The present invention generally relates to an intruder detection system, and more particularly to an intruder detection system with a millimeter-wave radar.
A passive infrared (PIR) sensor is an electronic device that detects infrared (IR) radiation emitted by objects in its field of view. PIR sensors are widely used for motion detection in security systems for detecting intruders.
One of the challenges of using PIR sensors for security purposes is that they cannot distinguish between different sources of infrared radiation. PIR sensors measure the changes in infrared light that occur when an object moves in front of them, but they do not provide any information on who or what is moving. Therefore, both humans and animals can trigger the PIR sensors, as they both emit infrared rays. Moreover, environmental factors such as changes in temperature or the flow of hot gas can also cause false alarms in the security system, as they can affect the infrared radiation levels in the sensor's field of view.
A need has thus arisen to propose a novel scheme to overcomes drawbacks of conventional intruder detection systems, such as high false alarm rates and susceptibility to environmental interference, and to enhance the security and reliability of intruder detection systems.
In view of the foregoing, it is an object of the embodiment of the present invention to provide an intruder detection system capable of accurately determining an intrusion by a human.
According to one embodiment, an intruder detection system includes a millimeter-wave (mmWave) radar, an image capture device and a processor. The millimeter-wave (mmWave) radar detects an object and determines a property thereof. The image capture device generates a captured image representing a scene under monitoring. The processor determines whether the object is a human according to an output signal of the mmWave radar.
According to one aspect of the embodiment, the intruder detection system 100 may include a millimeter-wave radar (mmWave radar) 11 configured to detect a (moving) object 10 and determine properties (e.g., size and position) thereof. Specifically, mmWave radar 11 is a radar operating in millimeter-wave band of 30-300 GHz (or wavelength of 1-10 mm), which is between the centimeter wave and the light wave. Therefore, the mmWave radar 11 has the advantages of microwave navigation and photoelectric navigation. It is particularly noted that, as the mmWave radar 11 operates in the millimeter-wave band of 30-300 GHz, the mmWave radar 11 is not affected by environmental interferences such as temperature and heat, as compared to conventional sensors, thereby substantially reducing false detections caused by environmental interferences.
Generally speaking, the mmWave radar 11 may include a transmitter 111 for producing electromagnetic waves in the millimeter-wave band, a (transmitting and receiving) antenna 112 for transmitting the electromagnetic waves produced by the transmitter 111, and a receiver 113 for determining properties (e.g., size and location (or distance)) of the object 10 according to reflected electromagnetic waves reflected from the object 10 (and received by the antenna 112).
The intruder detection system 100 of the embodiment may include an image capture device 12, such as an image sensor, configured to convert light waves into a captured image representing a scene (or field of view) under monitoring including the object 10. In one embodiment, the image captured device 12 is battery powered, and is adaptable to a smart home.
In the embodiment, the mmWave radar 11 may activate the image capture device 12 to generate the captured image when the object 10 is detected by the mmWave radar 11. In other words, the image capture device 12 is normally turned off, and is turned on (i.e., activated) only when the object 10 is detected by the mm Wave radar 11.
In one specific embodiment, the image capture device 12 is activated only when the object 10 enters a predetermined range (within the scene or the field of view). Therefore, a clear captured image may be obtained (due to match between the focal length of the image capture device 12 and the distance between the object 10 and the image capture device 12) and be utilized for subsequent image processing, thereby reducing power consumption (of the image capture device 12), increasing efficiency and improving the quality of the captured image. Furthermore, this can prevent unnecessary or unwanted images from being captured and processed, thereby enhancing the privacy and security of the intruder detection system 100.
The intruder detection system 100 of the embodiment may include an (image) processor 13 configured to determine whether the object 10 is a human according to an output signal of the mmWave radar 11. Further, the processor 13 may perform a subsequent image operation on the captured image (captured from the image capture device 12).
The subsequent image operation as mentioned above may include manual visual verification or image processing (such as enhancement, recognition or analysis) based on artificial intelligent (AI) filtering. AI-based image filtering or processing refers to a technique that uses AI algorithms to filter images to facilitate removing noise from images, enhancing image quality and performing other image processing tasks. AI filtering can be used to improve the accuracy of image recognition algorithms by removing noise and other artifacts from images.
In one embodiment, the intruder detection system 100 may optionally include a motion sensor 14, such as passive infrared (PIR) motion sensor, configured to detect motion of the object 10. The PIR motion sensor refers to a sensor that detects motion by sensing changes in the infrared (i.e., radiant heat) levels emitted by the object 10. The PIR motion sensor is passive because it does not emit any energy but detect the energy emitted by the object 10.
According to a further aspect of the embodiment, the motion sensor 14 activates the mm Wave radar 11 when motion of the object 10 is detected by the motion sensor 14. In other words, the mmWave radar 11 is normally turned off, and is turned on only when the motion sensor 14 has detected motion of the object 10. It is particularly noted that although the mmWave radar 11 generally consumes more power than the motion sensor 14 (but detects the object 10 more accurately than the motion sensor 14), the mmWave radar 11 in companion with the motion sensor 14 can substantially reduce power consumption and increase efficiency and performance.
In step 23, when the object 10 enters a predetermined range, the mmWave radar 11 activates the image capture device 12 to generate a captured image (step 24); otherwise, the flow goes back to step 22. Therefore, a clear captured image may be obtained (due to match between the focal length of the image capture device 12 and the distance between the object 10 and the image capture device 12) and be utilized for subsequent image processing, thereby reducing power consumption (of the image capture device 12), increasing efficiency and improving the quality of the captured image. Furthermore, this can prevent unnecessary or unwanted images from being captured and processed, thereby enhancing the privacy and security of the intruder detection method 200.
In step 25, the processor 13 determines whether the object 10 is a human. In one embodiment, the object 10 is initially determined as a human when a size as determined by the mm Wave radar 11 is greater than a predetermined threshold. Otherwise, the object 10 is determined as a nonhuman object such as an animal, and the flow goes back to step 21. Therefore, false detections can be substantially reduced.
After a human is initially determined (step 25), the processor 13 may further perform image processing (such as enhancement, recognition or analysis) on the captured image (captured by the image capture device 12), for example, based on artificial intelligent (AI) filtering (step 26), thereby facilitate removing noise from images, enhancing image quality and performing other image processing tasks. Accordingly, an intrusion by a human may be further confirmed, and subsequent alarm and personnel dispatching may be executed. In one embodiment, the image processing performed to confirm a human intruder may include, for example, analyzing shape and size of the object 10, analyzing color and texture of the object 10, or using a pattern recognition technique to identify specific features of the object 10.
Although specific embodiments have been illustrated and described, it will be appreciated by those skilled in the art that various modifications may be made without departing from the scope of the present invention, which is intended to be limited solely by the appended claims.