CN113449705A

CN113449705A - Space monitoring system and method, computer readable storage medium

Info

Publication number: CN113449705A
Application number: CN202111008335.4A
Authority: CN
Inventors: 赵耀坤; 翁庆忠
Original assignee: Guangzhou Yuexin Semiconductor Technology Co Ltd
Current assignee: Guangzhou Yuexin Semiconductor Technology Co Ltd
Priority date: 2021-08-31
Filing date: 2021-08-31
Publication date: 2021-09-28

Abstract

The invention provides a space monitoring system and a method thereof, and a computer readable storage medium, which can perform mask superposition on a picture to be identified, and further remove an area which does not need image identification in the picture (namely, an area which is easy to generate machine identification misjudgment in the picture is removed), so that when the image identification is directly performed on the picture after the mask superposition, the image identification accuracy of a specified object can be effectively improved, the monitoring efficiency is further improved, and the condition of no alarm or false alarm is avoided.

Description

Space monitoring system and method, computer readable storage medium

Technical Field

The present invention relates to the field of space monitoring technologies, and in particular, to a space monitoring system and method, and a computer-readable storage medium.

Background

With the development of computer vision technology, the space monitoring technology for moving objects such as people, animals or vehicles receives great attention, and has wide application background and economic value in the fields of human-computer interaction, video monitoring, intelligent vehicle systems, virtual reality and the like, and the monitoring principle is as follows: the method comprises the steps of collecting photos of a monitoring field through a monitoring camera, identifying specified objects (such as people or animals) in the photos, and giving an alarm according to an identification result. However, the current space monitoring system often causes a situation of false alarm or no alarm due to image recognition error.

Disclosure of Invention

The invention aims to provide a space monitoring system and method and a computer readable storage medium, which are used for solving the problem that a mirror image and other objects similar to a specified object exist in a collected image and the specified object cannot be accurately identified, so that the accuracy of image identification is effectively improved, and the condition of false alarm or no alarm in space monitoring is avoided.

To achieve the above object, the present invention provides a space monitoring system, comprising:

the system comprises a user operation layer, a monitoring area management layer and a monitoring area management layer, wherein the user operation layer is provided with a management platform and an alarm platform, the management platform is at least used for providing mask management operation required by the monitoring area for a user, and the alarm platform is used for displaying alarm information for the user;

the service layer is at least used for providing an image acquisition service, an image recognition service and an alarm sending service, the image acquisition service is used for controlling a camera in the monitoring area to acquire a video or a photo of the monitoring area so as to obtain a picture to be recognized, the image recognition service is used for firstly performing mask superposition on the picture to be recognized by using a mask set in the user operation layer so as to shield a specified area which does not need image recognition in the picture to be recognized, then performing image recognition on a specified object on the picture to be recognized after the mask superposition, and the alarm sending service is used for determining whether to generate and send an alarm according to a recognition result of the image recognition service;

and the basic component layer is provided with a picture storage module, and the picture storage module is at least used for storing the picture to be identified when the image identification service identifies that a specified object exists in the picture to be identified.

Optionally, the basic component layer further has a configuration database, and the configuration database

Optionally, the management platform is further configured to provide one of the following operations to a user: managing cameras in a monitoring area, managing monitoring time and managing alarm sending service; the management of the alarm sending service comprises at least one of management of alarm triggering conditions, management of alarm sending paths, management of alarm modes and management of alarm sending records.

Optionally, the image capturing service is configured to obtain camera configuration information in the monitoring area from the configuration database, and allocate a thread to each camera in the monitoring area, so that each camera captures a video stream or a photo in the monitoring area within the monitoring time set by the user operation layer to obtain a picture to be recognized.

Optionally, the base component layer further has a queue storage module, and the queue storage module is provided with an identified queue and an unidentified queue; the image acquisition service is also used for extracting the frames acquired by the cameras in the video stream according to a fixed time interval so as to obtain the pictures to be identified, storing the extracted pictures in the picture storage module, and storing the extracted records of the pictures in the unidentified queue of the queue storage module; the image identification service is also used for storing the picture to be identified without mask into the corresponding directory of the picture storage module and storing the corresponding picture storage record into the identified queue of the queue storage module when the specified object is identified in the picture to be identified.

Optionally, the alarm sending service is further configured to read the identified queue in a circulating manner, determine whether each picture storage record in the identified queue has sent an alarm, and generate and send an alarm corresponding to the picture storage record according to set management information of the alarm if any picture storage record has not sent an alarm.

Optionally, the service layer is further configured to provide one of the following services:

the platform management service is used for realizing the authority configuration and management of the administrator of the user operation layer and other users;

and the data cleaning service is used for realizing the timing cleaning or the custom cleaning of any data in the user operation layer and/or the basic component layer.

Based on the same inventive concept, the invention also provides a space monitoring method, which comprises the following steps:

s1, acquiring videos or photos of the monitoring area through a camera in the monitoring area to obtain a picture to be identified;

s2, performing mask superposition on the picture to be recognized to shield a designated area which does not need image recognition in the picture to be recognized;

and S3, performing image recognition of the specified object on the picture to be recognized after the mask is overlapped, and determining whether to generate and send an alarm according to the result of the image recognition.

Optionally, in step S1, the step of acquiring a video or a photo of the monitored area by a camera in the monitored area to obtain a picture to be recognized includes:

the method comprises the steps that cameras needed to be used in a monitoring area and monitoring time of each camera are configured in advance;

distributing a thread for each camera in the monitoring area according to the configuration information of the cameras in the monitoring area, and enabling each camera to collect video streams in the monitoring area within the set monitoring time;

and extracting frames in the video stream according to a fixed time interval to obtain a picture to be identified.

Optionally, in step S1, the extracted picture to be identified is cached, and the corresponding extraction record is placed in the unidentified queue; in step S2, the picture to be recognized corresponding to the cache is found according to the record in the unrecognized queue, and then the found picture to be recognized is subjected to mask overlapping to block the designated area in the picture to be recognized, which does not need image recognition.

Optionally, in step S3, when no specific object is identified in the image to be identified after the mask overlay, deleting the cached image to be identified, and deleting the extraction record corresponding to the image to be identified from the unidentified queue; when the specified object is identified from the images to be identified after the masks are overlapped, the images to be identified without the masks and with the specified object are stored, and the corresponding image storage records are added into the identified queue.

Optionally, in step S3, the identified queue is read in a loop, and it is determined whether each picture storage record in the identified queue has sent an alarm, and if no alarm has been sent by any picture storage record, an alarm corresponding to the record is generated and sent according to the management information of the alarm set in the user operation layer.

Based on the same inventive concept, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the spatial monitoring method of the present invention.

Compared with the prior art, the technical scheme provided by the invention has at least one of the following beneficial effects:

1. the system has simple structure and is easy to realize.

2. The mask can be utilized to shield the region which is not needed to be identified in the picture to be identified, so that the machine cannot identify the region which is easy to misjudge, the identification efficiency and the identification accuracy are improved, and the problem of no alarm or false alarm caused by error identification is avoided.

3. The method can be applied to any application scene needing to monitor a specified object (such as a human or an animal or a vehicle).

Drawings

Fig. 1 is a schematic diagram of a picture with mirror image collected in a known space monitoring scheme for a walkway.

Fig. 2 is a diagram illustrating a picture of an image with a human body-like shape acquired in a known spatial monitoring scheme for a walkway.

Fig. 3 is a system architecture diagram of a space monitoring system according to an embodiment of the invention.

Fig. 4 is a schematic view illustrating a usage flow of the space monitoring system according to an embodiment of the present invention.

Fig. 5 is a schematic flow chart illustrating image recognition performed in the space monitoring system according to an embodiment of the present invention.

Fig. 6 is a diagram of the corresponding picture effect in the flow shown in fig. 5.

Fig. 7 is a flowchart of a space monitoring method according to an embodiment of the invention.

Detailed Description

As described in the background, existing space monitoring systems often cause false alarms (i.e., should not alarm) or no alarms (i.e., should alarm but not alarm) due to image recognition errors.

The inventors have found that the above problems are caused by the following conditions:

(1) when the monitoring camera captures a mirror image or a reflected image of a designated object in a picture (or called an image) of a monitoring area, an image recognition device of the monitoring system recognizes the mirror image or the reflected image as the designated object when recognizing the designated object in the picture, thereby generating a recognition error. For example, in fig. 1, a human body image 101 actually walking on a walkway and a mirror image 102 on a walkway wall surface 100a exist in a picture taken by a monitoring camera in a walkway 100, and there is a problem that an erroneous alarm or no alarm is generated due to an error that the mirror image 102 is also recognized as an actual human body when an image recognition device of a monitoring system performs image recognition.

(2) When an object similar to the designated object exists in the monitored area, the image recognition device of the monitoring system recognizes the mirror image or the reflected image as the designated object when recognizing the designated object in the photograph, thereby causing a recognition error. For example, in fig. 2, a human body image 101 actually walking on a walkway and an object 103 (e.g., a human body model) similar to a human body on the walkway exist in a picture taken by a monitoring camera in the walkway 100, and there is a problem that an error occurs in recognizing the object 103 as an actual human body when an image recognition device of the monitoring system performs image recognition, thereby generating an erroneous alarm or no alarm.

Therefore, it is desirable to provide a new monitoring technology, which can eliminate the images that do not need to be identified, such as mirror reflection, reflection images, and similar object images, in the image to be identified, so as to improve the identification accuracy of the specified object and further improve the monitoring efficiency.

Referring to fig. 3, an embodiment of the invention provides a space monitoring system, which is a three-layer architecture system, including a user operation layer 20, a service layer 21 and a base component layer 22.

In this embodiment, the user operation layer 20 has a management platform 201 and an alarm platform 202.

Among other things, the management platform 201 may be used to provide one of the following operations to the user: management 201a of cameras, management 201b of monitoring time, management 201c of masks, and management 201d of alarm sending service. The management 201a of the cameras includes addition, deletion, and modification of the cameras in the monitored area, such as adjustment of the monitoring angle and the monitoring direction. The management 201b of the monitoring time includes defining a monitoring time period of each camera in the monitoring area. The management 201c of the masks includes defining the masks to block the areas of the pictures or videos shot by the respective cameras that do not need image recognition by the masks, where the areas blocked by the masks may be areas where mirror images and/or other objects similar to the designated object exist, that is, areas that do not need to be monitored are excluded from the respective cameras. The management 201d of the alert sending service includes one or more of management of alert triggering conditions, management of alert sending paths (e.g., defining a mailbox address, a telephone number, or an instant messaging app account for receiving alert information, etc.), management of alert modes (e.g., selecting a mode of notifying by mail, notifying by short message, notifying by telephone, notifying by instant messaging app information, etc.), and management of alert sending records (e.g., viewing camera monitoring records, etc.).

As can be seen from the above, the management platform 201 can achieve the purposes of performing initial configuration on the space monitoring system of the present invention and adjusting the configuration of the space monitoring system as needed.

It should be noted that, in the present embodiment, the user may be allowed to temporarily modify information such as monitoring time on the management platform 201 in a special case. For example, when the space monitoring system of this embodiment is used in a monitoring scene of a room area such as an office, a person in charge of the room usually resides in the room during the daytime working hours (9: 00 am to 5:00 pm), and therefore the monitoring time set for the room on the management platform 201 is the working hours (5: 00 pm to 9:00 am the second day) or the night sleep rest time (8: 00 evening to 8:00 am the second day), which are time periods when the person in charge usually leaves the room, so as to perform security monitoring on the room.

The alert platform 202 is capable of receiving alert information and presenting the alert information to a user, and may further provide the user with operations to query and retrieve relevant information in the base component layer 22. The alert platform 202 may include one or more of a mailbox, a cell phone, an instant messaging app, etc., the specific form of which depends on the alert mode and the alert path that are allowed to be set by the management 201a of the alert sending service on the management platform 201.

In this embodiment, the service layer 21 can provide an image capture service 211, an image recognition service 212, an alarm sending service 213, a platform management service 214, and a data cleaning service 215.

The image capturing service 211 is configured to control, according to parameters configured by the management 201a of the cameras on the management platform 201 of the user operation layer 21 and the management 201b of the monitoring time, the corresponding cameras in the monitoring area to capture videos or photos in the monitoring area within a set monitoring time period, so as to obtain a picture to be identified.

The image recognition service 212 is configured to first use the mask set in the management 201c of the mask of the user operation layer 20, perform mask overlapping on the corresponding to-be-recognized picture to block a specified area that does not need image recognition in the to-be-recognized picture, and then perform image recognition on a specified object (for example, a human body) on the to-be-recognized picture after the mask overlapping.

The alert transmission service 213 serves to determine whether to generate and transmit an alert according to the recognition result of the image recognition service 212. As an example, the warning information may be sent to the relevant person in the form of a mail, such as sending a mail to warn the responsible person in the monitoring area that the specified object is present in the monitoring area during the current monitoring time period.

The platform management service 214 is used to implement the configuration and management of permissions for administrators and other users of the user operations layer 20.

The data cleansing service 215 is used to implement timed or custom cleansing of any data in the user operations layer 20 and/or the base component layer 22, such as timed cleansing of expired alarm records in the user operations layer 20.

In this embodiment, the basic component layer 22 includes a configuration database 221, a queue storage module 222, and a picture storage module 223.

The configuration database 221 is used for storing various configuration information of the user operation layer 21 to the system, and/or storing a record of the alarm information sent by the alarm sending service 213, and the like.

The picture storage module 223 is at least used for storing some pictures to be recognized when the image recognition service 212 recognizes that there is a specific object in the pictures to be recognized.

The queue storage module 222 is provided with an identified queue and an unidentified queue, the image acquisition service 211 extracts frames acquired by each camera in the video stream at fixed time intervals to obtain a plurality of pictures to be identified, stores the extracted plurality of pictures to be identified in the picture storage module 223, and stores the extracted records of the pictures in the unidentified queue of the queue storage module 222. When a specific object is identified from some of the images to be identified after mask superposition, the image identification service 212 stores the images to be identified with the specific object and without mask into the image storage module 223, where the images to be identified with the specific object and without mask may be the images to be identified with the specific object before mask superposition, or the images to be identified without mask and marked with the specific object after image identification. The image recognition service 212 stores the pictures of the identified designated object and stores the picture storage records corresponding to the pictures in the identified queue of the queue storage module 222. The alarm sending service 213 reads the identified queue circularly, and determines whether each picture storage record in the identified queue has sent an alarm, if any picture storage record has not sent an alarm, then according to the configuration information in the management 201a of the alarm sending service of the user operation layer 20, generates the alarm information corresponding to the picture storage record, and sends the alarm information to the alarm platform 202. The picture storage record may contain a stored picture name, a storage time (including information of date, clock, etc.), a camera number, a monitoring time (including information of date, clock, etc.), a picture extraction time (including information of date, clock, etc.), a location name of a monitoring area, and the like.

As an example, the configuration database 221 may be a MySQL database, the queue storage module 222 may be a Redis message queue system, and the picture storage module 223 may be any suitable file storage system.

Optionally, the image recognition service 212 may also train or deep learn a large number of picture data sets that include or do not include the specified object to improve the accuracy of its image recognition.

As an example, a Python (e.g., Python 3.7) development tool, a JDK (e.g., JDK 1.8) development tool, an image recognition deep learning platform (e.g., paddlepaddlewheelble hundreds degree of flying oars hundreds degree open source deep learning platform, which may be paddlepaddlepaddlepaddleable 1.8) development tool, an image recognition pre-training model (e.g., paddlehub hundreds degree of flying oars pre-training model) development tool, etc. may also be deployed in the base component layer 22, so as to implement communication and information interaction between the user operation layer 20, the service layer 21, and the base component layer 22 inside the space monitoring system, and implement communication and information interaction between the space monitoring system and the external environment, and implement the image recognition service 212 to train or deep learn a large number of image datasets that include or do not include the specified object, so as to improve the accuracy of image recognition. Optionally, the image recognition service 212 may train or deep learn the pictures stored in the picture storage module 223 to improve the accuracy of its image recognition.

Referring to fig. 3 and fig. 4, the usage flow of the space monitoring system of the present embodiment is as follows:

first, the system is configured by the management platform 201, for example, including deploying cameras required for monitoring, setting a required mask to exclude an area that does not need image recognition in a picture taken by each camera, setting a monitoring period of each camera, setting an alarm receiver, and the like.

Then, the image capturing service 211 obtains configuration information (which may include, but is not limited to, address information and monitoring period setting information of each camera) for each camera on the management platform 201 of the user operation layer 21 from the base component layer 22, and allocates a thread to each camera according to the obtained configuration information of each camera so that each camera in the monitoring area can capture a video stream of the monitoring area. The image capturing service 211 further needs to determine whether the current time is within a monitoring time period configured by the user operation layer 20 for each camera, if not, the image capturing service 211 does not perform picture extraction on the video stream captured by the camera, and if so, the image capturing service 211 performs picture extraction on the frame of the video stream captured by the camera according to a fixed time to obtain a plurality of pictures to be identified. The image capturing service 211 buffers the extracted several pictures to be identified in the picture storage module 223, and adds a picture extraction record to the queue to be identified in the queue storage module 222, where the picture extraction record may include the name of the extracted picture, the extraction time (including information such as date and clock), the camera number, the monitoring time (including information such as date and clock), the location name of the monitoring area, and the like.

Then, the image recognition service 212 performs image recognition on each picture to be recognized obtained by the image capturing service 211, and when a specified object is recognized, the alarm sending service 213 issues an alarm. Referring to fig. 5 and fig. 6, the specific process is as follows: (1) and reading the picture to be identified. The corresponding extraction record buffered in the queue to be recognized of the queue storage module 222 may be read according to a queue processing rule (e.g., a first-in first-out principle or a first-in last-out principle, etc.), and according to information in the extraction record, the corresponding picture to be recognized, for example, the first picture in fig. 6, is found from the picture storage module 223. (2) And (5) overlapping masks. The mask set in the user manipulation layer 20 is superimposed on the picture to be recognized to block an area that is not to be recognized, for example, the mask 104 in fig. 6 blocks the entire wall surface 100a capable of generating the mirror image 102. (3) And (5) image recognition. And performing image recognition on the image to be recognized after the mask is superposed, and judging whether a specified object (such as a human body 101 in fig. 6) exists in the mask non-shielded area. If the specified object is not identified, the cached picture to be identified is deleted from the picture storage module 223, and the corresponding record of the picture is deleted from the queue storage module 222. If the specified object is detected, the position information of the specified object (i.e. the position information of the specified object in the original picture without mask being added) may be further obtained according to the detection result, and then the region where the specified object is located is marked in the original picture without mask being added according to the position information, as shown by a dotted frame 105 in the third picture in fig. 6, so as to obtain a new picture with an identification frame and needing an alarm, and the picture is moved to the corresponding storage directory in the picture storage module 223, or the picture to be identified without mask being superimposed (i.e. the original picture with the specified object) cached in the picture storage module 223 is directly transferred from the storage directory to another storage directory (e.g. the alarm picture directory). While storing the pictures identified as having the specified object, the associated picture storage information is also sent to the identified queue of the queue storage module 222. (4) And sending an alarm. The alarm sending service 213 circularly reads the identified queue of the queue storage module 222, and if the identified queue is not empty, takes out a record, and determines whether the alarm information has been sent in unit time according to the related information of the record (i.e., determines whether each picture storage record in the identified queue has sent an alarm), if the alarm information has been sent, does not process the record, and if the alarm information has not been sent, generates the alarm information corresponding to the record according to the configuration information in the management 201a of the alarm sending service of the user operation layer 20, and sends the alarm information to the alarm platform 202. (5) Historical data cleaning (or stale data cleaning). As an example, the alarm records before one week may be cleared once a day, for example, the alarm sending records before one week are deleted from the management 201d of the alarm sending service of the alarm platform 202, or the pictures stored in the picture storage module 223, from which the alarm information has been sent before one week, are deleted.

It should be noted that, in the space monitoring system of this embodiment, the image capturing service is allowed to capture a video stream of a period of time, and extract the video stream into a plurality of pictures to be recognized, and at the same time, the image recognizing service is allowed to recognize a designated object for each of the plurality of pictures to be recognized in the period of time, so as to synthesize recognition results of the pictures to be recognized, and finally determine whether the designated object exists in the monitoring area in the period of time, thereby further avoiding the situation of false alarm or no alarm by using the synthesized result.

Therefore, the space monitoring system of the embodiment removes the region which does not need to be subjected to image recognition (namely, the region which is easy to misjudge by machine recognition) by performing mask superposition on each picture to be recognized, so that the image recognition of the specified object can be performed only on the mask non-blocked region of the picture to be recognized, the image recognition accuracy rate is greatly improved, the space monitoring efficiency is effectively improved, and the condition of no warning or miswarning is avoided.

It is understood that the three-layer architectures of the user operation layer 20, the service layer 21 and the base component layer 22 may be combined and implemented in one device block, or any one of the three-layer architectures may be split into a plurality of devices, or at least part of the functions of one or more of the three-layer architectures may be combined with at least part of the functions of the other two-layer architectures and implemented in one device. According to an embodiment of the present invention, at least one of the three-tier architectures of the user operations layer 20, the service layer 21 and the base component layer 22 may be implemented at least partially as a hardware circuit, such as a Field Programmable Gate Array (FPGA), a Programmable Logic Array (PLA), a system on a chip, a system on a substrate, a system on a package, an Application Specific Integrated Circuit (ASIC), or may be implemented in hardware or firmware in any other reasonable manner of integrating or packaging a circuit, or in a suitable combination of software, hardware and firmware implementations. Alternatively, at least one of the three layers of the user operations layer 20, the service layer 21 and the base component layer 22 may be implemented at least partially as computer program modules, which when executed by a computer, may perform the functions of the respective module.

Based on the same inventive concept, referring to fig. 7, an embodiment of the present invention further provides a space monitoring method, including the following steps:

s11, pre-configuring cameras required to be used in the monitoring area and monitoring time of each camera;

s12, according to the configuration information of the cameras in the monitoring area, allocating a thread to each camera in the monitoring area, and enabling each camera to collect the video stream in the monitoring area within the set monitoring time;

and S13, judging whether the current time is the monitoring time, if not, not processing the current collected video stream, and if so, extracting the frames in the current collected video stream according to a fixed time interval to obtain a plurality of pictures to be identified.

The specific implementation of step S11 can refer to the above steps related to configuring the system in the usage flow of the space monitoring system, and will not be described in detail here. The steps S12-S13 can be embodied by referring to the steps of image acquisition in the usage process of the space monitoring system, and will not be described in detail herein. Steps S2-S3 refer to the steps of image recognition in the usage process of the space monitoring system, and are not described in detail herein.

Optionally, in step S1, the extracted picture to be identified is further buffered, and the corresponding extraction record is placed in the unidentified queue. In step S2, the picture to be recognized corresponding to the cache is found according to the record in the unrecognized queue, and then the found picture to be recognized is subjected to mask overlapping to block the designated area in the picture to be recognized, which does not need image recognition. In step S3, when the specified object is not identified in the image to be identified after the mask overlay, deleting the cached image to be identified, and deleting the extraction record corresponding to the image to be identified from the unidentified queue; when the specified object is identified from the images to be identified after the masks are overlapped, the images to be identified without the masks and with the specified object are stored, and the corresponding image storage records are added into the identified queue. As an example, when the specified object is identified from the image to be identified after the masks are superimposed, the position information of the specified object (i.e. the position information of the specified object in the original image without masks added) may be obtained according to the checking result, and then the area where the specified object is located is marked in the original image without masks added according to the position information, as shown by the dotted frame 105 in the third image in fig. 6, so as to obtain a new image (i.e. the image formed after the original image is marked) with the identification frame and needing an alarm, and then the marked image is stored, and the alarm information is sent to the outside. As another example, when a specific object is identified from the image to be identified after the mask is superimposed, the previously cached original image without the mask being superimposed is directly transferred from the cache directory to another storage directory (for example, an alarm image directory), and an alarm message is sent to the outside.

It should be noted that, when a plurality of pictures to be recognized are extracted from a video stream in step S1, steps S2 and S3 are actually a loop process as shown in fig. 5. The mask overlapping is performed after each picture to be recognized is read in step S2, only step S3 is then performed to perform image recognition on the image with the mask overlapped, and when the specified object is not recognized, step S2 is performed again to read the next picture to be recognized, thereby issuing an alarm until the specified object is recognized or emptying all extraction records in the queue which is not recognized.

Optionally, in step S3, the formed identified queue is further read in a loop manner, and it is determined whether an alarm has been issued from each picture storage record in the identified queue, and if an alarm has not been issued from any picture storage record, an alarm corresponding to the picture storage record is generated and issued according to the management information of the alarm set in step S1.

Based on the same inventive concept, an embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, where the computer program may include codes and computer-executable instructions, and when the computer program is executed by a processor, the computer program implements the spatial monitoring method and any variation thereof described in the present invention.

The computer-readable storage medium of the present embodiments may be any medium that can contain, store, communicate, propagate, or transport the instructions. For example, the computer readable storage medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, device, or propagation medium. Specific examples of the computer storage medium include a magnetic storage device such as a magnetic tape or a Hard Disk Drive (HDD); optical storage devices, such as Compact Discs (CDROMs); a memory, such as a Random Access Memory (RAM) or a flash memory; and/or a wired, wireless communication link.

In summary, the present invention provides a space monitoring system and method, and a computer readable storage medium, which can perform mask overlapping on a picture to be identified, and further remove an area that does not need image identification in the picture (i.e. exclude an area that is easy to generate machine identification misjudgment in the picture), so that when the image identification is directly performed on the picture after the mask overlapping, the image identification accuracy of a specified object can be effectively improved, and further the monitoring efficiency is improved, and the situation of no alarm or false alarm is avoided. The technical scheme of the invention can be applied to any suitable space monitoring scene, such as a monitoring scene needing people detection and people flow statistics, or a monitoring scene needing animal activity monitoring, or a monitoring scene needing vehicle (such as a vehicle, a ship, an airplane and the like) detection and flow statistics, and the like.

It should be noted that, although the present invention has been described with reference to the preferred embodiments, the present invention is not limited to the embodiments. It will be apparent to those skilled in the art from this disclosure that many changes and modifications can be made, or equivalents modified, in the embodiments of the invention without departing from the scope of the invention. Therefore, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention are still within the protection scope of the technical solution of the present invention, unless the content of the technical solution of the present invention is departed from.

As used herein, the term "and/or" means either or both. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.

Claims

1. A space monitoring system, comprising:

2. The space monitoring system according to claim 1, wherein the base component layer further has a configuration database for storing various configuration information of the user operation layer to the system and/or storing a record of the alarm transmission service transmitting the alarm information.

3. The space monitoring system of claim 2, wherein the management platform is further configured to provide one of the following to a user: managing cameras in a monitoring area, managing monitoring time and managing alarm sending service; the management of the alarm sending service comprises at least one of management of alarm triggering conditions, management of alarm sending paths, management of alarm modes and management of alarm sending records.

4. The space monitoring system according to claim 3, wherein the image capturing service is configured to obtain the configuration information of the cameras in the monitored area from the configuration database, and allocate a thread to each camera in the monitored area, so that each camera captures the video stream or the photo in the monitored area within the monitoring time set by the user operation layer to obtain the picture to be recognized.

5. The space monitoring system according to claim 4, wherein the base component layer further has a queue storage module, and the queue storage module is provided with an identified queue and an unidentified queue; the image acquisition service is also used for extracting the frames acquired by the cameras in the video stream according to a fixed time interval so as to obtain the pictures to be identified, storing the extracted pictures in the picture storage module, and storing the extracted records of the pictures in the unidentified queue of the queue storage module; the image identification service is also used for storing the picture to be identified without mask into the corresponding directory of the picture storage module and storing the corresponding picture storage record into the identified queue of the queue storage module when the specified object is identified in the picture to be identified.

6. The space monitoring system according to claim 5, wherein the alarm sending service is further configured to read the identified queue in a loop, determine whether each picture storage record in the identified queue has sent an alarm, and if any picture storage record has not sent an alarm, generate and send an alarm corresponding to the picture storage record according to management information of the alarm set in the user operation layer.

7. The space monitoring system according to any one of claims 1-6, wherein the service layer is further configured to provide one of:

8. A method of spatial surveillance, comprising:

9. The spatial monitoring method according to claim 8, wherein in step S1, the step of acquiring a video or a photo of the monitored area by a camera in the monitored area to obtain a picture to be recognized comprises:

10. The spatial monitoring method according to claim 9, wherein in step S1, the extracted pictures to be identified are buffered, and the corresponding extraction records are put into an unidentified queue; in step S2, the picture to be recognized corresponding to the cache is found according to the record in the unrecognized queue, and then the found picture to be recognized is subjected to mask overlapping to block the designated area in the picture to be recognized, which does not need image recognition.

11. The spatial monitoring method according to claim 10, wherein in step S3, when no specific object is identified in the mask-superimposed pictures to be identified, the cached pictures to be identified are deleted, and the extraction record corresponding to the pictures to be identified is deleted from the unrecognized queue; when the specified object is identified from the images to be identified after the masks are overlapped, the images to be identified without the masks and with the specified object are stored, and the corresponding image storage records are added into the identified queue.

12. The space monitoring method according to claim 11, wherein in step S3, the identified queue is read circularly, and it is determined whether each picture storage record in the identified queue has issued an alarm, and if any picture storage record has not issued an alarm, an alarm corresponding to the picture storage record is generated and issued according to the set management information of the alarm.

13. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the spatial monitoring method according to any one of claims 8 to 12.