JP2018088157A - Detection recognizing system - Google Patents

Abstract

A detection / recognition system capable of detecting features included in an image and recognizing a recognition target set from the features by updating firmware for detection / recognition. provide. A detection and recognition system includes an imaging unit, a detection / recognition unit, and a server. The detection / recognition unit 21 detects features included in the image from the image acquired by the imaging unit 20 under the control of the detection / recognition firmware, recognizes the set recognition target, and sets a new detection / recognition firmware. The detection / recognition firmware can be updated. The server 3 includes machine learning means 30 that generates a detection / recognition algorithm by machine learning using the image acquired by the imaging means 20 as teacher data, and the detection / recognition firmware that generates new detection / recognition firmware from the detection / recognition algorithm. Generating means 31. [Selection] Figure 1

Description

  The present invention relates to a detection recognition system.

  Conventionally, a camera that detects / recognizes an object such as a person or an object is known (for example, see Patent Documents 1 and 2). Such a camera is used as a surveillance camera for the purpose of crime prevention, for example, and issues an alarm when an abnormality is detected by detection / recognition.

  In recent years, machine learning has been frequently used in fields such as image recognition. As a machine learning technique, for example, deep learning is known. Deep learning is a technique for learning data characteristics using a neural network having a multilayer structure, and it is known that high-accuracy image recognition is possible by using this.

JP 2012-208551 A JP 2010-160743 A

By the way, surveillance cameras, etc. are often used for a long time after installation. However, technologies for detecting / recognizing objects etc. from images are advancing day by day. The detection / recognition technology used can be outdated.
Also, the detection / recognition algorithm varies depending on the environment such as the place where it is used, the subject of photography, etc., so that the detection / recognition firmware originally provided in the camera before the installation of the camera is used. There is a possibility that sufficient detection / recognition cannot be performed with the detection / recognition algorithm used.

  The present invention has been made in view of the above circumstances, and detects the feature included in the image and updates the detection / recognition performance of the detection / recognition performance for recognizing the recognition target set from this feature. It is an object of the present invention to provide a detection recognition system that can be improved.

In order to achieve the above object, the detection recognition system of the present invention includes:
An imaging unit for imaging, a detection / recognition unit, and a server;
The detection / recognition unit includes detection / recognition firmware, detects a feature included in the image from the image acquired by the imaging unit, and recognizes a set recognition target. In addition, the detection / recognition firmware can be updated to a new detection / recognition firmware generated by the detection / recognition firmware generation unit.
The server generates machine detection means for generating a detection / recognition algorithm by machine learning using the image acquired by the image pickup means as teacher data, and generates new detection / recognition firmware of the detection / recognition means from the detection / recognition algorithm. And detecting / recognizing firmware generation means.

In the present invention, the imaging means performs imaging. Then, the detection / recognition unit detects a feature included in the image from the image obtained by the image pickup by the image pickup unit, and recognizes the set recognition target. The machine learning means of the server generates a detection / recognition algorithm by machine learning using the image acquired by the imaging means as teacher data. Then, the generated detection / recognition algorithm is converted into firmware (detection / recognition firmware) suitable for the detection / recognition means by the detection / recognition firmware generation means of the server. Then, the detection / recognition firmware of the detection / recognition unit is updated to a new detection / recognition firmware generated by the detection / recognition firmware generation unit.
Therefore, a detection / recognition algorithm that can be detected and recognized with higher accuracy is generated from the image obtained by the imaging means by machine learning, and the detection / recognition algorithm is converted into firmware suitable for the detection / recognition means. Since the detection / recognition firmware of the detection / recognition means can be updated, the detection / recognition performance can be improved.

  In the configuration of the present invention, it is preferable that the machine learning unit performs machine learning using, as teacher data, an image when the detection / recognition unit erroneously recognizes the set recognition target.

  According to such a configuration, the machine learning means learns a new recognition target so that the recognition target is not erroneously recognized with respect to an image when the recognition target set with the detection / recognition means is erroneously recognized. Since the detection / recognition algorithm can be generated and new detection / recognition firmware can be generated, the detection / recognition performance can be reliably improved.

Further, in the above configuration of the present invention,
With at least one camera,
The camera preferably includes the imaging unit and the detection / recognition unit.

  According to such a configuration, it is not necessary to provide a terminal or the like that includes detection / recognition means separately from the camera, and thus the entire system can be reduced in size. In addition, since the camera can update the detection / recognition firmware of the detection / recognition means to a new detection / recognition firmware created as a result of machine learning in the server, the detection / recognition performance of the camera is improved. be able to. Therefore, the detection / recognition performance of the camera can be easily improved even with a camera or the like after installation.

Further, in the above configuration of the present invention,
It is preferable to include a plurality of the cameras in which at least a part of an imaging range by the imaging unit and the set recognition target overlap.

  According to such a configuration, a predetermined range can be imaged with a plurality of cameras, and detection / recognition can be performed. Therefore, since the same object and the same phenomenon can be detected / recognized by a plurality of cameras, the accuracy of detection / recognition can be improved.

Further, in the above configuration of the present invention,
When the machine learning means recognizes a recognition target that is overlapped by some of the plurality of cameras, the other camera of the plurality of cameras does not recognize the recognition target that is overlapped. It is preferable to perform machine learning using the acquired images as teacher data.

  According to such a configuration, when at least one camera recognizes an overlapping recognition target, machine learning is performed using, as teacher data, an image in which another camera could not recognize the overlapping recognition target. be able to. Therefore, it is possible to perform machine learning using, as teacher data, an image that is likely to be recognized but wanted to be recognized, and the efficiency of machine learning can be increased.

  In the configuration of the present invention, it is preferable that at least one of the plurality of cameras is a camera having a different imaging unit.

  According to such a configuration, it is difficult to detect / recognize from an image picked up by a certain image pickup means, and even if a recognition target cannot be recognized by a camera including the image pickup means, It becomes possible to recognize a recognition target with a camera provided with means. As a result, it is possible to easily know that the recognition target could not be recognized, and machine learning can be performed using the image that could not be recognized as teacher data. It is possible to perform machine learning so that it can be recognized even from difficult images.

  According to the present invention, it is possible to improve the detection / recognition performance of detecting a feature included in an image and recognizing a recognition target set from the feature by updating the firmware for detection / recognition.

1, showing an embodiment of the present invention, is a block diagram showing a detection recognition system. FIG. It is a block diagram which shows the camera of a detection recognition system. It is a block diagram which shows the server of a detection recognition system. 4 is a flowchart for explaining a detection recognition firmware update method by the detection recognition system.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
The detection and recognition system according to the present embodiment is used to notify when a recognition target set from an image captured by a camera is recognized, for example.
In the following description, an image simply includes both a moving image and a still image.

  As shown in FIG. 1, the detection recognition system 1 includes a plurality of cameras 2, a server 3, and a terminal 4. The plurality of cameras 2, the server 3, and the terminal 4 are connected by a wired or wireless network 5.

  For example, when the camera 2 is installed in a building such as a convenience store or outdoors as a surveillance camera, and the object recognized by the camera 2 is recognized as a suspicious person due to its external shape and movement, for example, In addition, it can be used to notify the terminal 4 at another location. In addition, when a suspicious person is recognized in this way, a terminal or a system management device held by the administrator 6 of the detection recognition system 1 may be notified.

As shown in FIG. 2, the camera 2 includes an imaging unit 20, a detection / recognition unit 21, a recording unit 22, a communication unit 23, and a control unit 24.
The imaging means 20 has a lens and a solid-state image sensor, for example, and acquires an image by imaging. The detection / recognition means 21 includes an arithmetic processing unit and a memory, and performs image recognition. Specifically, a feature included in an image captured by the imaging unit 20 is detected by control by the detection / recognition firmware provided in the memory of the detection / recognition unit 21, and a recognition target set from this feature is recognized. To do. In the following description, in the case of simple detection / recognition, basically, the feature included in the image picked up by the image pickup means 20 is detected as described above, and the recognition target set from this feature is recognized. Say.

  The recording unit 22 also includes a reference image and other information for detection / recognition by the detection / recognition unit 21 and an image at the time of abnormality (for example, when the recognition target set by the detection / recognition unit 21 is recognized). Record other information (for example, voice, etc.). The communication unit 23 communicates with the server 3 via the network 5 to transmit images and other information at the time of abnormality to the server 3 and to receive commands and detection recognition firmware from the server 3. The communication means 23 is also connected to the terminals 4 and the terminals of the administrator 6 via the network 5 and transmits an alarm signal or the like to these terminals or the server 3 when there is an abnormality. The terminal 4 or the terminal of the manager 6 receives this alarm signal, or issues an instruction to sound an alarm from the server 3 that has received the alarm signal, etc.

  The control unit 24 includes an arithmetic processing unit and a memory, and controls the imaging unit 20, the detection / recognition unit 21, the recording unit 22, and the communication unit 23. The control unit 24 may share the arithmetic processing unit or the memory with the detection / recognition unit 21.

  Note that the imaging unit 20, the detection / recognition unit 21, the recording unit 22, the communication unit 23, and the control unit 24 may not all be provided in the camera 2. For example, the detection recognition system 1 includes a terminal connected to the camera 2 by wire or wirelessly and capable of controlling the camera 2 and displaying an image taken by the camera 2. The detection / recognition means 21, the recording means 22, the communication means 23, and the control means 24 are arranged in the terminal, and the terminal detects and recognizes an image taken by the imaging means 20 provided in the camera 2. You may do it.

The camera 2 has, for example, the same configuration as a general surveillance camera. For example, the imaging unit 20 captures an imaging range corresponding to the set angle of view according to the orientation of the camera 2. As the plurality of cameras 2 included in the detection recognition system 1, the same type of camera may be used, or different types of cameras may be used. Further, the imaging ranges of the respective cameras 2 may overlap or may be completely different.
In the present embodiment, a total of four different types of cameras 2 including two stereo cameras 2a, one infrared camera 2b, and one monocular camera 2c are used as the cameras 2. It is assumed that the imaging ranges of the four cameras 2 overlap each other.
By using the stereo camera 2a that can calculate the distance, size, 3D structure, etc. from the parallax as the camera 2, since the distance, size, 3D structure, etc. can be calculated from the parallax, an arithmetic processing unit for detection / recognition, etc. Therefore, even if the camera does not include a high-performance arithmetic processing unit or the like, detection / recognition can be easily performed.

  Further, by using an infrared camera (near infrared camera or far infrared camera) 2b as the camera 2, a near infrared or far infrared image can be taken, and what cannot be seen by human eyes is also detected. / Can be recognized. Also, detection / recognition in a dark environment such as at night becomes easy.

  Moreover, the kind of camera 2 is not restricted to these. For example, a distance image sensor may be used as the camera 2. As the distance image sensor, for example, TOF (Time Of Flight) can be used. The TOF measures the distance from the time taken for the projected laser to reciprocate to the target.

  That is, the camera 2 may be one in which the imaging unit 20 captures one two-dimensional image and performs detection / recognition from this image, and the imaging unit 20 captures two images, and from the parallax of these images, The distance, size, 3D structure, etc. may be calculated and detected / recognized, or the imaging means 20 may capture a 3D distance image with a TOF sensor or the like, and detect / recognize from this 3D distance image. The imaging means 20 may pick up near-infrared or far-infrared images and perform detection / recognition from these images. One camera 2 may include a plurality of the imaging means 20 described above. That is, a single camera 2 may include, for example, an imaging function of a stereo camera and an infrared camera, and detection / recognition may be performed from an image obtained by these functions.

The detection / recognition means 21 recognizes a set recognition target, and the recognition target may be a specific object (including a person and a thing other than a person) or an abstract phenomenon. Conceivable. In other words, the recognition target may be an object such as a person such as a robber, a thief, or a arson, or an object such as a handgun, or a phenomenon such as a crime or fire.
For example, when a robbery is set as a recognition target, when an image of a person holding a knife or a handgun is taken by the imaging means 20 of the camera 2 installed in a convenience store, the detection / recognition means It is conceivable that the person 21 recognizes this person as a burglar by detecting a person holding a knife or a handgun or detecting the movement of the person from this image. In addition, for example, when a fire is set as a recognition target, it is detected that the temperature of a certain place is abnormally high from an image obtained by an infrared camera and recognizes that a fire has occurred. Can be considered. For example, if the infrared camera uses far-infrared rays, the temperature can be detected, and the handgun, knife, etc. hidden in the pocket of clothes due to the temperature difference between the handgun, knife and other weapons and body temperature. It is also conceivable to detect and recognize the weapons of this. However, since the detection / recognition firmware of the detection / recognition means 21 is generated by machine learning in the machine learning means 30 to be described later, the detection / recognition means 21 is actually easy for such a person to understand ( It doesn't always make a way of recognition.
That is, the detection / recognition unit 21 detects a feature included in the image captured by the imaging unit 20 under the control of the detection / recognition firmware, and recognizes a recognition target set from the feature.
The detection / recognition means 21 may perform detection / recognition using not only images but also sound. For example, the camera 2 includes a voice input unit such as a microphone, and the detection / recognition accuracy can be improved by performing detection / recognition using the voice acquired by the voice input unit. Similarly, voice may be used in detection / recognition by the server-side detection / recognition means 32 described later.

  The detection / recognition firmware of the detection / recognition means 21 is updated by new detection / recognition firmware generated by the machine learning means 30 and the detection / recognition firmware generation means 31 described later. First, the detection / recognition firmware provided in the detection / recognition unit 21 may be generated by the machine learning unit 30 and the detection / recognition firmware generation unit 31, and is detected by another machine-learning device. / It may be incorporated in the recognition means 21. Alternatively, the detection / recognition means 21 may be first provided with detection / recognition firmware generated by a method other than machine learning.

Further, it is assumed that the setting of the target recognized by the detection / recognition means 21 is included in the detection / recognition firmware. For example, in the case where the detection / recognition firmware is generated by the machine learning means 30 and the detection / recognition firmware generation means 31, if the object to be recognized is a burglar at a convenience store, the machine learning teacher data is, for example, a convenience store The machine learning means 30 is provided with a plurality of images showing robbers who have been robbed in the past and information that these images are images showing burglars as teacher data (tag the image as a burglar) . Then, by machine learning, it is learned where a given image (teacher data) can be recognized to recognize a burglar. As a result of machine learning, a detection / recognition algorithm with a high probability of being able to recognize a burglar from an image is generated. The detection / recognition algorithm is converted by the detection / recognition firmware generation means 31 to generate detection / recognition firmware. That is, the detection / recognition firmware (detection / recognition algorithm) obtained by this learning can recognize where in the image the robber is included, and can be recognized as an object to be recognized. That is, it can be said that robbery is set. Note that tagging an image is not always necessary when performing this machine learning. For example, if only the image showing the burglar is given as teacher data, it recognizes an image whose characteristics are similar to the image given as the teacher data even if there is no information that it is an image showing the burglar It is possible to generate an algorithm for recognizing a burglar by generating an algorithm.
Note that the number of recognition targets (targets recognized by the detection / recognition firmware) set in the detection / recognition firmware is not limited to one, and a plurality of recognition targets may be set.

  As described above, the detection / recognition firmware recognizes a specific target. When the detection / recognition unit 21 recognizes the specific target by the detection / recognition firmware, a signal indicating that the detection / recognition unit 21 recognizes the specific target, and the like. (For example, an alarm signal) is output. A signal indicating that the recognition has been performed is sent to the server 3, the terminal 4, the terminal of the administrator 6, and the like via the communication unit 23, and notification that the setting target has been recognized is made to these terminals. The The signal indicating that the recognition has been performed is sent only to the server 3, and the server 3 recognizes the recognition target from the server 3 to the terminal 4 after comprehensively judging information from each camera 2. Alarm information such as an e-mail or an instruction to sound an alarm may be sent.

  Also, the four cameras 2 have overlapping imaging ranges, and the four cameras 2 can simultaneously recognize overlapping portions of the recognition targets set in the detection / recognition firmware. ing. That is, for example, when a burglar is set as an overlapping recognition target, it is possible to simultaneously recognize a specific burglar who performs a specific burglary with four cameras.

  As shown in FIG. 3, the server 3 includes machine learning means 30, detection / recognition firmware generation means 31, server-side detection / recognition means 32, server-side recording means 33, server-side communication means 34, server Side control means 35. The machine learning means 30, the detection / recognition firmware generation means 31, the server-side detection / recognition means 32, and the server-side control means 35 each have an arithmetic processing device and a memory, but each has an individual arithmetic processing device or memory. You may have and you may share an arithmetic processing unit or memory.

  For example, the machine learning means 30 performs machine learning such as deep learning to generate a detection / recognition algorithm. Here, the detection / recognition algorithm is an algorithm for recognizing a set recognition target from an image captured by the imaging unit 20 of the camera 2.

  The detection / recognition firmware generation unit 31 converts the detection / recognition algorithm generated by the machine learning unit 30 into firmware that can be executed by each camera 2 to generate detection / recognition firmware. Each camera 2 has a resolution of an image that can be acquired by the imaging means 20, the performance of the arithmetic processing unit of the detection / recognition means 21, the presence / absence of a GPU (Graphics Processing Unit) for the detection / recognition means 21, and an audio input means such as a microphone. And the type of camera (stereo camera, TOF sensor, etc.) are different, and the firmware that can be executed by each camera 2 is also different. The detection / recognition firmware generation means 31 converts a detection / recognition algorithm generated by machine learning into firmware that can be executed by each camera 2, so that a new detection / recognition program can be installed in each camera 2. It becomes.

  The server-side detection / recognition means 32 performs detection / recognition by comprehensively judging the situation from the images and information of each camera 2. For example, the detection / recognition means 21 of each camera 2 performs detection / recognition using an image acquired by the imaging means 20 of the camera 2, but the server-side detection / recognition means 32 is acquired by a plurality of cameras 2. Detection / recognition is performed using the selected image. Further, when the processing is heavy for each camera 2, the server side detection / recognition means 32 may perform a part of the processing. Further, the detection / recognition firmware of the server side detection / recognition means 32 is provided in the memory of the server side detection / recognition means 32. Further, the detection / recognition firmware of the server-side detection / recognition means 32 can also be updated by the detection / recognition firmware generated by the machine learning means 30 and the detection / recognition firmware generation means 31.

  Further, the server side detection / recognition means 32 determines whether the recognition target of each camera 2 is correct from the recognition results of the detection / recognition means 21 of the four cameras 2 (cameras 2a, 2b, 2c), or The probability or the like that the recognition target of each camera 2 is recognized may be determined. Then, from this determination result, alarm information or the like may be sent to the terminal 4 or the like. For example, when there is a notification that all four cameras have recognized the setting target (for example, robbery), the server-side detection / recognition means 32 determines that the setting target is recognized correctly and notifies the terminal 4 or the like of an alarm. May be ordered to sound. Further, the content of the alarm information may be changed depending on the number of cameras that have recognized the setting target. For example, if all four cameras recognize the setting target, the server side detection / recognition means 32 determines that the recognition is correct and instructs the terminal 4 to sound a loud alarm sound. When only the following cameras are recognized, the server side detection / recognition means 32 determines that the recognition may be correct, and instructs the terminal 4 to sound a small alarm sound. It may be.

Further, the server side detection / recognition means 32 determines misrecognition (recognition error) of the camera 2 from the recognition results of the detection / recognition means 21 of the plurality of cameras 2. For example, when there is a notification that the setting target is recognized by the detection / recognition means 21 from the three cameras 2 out of the four cameras 2 and there is no notification from the one camera 2, this one The camera 2 determines that a recognition error (recognition error) has occurred. Conversely, out of the four cameras 2, three cameras 2 did not receive a notification that the setting target was recognized by the detection / recognition means 21, but one camera 2 received a notification. In addition, it may be determined that the one camera 2 has misrecognized (recognition error).
Even if the result of detection / recognition by the server-side detection / recognition means 32 is compared with the result of detection / recognition by each camera 2, the erroneous recognition (recognition error) of each camera 2 is determined. Good.

  The server-side recording unit 33 records teacher data for machine learning performed by the machine learning unit 30. The server side communication means 34 communicates with each camera 2 via the network 5 to receive images and other information from each camera 2, send commands and detection / recognition firmware to each camera 2, The alarm information (when the setting target is recognized) is transmitted to the terminal 4 or the administrator 6.

Next, a method for updating the detection recognition firmware of the detection recognition system 1 will be described with reference to the flowchart of FIG.
The camera 2 acquires an image by the imaging unit 20 and performs recognition (detection / recognition) of a recognition target set by control by the detection / recognition firmware of the detection / recognition unit 21. When the recognition is wrong, an image when the recognition is wrong is transmitted to the server 3 (step S1). Note that audio data or the like when the recognition is wrong may be transmitted simultaneously with the image when the recognition is wrong.
The server side detection / recognition means 32 determines whether or not the recognition is wrong from the recognition results of the plurality of cameras 2 as described above. For example, when the camera 2 recognizes a set recognition target (for example, a burglar), in the system that notifies the server 3 of the recognition (for example, the burglar is recognized), the camera 2a and the camera 2b Is notified to the server 3, but if there is no notification from the camera 2c, the server-side detection / recognition means 32 recognizes that the camera 2c has erroneously recognized from these notification results ( Judgment was not possible. At this time, the control means of the server 3 gives the camera 2c the same time as or the time before and after the time when the camera 2a and the camera 2b recognized the recognition target (for example, several seconds to several minutes before and after). The camera 2c is instructed to transmit the image acquired by the camera 2c to the server 3 as an image when the recognition is wrong. Upon receiving this command, the camera 2c transmits an image when the recognition is wrong to the server 3.
A person may determine whether or not the recognition is wrong. For example, the detection and recognition system 1 includes a display unit that displays an image captured by the camera 2 and a terminal that includes an input unit such as a pointing device or a keyboard. The camera 2 can recognize a burglar. If there is not, the person confirms the image taken by the camera 2 from the display means of this terminal, the person selects the image that he wanted to recognize the burglar using the input means of this terminal, and recognizes It is good also as transmitting to the server 3 as an image at the time of a mistake.

The server-side control unit 35 records an image sent from the camera 2 when the recognition is wrong in the server-side recording unit 33 as teacher data (education data). Further, along with the recording of the image when the recognition is wrong, the recognition result desired to be output to the detection / recognition means 21 (for example, the fact that the robber was desired to be recognized from the image) is stored in the server side recording means 33 as teacher data. Record.
The recognition result desired to be output to the detection / recognition means 21 recorded as the teacher data may be created by the server 3 or sent from the camera 2. For example, when the server-side detection / recognition means 32 determines whether or not the recognition is wrong from the recognition results of the plurality of cameras 2, the server-side detection / recognition means 32 recognizes the recognition result (detection / recognition means) that would be correct. 21) may be created as teacher data, and this teacher data may be recorded in the server-side recording means 33. In addition, for example, when checking an image taken by a person with the camera 2 and determining whether or not the recognition is wrong, when selecting an image that the person wanted to recognize a burglar from the above-mentioned terminal, The fact that he / she wanted to recognize the burglar (the fact that the image indicates a burglar) is also input using the input means of this terminal, and is sent to the server 3 together with the image when the recognition is wrong. The server-side control means 35 may record the data in the server-side recording means 33 as teacher data.

  The machine learning means 30 reads the teacher data recorded in the server side recording means 33 (step S2). Then, the machine learning means 30 extracts feature points by convolution calculation from the image when the recognition included in the read teacher data is erroneous (step S3). The machine learning means 30 performs machine learning from the information of the extracted feature points and the recognition result that the detection / recognition means 21 wants to output (step S4). As a result of machine learning, a detection / recognition algorithm that is a neural network that performs detection and recognition processing is generated (step S5).

  The machine learning in the machine learning means 30 is performed so that the detection / recognition algorithm (detection / recognition firmware) is optimized for each camera 2. Each camera 2 may have a different type of camera, or even a camera with exactly the same characteristics, but the installation location and environment used may differ. This is because it may come. Based on the original detection / recognition algorithm and the teacher data, the machine learning means 30 recognizes the detection / recognition means 21 included in the teacher data from the image when the recognition included in the teacher data is wrong. Generate a new detection / recognition algorithm that can produce a result. It should be noted that the original detection / recognition algorithm used for machine learning may be recorded in the server-side recording means 33. The detection / recognition firmware is transmitted from the camera 2, and the detection / recognition firmware is detected / recognized. It may be converted into an algorithm and used. In other words, the machine learning means 30 generates a new detection / recognition algorithm from the detection / recognition algorithm used in the detection / recognition firmware of the camera 2 with erroneous detection / recognition and the teacher data.

The detection / recognition firmware generation means 31 converts the detection / recognition algorithm generated by the machine learning means 30 into detection / recognition firmware that is detection / recognition software for each camera (step S6). That is, the detection / recognition algorithm is converted by the detection / recognition firmware generation unit 31 into software in a format that can be executed by each camera.
The server-side communication unit 34 transmits detection / recognition firmware, which is detection / recognition software generated by the detection / recognition firmware generation unit 31, to the camera 2 (step S7). When the camera 2 receives the detection / recognition firmware, the control unit 24 of the camera 2 updates the firmware of the detection / recognition unit 21 to the new detection / recognition firmware.

According to the detection recognition system of this embodiment, the detection / recognition firmware of the detection / recognition means 21 of the camera 2 is the new detection / recognition firmware generated by the machine learning means 30 and the detection / recognition firmware generation means 31 of the server 3. Can be updated.
The machine learning by the machine learning means 30 is performed using, as teacher data, an image when the detection / recognition means 21 of the camera 2 erroneously recognizes the set recognition target. In machine learning using this teacher data, The detection / recognition algorithm is improved so that the recognition target set for the image is not erroneously recognized. Therefore, the detection / recognition performance of the camera 2 can be improved.

  In addition, since machine learning is performed by the server 3 and the camera 2 only needs to execute the detection recognition firmware generated by the server 3, the detection recognition firmware is updated even if the calculation capability of the camera 2 is not so high. Thus, highly accurate detection / recognition can be performed. In addition, the camera does not have a relatively low performance as compared with other cameras over the years, but rather the performance can be gradually improved with use. In addition, the performance of the camera 2 can be improved so that detection / recognition suitable for the environment in which the camera 2 is used can be performed.

  In addition, since the machine learns by itself, it is possible to recognize the set recognition target even when the person cannot notice it. For example, as educational data, instead of giving an image when the burglar actually robbed, giving an image of the robber before the robber was actually taken. It is not an algorithm for recognizing a burglar when a burglar is actually being carried out. It is also possible to generate a detection / recognition algorithm that finds the characteristics and recognizes such a person as a burglar (a person who is likely to be a burglar). Note that since the machine learning means 30 determines which features are actually focused and recognized, it does not always recognize a person who is likely to be a robber from a behavior.

Further, according to the detection recognition system of the present embodiment, since the four cameras 2 have the same imaging range, the overlap between the recognition targets set in the detection / recognition firmware is 4 Can be recognized simultaneously by two cameras 2. Therefore, even if several of the four cameras 2 cannot recognize the recognition target, the other cameras of the four cameras 2 can recognize the recognition target. The possibility of being able to be recognized / recognized can be increased, and the accuracy of detection / recognition as a whole system can be increased.
Further, the four cameras 2 include cameras of different types of imaging means 20, such as a stereo camera 2a, an infrared camera 2b, and a monocular camera 2c. Therefore, for example, even when it is difficult to detect / recognize with the stereo camera 2a, the entire system can be compared with the case where the same type of camera 2 is used because the infrared camera 2b can detect / recognize. The accuracy of detection / recognition can be improved.
The plurality of cameras 2 may be installed in places where the imaging ranges are completely different from each other, or may recognize a completely different recognition target.

Further, the server-side detection / recognition means 32 determines whether the recognition of the recognition target in each camera 2 is correct from the recognition results of the detection / recognition means 21 of the four cameras 2 or recognizes the recognition target in each camera 2 It is possible to determine the probability of correct recognition of the camera 2 or to determine whether the camera 2 is erroneously recognized (recognition error). Therefore, the alarm sound can be emitted from the terminal 4 or the like only when the server-side detection / recognition means 32 determines that the recognition is correct from the detection / recognition results of the individual cameras 2.
Further, it is possible to automatically determine the misrecognition of the camera 2 and automatically perform machine learning so as to improve the detection / recognition ability of the misrecognized camera 2. In addition, the image used as the machine learning teacher data at this time can be the image that was misrecognized, so learning is performed so as not to misrecognize the image that has been misrecognized. It becomes possible. Therefore, it is possible to automatically determine misrecognition and to improve the accuracy of detection / recognition as the camera 2 is used.

Note that the timing at which machine learning is performed may be adjusted as necessary. For example, the teacher data may be stored in the recording unit 22 or the server-side recording unit 33, and machine learning may be performed when a certain number or more have been accumulated, or when a certain period has elapsed.
The machine learning may be performed using an image other than the image captured by the imaging unit 20. When the number and quality of the teacher data is not sufficient with only the image captured by the imaging unit 20, the machine learning effect can be improved by providing another image to the machine learning unit 30.

  Note that the recognition target recognized by the camera 2 is not limited to that described above, and any recognition target can be used as long as it can be detected / recognized from the image captured by the imaging unit 20.

1 Detection recognition system 2 Camera 2a Stereo camera (camera)
2b Infrared camera (camera)
2c Monocular camera (camera)
3 Server 20 Imaging means 21 Detection / recognition means 30 Machine learning means 31 Recognition firmware generation means

Claims (6)

An imaging unit for imaging, a detection / recognition unit, and a server;
The detection / recognition unit includes detection / recognition firmware, detects a feature included in the image from the image acquired by the imaging unit, and recognizes a set recognition target. In addition, the detection / recognition firmware can be updated to a new detection / recognition firmware generated by the detection / recognition firmware generation unit.
The server generates machine detection means for generating a detection / recognition algorithm by machine learning using the image acquired by the image pickup means as teacher data, and generates new detection / recognition firmware of the detection / recognition means from the detection / recognition algorithm. And a detection / recognition firmware generation means.   The detection / recognition system according to claim 1, wherein the machine learning unit performs machine learning using, as teacher data, an image when the detection / recognition unit erroneously recognizes the set recognition target. With at least one camera,
The detection recognition system according to claim 1, wherein the camera includes the imaging unit and the detection / recognition unit.   The detection recognition system according to claim 3, further comprising a plurality of the cameras in which at least a part of an imaging range by the imaging unit and the set recognition target overlap.   When the machine learning means recognizes a recognition target that is overlapped by some of the plurality of cameras, the other camera of the plurality of cameras does not recognize the recognition target that is overlapped. The detection recognition system according to claim 4, wherein machine learning is performed using the obtained image as teacher data.   The detection recognition system according to claim 4 or 5, wherein at least one of the plurality of cameras is a camera having a different imaging unit.

Images