactive source characteristic database is the same as the algorithm for calculating the characteristic value of the extracted characteristic, namely the same algorithm is used for calculating the characteristic value of the extracted characteristic and the characteristic value corresponding to the characteristic in the radioactive source characteristic database, so that the matching reference is the same when the matching is carried out subsequently.
S262, comparing the similarity of the calculated characteristic value and the characteristic value in the radioactive source characteristic database.
After calculating the characteristic values corresponding to the extracted features, the electronic device sequentially calculates the similarity between the characteristic values and all the characteristic values in the radiation source characteristic data, that is, the distance between the two characteristic values can be calculated. If the distance is within the preset range, the extracted features are matched with the features in the radioactive source feature database, namely the extracted features belong to the features of the radioactive source; otherwise, it indicates that the extracted features do not belong to the features of the radiation source.
And S263, when the similarity is within a preset range, determining that the extracted features are matched with features in the radioactive source feature database.
When the similarity of the two characteristic values (the characteristic value corresponding to the extracted characteristic and the characteristic value of the characteristic in the radioactive source characteristic database) is calculated to be within a preset range, the electronic equipment shows that the extracted characteristic is matched with the characteristic in the radioactive source characteristic database.
The preset range can be specifically set according to specific detection precision, the detection precision is high, and the corresponding preset range is small; the detection precision is low, and the corresponding preset range is large.
And S27, when the extracted features are matched with the features in the radioactive source feature database, determining that the video image corresponding to the extracted features is a source-involved image containing the radioactive source.
When the electronic equipment determines that the features are matched, the extracted features belong to the features of the radioactive source, and meanwhile, the video image corresponding to the extracted features is a source-related image containing the radioactive source.
Compared with the embodiment shown in fig. 1, the method for monitoring the radioactive source, which is provided by the embodiment of the invention, simplifies the matching steps by comparing the characteristic values corresponding to the extracted features with the characteristic values in the radioactive source feature database, can determine the source-related image only through the characteristic values, and improves the efficiency of radioactive source monitoring.
The embodiment of the present invention further provides a method for monitoring a radioactive source, as shown in fig. 3, the method includes:
and S31, acquiring the video to be detected. Please refer to S23 in fig. 2 for details, which are not described herein.
And S32, extracting the video image in the video to be detected. Please refer to S24 in fig. 2 for details, which are not described herein.
And S33, performing feature extraction on the video image by using the radioactive source classifier. Please refer to S25 in fig. 2 for details, which are not described herein.
And S34, matching the extracted features with features in the radioactive source feature database. Please refer to S26 in fig. 2 for details, which are not described herein.
The radioactive source feature database is a feature database of a radioactive source obtained by performing feature training by using an image of the radioactive source.
And S35, when the extracted features are matched and consistent with the features in the radioactive source feature database, determining that the video image corresponding to the extracted features is a source-involved image containing a radioactive source. Please refer to S27 in fig. 2 for details, which are not described herein.
In this embodiment, after the source-related image is determined, a method for detecting and tracking source-related personnel based on the source-related image is provided, which specifically includes:
and S36, extracting the source-related image containing the radiation source in the video to be detected.
After the source concerning image in the video to be detected is determined, the electronic equipment extracts the source concerning image containing the radioactive source, and detection of source concerning personnel is carried out on the basis of the source concerning image.
And S37, detecting source related personnel in the source related image by utilizing the personnel classifier.
The personnel classifier is obtained by training personnel images. The person image is an image of a worker of the monitoring unit or, further, includes basic information (e.g., name, age, position, etc.) of the person.
In this embodiment, the electronic device may detect the source-related person in the source-related image through the person classifier.
S38, tracking the radioactive source based on the image of the personnel involved in the source.
After the source-related image and the source-related personnel in the source-related image are detected, the motion trail of the source-related personnel can be tracked, so that the tracking of the radioactive source is realized. For example, the motion trail of the source-related person can be determined through the position information of the source-related person in the video image of the adjacent frame in the video to be detected.
Compared with the embodiment shown in fig. 2, the method for monitoring the radioactive source provided by the embodiment of the invention determines the source-related personnel in the source-related image through feature extraction and feature matching.
As an alternative implementation of this embodiment, before S31, the method further includes the same steps as S21 and S22 of the embodiment shown in fig. 2, that is, the method further includes the step of forming a radioactive source feature database.
The embodiment of the present invention further provides a method for monitoring a radioactive source, as shown in fig. 4, the method includes:
s401, images of a plurality of preset persons are obtained.
The image of the preset person can be an image of a worker in a monitoring unit, and the image is specifically an appearance image of the worker. The images of the staff members are provided to an electronic device, or other device, for subsequent formation of a radiation source characteristic database based on the staff images.
In addition, the preset personnel image may also be an image of other personnel, and is not limited to the staff of the monitoring unit.
S402, performing feature training based on images of preset people to form a people database.
After the preset personnel image is acquired, feature training is performed on the preset personnel image by using a deep learning algorithm, such as a CNN algorithm, or other algorithms to form a personnel database.
And S403, acquiring the video to be detected. Please refer to S31 in fig. 3 for details, which are not described herein.
S404, extracting the video image in the video to be detected. Please refer to S32 in fig. 3 for details, which are not described herein.
S405, extracting the characteristics of the video image by using the radioactive source classifier. Please refer to S33 in fig. 3 for details, which are not described herein.
And S406, matching the extracted features with features in a radiation source feature database. Please refer to S34 in fig. 3 for details, which are not described herein.
The radioactive source feature database is a feature database of a radioactive source obtained by performing feature training by using an image of the radioactive source.
S407, when the extracted features are matched and consistent with the features in the radioactive source feature database, determining that the video image corresponding to the extracted features is a source-related image containing a radioactive source. Please refer to S35 in fig. 3 for details, which are not described herein.
And S408, extracting the source involved image containing the radioactive source in the video to be detected. Please refer to S36 in fig. 3 for details, which are not described herein.
And S409, detecting source-related personnel in the source-related image by utilizing a personnel classifier.
In this embodiment, the step of detecting the source-related person in the source-related image by using the person classifier is similar to the step of detecting the source-related image in the video image by using the radioactive source classifier, and mainly includes: feature extraction and feature matching. The method specifically comprises the following steps:
and S4091, extracting the person image in the source image.
And S4092, extracting the characteristics of the personnel image by utilizing a personnel classifier.
And S4093, matching the extracted features with features in a personnel feature database.
The personnel feature database is a feature database of the preset personnel obtained after feature training is carried out by utilizing images of the preset personnel.
And S4094, when the extracted features are matched with the features in the feature database of the preset personnel, determining the personnel image corresponding to the extracted features as the image of the source-related personnel.
S4091 to S4094 are similar to S24 to S27 of the embodiment shown in fig. 2, and please refer to S24 to S27 of the embodiment shown in fig. 2, which are not repeated herein.
And S410, tracking the radiation source based on the image of the personnel involved in the source.
In this embodiment, the tracking of the radiation source is realized by tracking the movement locus of the personnel involved in the source. The method specifically comprises the following steps:
s4101, sequentially identifying source related personnel in each frame of video image in the video to be detected.
After determining source-related personnel in the source-related image, the electronic equipment sequentially identifies the source-related personnel in each frame of video image so as to determine the positions of the source-related personnel.
S4102, recording the location of the person involved in the source.
The electronic equipment sequentially records the positions of source-related personnel according to the sequence of video images in the video to be detected, and the motion trail of the source-related personnel can be determined through a plurality of position points of the source-related personnel.
And S4103, combining all recorded positions of the source personnel to form a motion track of the source personnel.
And the electronic equipment combines the positions of all recorded source-related personnel to form the motion trail of the source-related personnel.
Compared with the embodiment shown in fig. 3, in the embodiment, the source-related personnel in the source-related image are determined through feature extraction and feature matching, in the method, the source-related personnel are determined through feature matching after the image of the preset personnel is used for machine learning, so that high detection of the source-related personnel can be ensured, and the detection precision is improved; in addition, the tracking of the radioactive source is convenient to realize through the motion trail of source-related personnel.
The embodiment of the present invention further provides a method for monitoring a radioactive source, as shown in fig. 5, the method includes:
and S51, acquiring the video to be detected. Please refer to S11 in fig. 1, which is not described herein again.
And S52, extracting the video image in the video to be detected. Please refer to S12 in fig. 1, which is not described herein again.
And S53, detecting the video image by using the radioactive source classifier to determine a source-related image containing the radioactive source.
The electronic equipment detects the video image by using the radioactive source classifier, and determines a source image containing the radioactive source in the video to be detected. In particular, the radiation source classifier detector performs machine learning based on images of a preset radiation source, i.e., enables the radiation source classifier to know which images are images of the radiation source.
And S54, detecting the source-involved image by using the personnel classifier to determine source-involved personnel in the source-involved image. Please refer to S37 in fig. 3 for details, which are not described herein.
S55, tracking the radioactive source based on the image of the personnel involved in the source. Please refer to S38 in fig. 3 for details, which are not described herein.
The embodiment of the invention also provides a specific application example of the radioactive source monitoring method, which comprises the following steps:
the monitoring method for the radioactive source provided by the embodiment includes detection and identification of the radioactive source, motion detection and face identification of source-related personnel, can identify the radioactive source and track the moving path of the radioactive source in a monitoring range, can also identify related source-related personnel, and ensures comprehensive safe handling and control of the radioactive source and the related personnel. The scheme technically comprises target recognition and behavior detection.
Under the deep learning framework, a program algorithm autonomously and directly learns and extracts key features from an input image, and a plurality of processing layers each comprise a certain number of connection points with artificial neurons, so that more and more meaningful data vectors are learned.
And extracting low-level features by using the neurons of the lower layer, then transmitting the low-level features to the neurons of the higher layer, and extracting layer by layer until expected features conforming to the mode classification are identified. Specifically, the functions are realized by adopting a deep convolutional neural network algorithm, and the general process of target identification is as follows: loading a data image; setting a CNN network structure; determining the number of convolution layers, the number of convolution kernels, the kernel size and the pooling size in the step; calling a CNN initialization function to configure the CNN setting; initializing relevant training parameters including a learning rate, sample parameters and training iteration cycle times; the last is the training and testing of CNNs. Wherein, the training process is to teach neural network to learn various radioactive sources.
Early training requires a large amount of radiation source image data to learn the key features of the radiation source to achieve high recognition accuracy. The test process is to test whether the system can accurately detect and identify the radioactive source in the dynamic image. The motion detection principle based on deep learning is actually similar to target recognition. Firstly, preprocessing all dynamic images in a training sample set and a testing sample set, and extracting a target motion foreground through a Gaussian mixture model. Secondly, establishing a sample library for various target behaviors in the training sample set, and defining different types of recognition behaviors as priori knowledge for training the deep learning network. And finally, classifying and identifying various behaviors in the test sample set by combining with the network model obtained by deep learning.
As shown in fig. 6, the convolutional neural network CNN is a structural model diagram, wherein the convolutional neural network mainly includes a feature extraction layer, a sub-sampling layer, and a feature mapping layer. The C1 layer convolution is to extract the input picture characteristics, and obtain a preliminary characteristic diagram through convolution as the input of the subsampled C2 layer. And the last layer is a full connection layer, and a softmax classifier is adopted to distinguish and classify the characteristics obtained in the front.
Specific deep learning frameworks can be exemplified by tenoflow of ***, cafe of facebook, MatConvnet of matlab, and the like. The deep learning framework adopted in the embodiment is MatConvnet, which is a MATLAB toolbox for the computer vision field. Which contains many CNN calculation blocks such as convolution, normalization, pooling, etc.
The monitoring method of the radioactive source provided by the embodiment can provide a reliable technical means for the management of the radioactive environment by the domestic environmental protection department, and greatly improves the monitoring and management capability of the radioactive source. Meanwhile, the application of the technology can enhance the emergency treatment and disposal capability of the radioactive source emergency, and plays a positive role in the aspects of ensuring the social public safety and the health of people. The method specifically comprises the following steps:
1. determining a face database, a radiation source database and a motion dynamic image training set for CNN early-stage training, and preparing small samples for face images and radiation sources to be tested; the face library can adopt a CIFAR data set, an ImageNet data set and the like; the human behavior training set comprises a KTH database, HOHA and the like. In the process of target recognition, all images of a training set are used for learning the neural network, part of the images in the test set are used for recognizing the neural network, and the other part of the images in the test set are used for testing recognition.
2. Code compiling and compiling are completed, wherein the code compiling is based on deep learning (CNN algorithm) for identifying a radioactive source, a human face and behavior detection; the codes can realize early-stage training and testing, can give the recognition accuracy and the accuracy of the testing result to the training result, and can draw a line graph of which the accuracy changes due to parameter changes so as to compare the testing performance;
3. running codes, firstly training a database, putting training results into a data set, and starting testing after training is finished; in the testing process, a function is called to detect a target, then the selected image is compared with parameters in a data set for putting a training result, and the identity with the highest probability of conformity is taken as a recognition result.
4. And (3) inputting a video as data (the video comprises the radioactive source carried by the person and has movement), starting the test, and enabling the test result to see the result of the radioactive source identified by the system, mark the moving path of the person carried by the person and identify the identity of the carrier. In the target recognition, parameters such as epoch, bachsize and the like of a neural network are modified, and in the human behavior detection, three parameters including the training iteration times, the number of layers of a hidden layer of the network and the number of neurons in each layer are modified and tested respectively, so that the parameter setting for the optimal performance can be compared and analyzed. From the experimental results, it can be seen that the more the iteration times, the more accurate the training is, but the more time is consumed, and as the training times increase, the accuracy of the network tends to be stable.
An embodiment of the present invention further provides a monitoring apparatus for a radiation source, as shown in fig. 7, the apparatus includes:
the acquiring module 61 is configured to acquire a video to be detected.
And the first detection module 62 is configured to detect the radioactive source in the video to be detected by using the radioactive source classifier.
And the image extraction module 63 is configured to extract a video image in the video to be detected.
And the feature extraction module 64 is used for performing feature extraction on the video image by using the radioactive source classifier.
And the feature matching module 65 is used for matching the extracted features with features in the radioactive source feature database. The radioactive source feature database is a feature database of a radioactive source obtained by performing feature training by using an image of the radioactive source.
And the determining module 66 is configured to determine that the video image corresponding to the extracted feature is a source-related image containing the radiation source when the extracted feature matches and coincides with the feature in the radiation source feature database.
The monitoring device for the radioactive source provided by the embodiment of the invention can be used for carrying out real-time radioactive source detection on video images in a video stream while carrying out video monitoring, and particularly determining a source image containing the radioactive source through feature extraction and feature matching.
In some optional embodiments of this embodiment, as shown in fig. 8, the characteristic value corresponding to the characteristic of the plurality of radioactive sources is stored in the radioactive source characteristic database; wherein, the feature matching module 65 includes:
the calculation unit 651 calculates a feature value of the extracted feature.
The comparison unit 652 compares the calculated feature values with the similarity of the feature values in the radiation source feature database.
The determining unit 653 determines that the extracted features match the features in the radiation source feature database when the similarity is within the preset range.
In some optional implementations of this embodiment, as shown in fig. 9, the apparatus further includes:
and the extracting module 71 is used for extracting the source-related image containing the radiation source in the video to be detected.
And the second detection module 72 detects the source related personnel in the source related image by utilizing the personnel classifier.
And the tracking module 73 is used for tracking the radioactive source based on the image of the personnel involved in the source.
An embodiment of the present invention further provides an electronic device, as shown in fig. 10, the electronic device may include a processor 81 and a memory 82, where the processor 81 and the memory 82 may be connected by a bus or in another manner, and fig. 10 illustrates the connection by the bus as an example.
Processor 81 may be a Central Processing Unit (CPU). The Processor 81 may also be other general purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components, or combinations thereof.
The memory 82, which is a non-transitory computer readable storage medium, may be used to store non-transitory software programs, non-transitory computer executable programs, and modules, such as program instructions/modules corresponding to the monitoring method of the radiation source in the embodiment of the present invention (for example, the acquisition module 61, the first detection module 62, the image extraction module 63, the feature extraction module 64, the feature matching module 65, and the determination module 66 shown in fig. 7). The processor 81 executes various functional applications and data processing of the processor by executing non-transitory software programs, instructions and modules stored in the memory 82, namely, the monitoring method of the radiation source in the above method embodiment.
The memory 82 may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created by the processor 81, and the like. Further, the memory 82 may include high speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, the memory 82 may optionally include memory located remotely from the processor 81, which may be connected to the processor 81 via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The one or more modules are stored in the memory 82 and, when executed by the processor 81, perform a method of monitoring a radiation source as in the embodiment of fig. 1-5.
The details of the electronic device may be understood by referring to the corresponding descriptions and effects in the embodiments shown in fig. 1 to fig. 5, which are not described herein again.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. The storage medium may be a magnetic Disk, an optical Disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a Flash Memory (Flash Memory), a Hard Disk (Hard Disk Drive, abbreviated as HDD) or a Solid State Drive (SSD), etc.; the storage medium may also comprise a combination of memories of the kind described above.
Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations fall within the scope defined by the appended claims.
Claims (10)
1. A method of monitoring a radiation source, comprising:
acquiring a video to be detected;
extracting a video image in the video to be detected;
performing feature extraction on the video image by using a radioactive source classifier;
matching the extracted features with features in a radioactive source feature database; the radioactive source feature database is obtained by performing feature training on images of radioactive sources;
and when the extracted features are matched and consistent with the features in the radioactive source feature database, determining that the video image corresponding to the extracted features is a source image containing the radioactive source.
2. The method according to claim 1, wherein the radiation source feature database stores feature values corresponding to features of a plurality of radiation sources;
wherein, the matching of the extracted features with the features in the radiation source feature database comprises:
calculating a feature value of the extracted feature;
comparing the calculated characteristic values with the similarity of the characteristic values in the radioactive source characteristic database;
and when the similarity is within a preset range, determining that the extracted features are matched with features in the radioactive source feature database.
3. The method of claim 1, wherein the step of determining the extracted features as the corresponding video image is a source-related image including a radiation source further comprises:
extracting a source image containing the radioactive source in the video to be detected;
detecting source-related personnel in the source-related image by utilizing a personnel classifier;
tracking the radiation source based on the image of the source-related person.
4. The method of claim 3, wherein the detecting source-related persons in the source-related image using a face classifier comprises:
extracting a person image in the source-related image;
extracting the features of the personnel images by utilizing the personnel classifier;
matching the extracted features with features in a personnel feature database; the personnel feature database is a feature database of a preset personnel obtained after feature training is carried out by utilizing an image of the preset personnel;
and when the extracted features are matched with the features in the feature database of the preset personnel in a consistent manner, determining the personnel image corresponding to the extracted features as the image of the source-related personnel.
5. The method of claim 3, wherein tracking the radiation source based on the image of the source-related person comprises:
sequentially identifying the source related personnel in each frame of video image in the video to be detected;
recording the position of the source-related personnel;
combining all recorded positions of the source personnel to form a motion track of the source personnel.
6. The method according to claim 1, wherein the step of obtaining the video to be detected further comprises, before the step of obtaining the video to be detected:
acquiring images of a plurality of preset radioactive sources;
performing feature training based on the image of the preset radioactive source to form the radioactive source feature database;
and/or the presence of a gas in the gas,
before the step of obtaining the video to be detected, the method further comprises the following steps:
acquiring images of a plurality of preset persons;
and performing feature training based on the images of the preset personnel to form a personnel database.
7. A monitoring device for a radioactive source, comprising:
the acquisition module is used for acquiring a video to be detected;
the first detection module is used for detecting the radioactive source in the video to be detected by utilizing the radioactive source classifier;
the image extraction module is used for extracting a video image in the video to be detected;
the characteristic extraction module is used for extracting the characteristics of the video image by using the radioactive source classifier;
the characteristic matching module is used for matching the extracted characteristics with the characteristics in the radioactive source characteristic database; the radioactive source feature database is obtained by performing feature training on images of radioactive sources;
and the determining module is used for determining that the video image corresponding to the extracted features is a source image containing the radioactive source when the extracted features are matched and consistent with the features in the radioactive source feature database.
8. The apparatus according to claim 7, wherein the radiation source feature database stores feature values corresponding to features of a plurality of radiation sources;
the feature matching module comprises:
a calculation unit configured to calculate a feature value of the extracted feature;
the comparison unit is used for comparing the similarity between the calculated characteristic value and the characteristic value in the radioactive source characteristic database;
and the determining unit is used for determining that the extracted features are matched with the features in the radioactive source feature database when the similarity is within a preset range.
9. An electronic device, comprising:
a memory and a processor, the memory and the processor being communicatively coupled to each other, the memory having stored therein computer instructions, the processor executing the computer instructions to perform the method for monitoring a radiation source according to any one of claims 1 to 6.
10. A computer-readable storage medium storing computer instructions for causing a computer to perform the method for monitoring a radiation source according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810803761.9A CN110807753A (en) | 2018-07-20 | 2018-07-20 | Radioactive source monitoring method and device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810803761.9A CN110807753A (en) | 2018-07-20 | 2018-07-20 | Radioactive source monitoring method and device and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110807753A true CN110807753A (en) | 2020-02-18 |
Family
ID=69486752
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810803761.9A Pending CN110807753A (en) | 2018-07-20 | 2018-07-20 | Radioactive source monitoring method and device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110807753A (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH1115945A (en) * | 1997-06-19 | 1999-01-22 | N T T Data:Kk | Device and method for processing picture and system and method for detecting dangerous substance |
| CN103902960A (en) * | 2012-12-28 | 2014-07-02 | 北京计算机技术及应用研究所 | Real-time face recognition system and method thereof |
| CN106203264A (en) * | 2016-06-28 | 2016-12-07 | 深圳市元征科技股份有限公司 | Dangerous goods detection method and device |
| CN108108744A (en) * | 2016-11-25 | 2018-06-01 | 同方威视技术股份有限公司 | For the method and its system of radiation image assistant analysis |
-
2018
- 2018-07-20 CN CN201810803761.9A patent/CN110807753A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH1115945A (en) * | 1997-06-19 | 1999-01-22 | N T T Data:Kk | Device and method for processing picture and system and method for detecting dangerous substance |
| CN103902960A (en) * | 2012-12-28 | 2014-07-02 | 北京计算机技术及应用研究所 | Real-time face recognition system and method thereof |
| CN106203264A (en) * | 2016-06-28 | 2016-12-07 | 深圳市元征科技股份有限公司 | Dangerous goods detection method and device |
| CN108108744A (en) * | 2016-11-25 | 2018-06-01 | 同方威视技术股份有限公司 | For the method and its system of radiation image assistant analysis |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| He et al. | Application of deep learning in integrated pest management: A real-time system for detection and diagnosis of oilseed rape pests | |
| Cao et al. | An attention enhanced bidirectional LSTM for early forest fire smoke recognition | |
| Subetha et al. | A survey on human activity recognition from videos | |
| CN108470332B (en) | Multi-target tracking method and device | |
| WO2021047232A1 (en) | Interaction behavior recognition method, apparatus, computer device, and storage medium | |
| CN111709311B (en) | Pedestrian re-identification method based on multi-scale convolution feature fusion | |
| CN110751022A (en) | Urban pet activity track monitoring method based on image recognition and related equipment | |
| CN111079658B (en) | Multi-target continuous behavior analysis method, system and device based on video | |
| US11048917B2 (en) | Method, electronic device, and computer readable medium for image identification | |
| Niu et al. | DefectGAN: Weakly-supervised defect detection using generative adversarial network | |
| Xiao et al. | Apple ripeness identification using deep learning | |
| CN110688955A (en) | Building construction target detection method based on YOLO neural network | |
| CN111210412B (en) | Packaging detection method and device, electronic equipment and storage medium | |
| Patil et al. | An approach of understanding human activity recognition and detection for video surveillance using HOG descriptor and SVM classifier | |
| CN108364303A (en) | A kind of video camera intelligent-tracking method with secret protection | |
| CN108875500A (en) | Pedestrian recognition methods, device, system and storage medium again | |
| Abbass et al. | Violence detection enhancement by involving convolutional block attention modules into various deep learning architectures: comprehensive case study for ubi-fights dataset | |
| CN109492702B (en) | Pedestrian re-identification method, system and device based on ranking measurement function | |
| Chala et al. | The Use of Neural Networks for the Technological Objects Recognition Tasks in Computer-Integrated Manufacturing | |
| Shf et al. | Review on deep based object detection | |
| CN110866458A (en) | Multi-user action detection and identification method and device based on three-dimensional convolutional neural network | |
| CN110807753A (en) | Radioactive source monitoring method and device and electronic equipment | |
| Görgülü et al. | Human activity recognition and temporal action localization based on depth sensor skeletal data | |
| CN109359544B (en) | Portrait retrieval method and device | |
| CN114005017A (en) | Target detection method and device, electronic equipment and 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 |