Lighting control device

Abstract

To provide a lighting control device that can precisely control lighting in accordance with the movement of a moving object. [Solution] The lighting control device according to the embodiment comprises a first acquisition unit, a second acquisition unit, and a determination unit. The first acquisition unit acquires a first detection result of a moving object based on the sensing results of a predetermined area. The second acquisition unit acquires a second detection result of an object to be illuminated based on the sensing results. The determination unit determines the control content of the lighting device that illuminates the predetermined area according to the combination of the first detection result and the second detection result.

Description

Technical Field

[0001] Embodiments of the present invention relate to a lighting control device.

Background Art

[0002] Conventionally, lighting devices incorporating a human sensor that detects the movement of a person such as an infrared sensor are known. As another example, there is a lighting device that turns on the illumination upon detection of a person through an image sensor and AI analysis, and keeps the illumination on as long as the person is detected. Further, Patent Document 1 discloses a technique that performs image recognition triggered by detection from a human sensor to identify an individual and execute lighting control.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of Invention

Problems to be Solved by the Invention

[0004] However, in the conventional technology, the infrared sensor cannot detect a person when the person is stationary, and thus turns off the lighting device. In addition, AI analysis using images requires time to detect a person, so there is a risk that the lighting device cannot be turned on immediately. As described above, conventionally, there has been room for improvement in performing lighting control with high accuracy in accordance with the movement of a moving object such as a person.

[0005] The problem to be solved by the present invention is to provide a lighting control device capable of performing lighting control with high accuracy in accordance with the movement of a moving object.

Means for Solving the Problems

[0006] The lighting control device according to the embodiment comprises a first acquisition unit, a second acquisition unit, and a determination unit. The first acquisition unit acquires a first detection result of a moving object based on the sensing results of a predetermined area. The second acquisition unit acquires a second detection result of an object to be illuminated based on the sensing results. The determination unit determines the control content of the lighting device that illuminates the predetermined area according to the combination of the first detection result and the second detection result. [Brief explanation of the drawing]

[0007] [Figure 1] Figure 1 shows an example of the operation of a lighting device according to an embodiment. [Figure 2] Figure 2 is a block diagram showing an example configuration of a lighting device according to an embodiment. [Figure 3] Figure 3 is a flowchart showing the processing procedure performed in the lighting device according to the embodiment. [Figure 4] Figure 4 is a flowchart showing the processing procedure performed in the lighting device according to the embodiment. [Modes for carrying out the invention]

[0008] The lighting control device (lighting device 1) described below comprises a first acquisition unit 21, a second acquisition unit 22, and a determination unit 23. The first acquisition unit 21 acquires a first detection result of a moving object based on the sensing results of a predetermined area. The second acquisition unit 22 acquires a second detection result of an object to be illuminated based on the sensing results. The determination unit 23 determines the control content of the lighting device 1 to illuminate the predetermined area according to the combination of the first and second detection results.

[0009] The decision unit 23, described below, determines the control content to continue the lighting state when the lighting device 1 is lit, the first detection result is not detected, and the second detection result detects the target object.

[0010] The decision unit 23, described below, determines the control content to change the lighting device 1 from a lit state to an off state if the first detection result is not detected and the second detection result is not detected, while the lighting device 1 is lit.

[0011] The object detection process for obtaining the second detection result, as described below, is performed after the first detection result indicates that no object has been detected.

[0012] The object detection process for obtaining the second detection result, as described below, is performed after a predetermined period of time has passed since the first detection result showed no detection.

[0013] The decision unit 23, described below, prohibits the determination process for the control content if the position of the target object detected in the second detection result is a position where no moving object has been detected in the first detection result in the past.

[0014] The first detection result described below is the detection result of the moving object by the infrared sensor. The second detection result is the detection result of the AI ​​model that detects the target object based on the image captured by the camera.

[0015] The lighting control device described below further comprises a learning unit 24. The learning unit 24 generates training data based on a combination of the first detection result and the second detection result to train the AI ​​model.

[0016] The learning unit 24, described below, generates training data by assigning error labels to the image corresponding to the second detection result if the position of the target object detected in the second detection result is a position where no moving object has been detected in the first detection result in the past.

[0017] The learning unit 24, described below, if it detects a moving object in the first detection result and the second detection result is no detection, generates training data by assigning the correct label of the target object to the image corresponding to the second detection result.

[0018] Hereinafter, the lighting control device according to the embodiment will be described with reference to the drawings. In the embodiments, identical portions are denoted by identical reference numerals, and duplicate descriptions are omitted.

[0019] Figure 1 is a diagram illustrating an operation example of the lighting device 1 according to the embodiment. The lighting device 1 shown in Figure 1 is disposed, for example, in a relatively large facility (an example of an area) such as a factory where a large number of people move around. Note that the present disclosure illustrates a case where the lighting control device is the lighting device 1, that is, an example in which the lighting control device and the lighting device 1 are configured integrally, but the lighting control device and the lighting device 1 may be separate bodies. In this case, the lighting control device is connected to the lighting device 1 via a network of a communication standard for lighting control such as DALI (Digital Addressable Lighting Interface) (registered trademark), acquires information detected by the camera 11 and the infrared sensor 12 of the lighting device 1, and performs lighting control of the lighting device 1.

[0020] As shown in Figure 1, the lighting device 1 includes a light source unit 10, a camera 11, and an infrared sensor 12. The light source unit 10 includes, for example, a light-emitting element such as an LED (Light Emitting Diode), and irradiates light from the ceiling toward the floor surface.

[0021] The camera 11 is a camera that captures an image of a predetermined area, and is built into the lighting device 1. The camera 11 has an angle of view directed from the ceiling toward the floor surface for the predetermined area. In addition, the camera 11 has an AI model that detects an object based on the captured image, and the AI model is a learning model trained to detect a target object to be illuminated from the image. The target object to be illuminated is a moving body in a stationary state. That is, the AI model detects a moving object such as a stationary person as the target object. In other words, the AI model detects a target to be lit in lighting control as the target object. Note that the camera 11 is not limited to being built into the lighting device 1, and may be configured as a separate body from the lighting device 1.

[0022] The infrared sensor 12 is a sensor that detects a moving object (such as a person) in a predetermined area, and is incorporated in the lighting device 1. Specifically, the infrared sensor 12 detects a moving object that is in a moving state while moving through the predetermined area. In other words, when the moving object is in a stationary state, the infrared sensor 12 is in an undetected state. Note that the infrared sensor 12 is not limited to being incorporated in the lighting device 1, and may be configured as a separate body from the lighting device 1.

[0023] In the present disclosure, the lighting device 1 (lighting control device) determines the control content in lighting control based on the combination of the detection result of the moving object by the infrared sensor 12 (first detection result) and the detection result of the target moving object by the camera 11 (second detection result). The details of this point will be specifically described with reference to FIG. 1.

[0024] As shown in the upper diagram of FIG. 1, when a person is moving in a predetermined area, the infrared sensor 12 detects the person as a moving object due to the movement of the person (IR: detected). Then, when the infrared sensor 12 detects a moving object, the lighting device 1 drives the light source unit 10 to turn it on. Note that when the light source unit 10 is off, the detection processing of the target object by the camera 11 is not performed, and the detection processing of the target object is started after the light source unit 10 is turned on. This is because the target object detection processing by the AI model requires time, and does not satisfy the necessary condition in a situation where it is a necessary condition to detect a person and turn on the light immediately. In other words, by not performing the target object detection processing by the camera 11 when the light source unit 10 is off, the processing load of the lighting device 1 and the heat generation accompanying the operation of the AI model can be reduced.

[0025] Next, as shown in the lower diagram of Figure 1, if a person remains stationary in a predetermined area after the light source unit 10 has been turned on, the infrared sensor 12 will enter a non-detection state in the moving object detection process (IR: non-detection). On the other hand, the camera 11 starts the detection process after the light source is turned on, and therefore detects the stationary moving object as the target object. The lighting device 1 then continues to turn on the light source unit 10 if the target object is detected by the camera 11. In other words, the lighting device 1 continues to turn on the light source unit 10 if the first detection result from the infrared sensor 12 is non-detection, and the second detection result from the camera 11 detects the target object. Although not shown in the diagram, the lighting device 1 also turns off the light source unit 10 if the first detection result from the infrared sensor 12 is non-detection and the second detection result from the camera 11 is also non-detection.

[0026] Thus, in this disclosure, the lighting device 1 determines the control content based on a combination of the first detection result and the second detection result, enabling, for example, immediate illumination of the light source unit 10 by the infrared sensor 12 and continued illumination by the camera 11 when the infrared sensor 12 does not detect anything. In other words, the lighting device 1 can avoid a situation where the lights are mistakenly turned off because the infrared sensor 12 fails to detect a moving object, by detecting the target object with the camera 11 and continuing to illuminate. In other words, by determining the control content by combining the first detection result and the second detection result, the lighting device 1 can maximize the advantages of both the infrared sensor 12 and the camera 11 while complementing each other's disadvantages. Thus, according to this disclosure, the lighting device 1 can perform accurate lighting control in accordance with the movement of a moving object.

[0027] The embodiment shown in Figure 1 is just one example, and in this disclosure, various lighting controls can be performed using the first detection result and the second detection result, but other embodiments will be described later.

[0028] Furthermore, this disclosure makes it possible to generate training data for the AI ​​model based on the first and second detection results and retrain the AI ​​model (adjust parameters), but details regarding this will be described later.

[0029] Next, an example of the configuration of the lighting device 1 will be described using Figure 2. Figure 2 is a block diagram showing an example of the configuration of the lighting device 1 according to the embodiment. As shown in Figure 2, the lighting device 1 comprises a light source unit 10, a camera 11, an infrared sensor 12, a control unit 20, and a storage unit 30. The control unit 20 comprises a first acquisition unit 21, a second acquisition unit 22, a determination unit 23, and a learning unit 24.

[0030] Here, the lighting device 1 includes, for example, a computer and various circuits having a CPU (Central Processing Unit), ROM (Read Only Memory), RAM (Random Access Memory), flash memory, input / output ports, etc.

[0031] The computer's CPU functions as the first acquisition unit 21, second acquisition unit 22, determination unit 23, and learning unit 24 of the control unit 20, for example, by reading and executing a program stored in ROM.

[0032] Furthermore, at least one or all of the first acquisition unit 21, second acquisition unit 22, determination unit 23, and learning unit 24 of the control unit 20 can be configured with hardware such as an ASIC (Application Specific Integrated Circuit) or FPGA (Field Programmable Gate Array).

[0033] Furthermore, the storage unit 30 can, for example, correspond to RAM or flash memory. RAM and flash memory can store information about various programs, etc. The lighting device 1 may also acquire the above-mentioned programs and various information via other computers or portable recording media connected by a wired or wireless network.

[0034] The first acquisition unit 21 acquires a first detection result of a moving object based on the sensing results of the infrared sensor 12. The first detection result includes information on the presence or absence of a moving object and the location where the moving object was detected.

[0035] The second acquisition unit 22 acquires a second detection result of the target object based on the sensing results of the camera 11. Specifically, the second acquisition unit 22 uses an AI model to acquire a second detection result, which is the result of detecting the target object from the image of the camera 11. The second detection result includes information on the presence or absence of the target object and the location where the target object was detected.

[0036] The determination unit 23 determines the control content of the light source unit 10 according to the combination of the first detection result and the second detection result. The determination process of the determination unit 23 when the light source unit 10 is off and the determination process of the determination unit 23 when it is on will be explained below.

[0037] (When the lights are off) When the light source unit 10 is off, the decision unit 23 determines that if the infrared sensor 12 detects a moving object, it will turn on the light source unit 10, which is currently off. If the infrared sensor 12 does not detect a moving object, the decision unit 23 determines that it will keep the light source unit 10 off. Then, the decision unit 23 determines that it will turn on the light source unit 10, which is currently off, and starts the process of detecting the target object using the AI ​​model with the camera 11.

[0038] (When lit) When the light source unit 10 is lit, the decision unit 23 determines the control to continue the lit state of the light source unit 10 if the infrared sensor 12 does not detect the target object and the AI ​​model detects the target object. Furthermore, if the infrared sensor 12 does not detect the target object and the AI ​​model also does not detect the target object, the decision unit 23 determines the control to turn off the light source unit 10. This allows for accurate turning off of the light when the moving object to be illuminated is no longer present in the area. However, if the infrared sensor 12 does not detect the target object and the AI ​​model detects the target object, and the infrared sensor 12 has never previously detected a moving object at the target object's location, the decision unit 23 prohibits the determination of the control content because there is a possibility of false detection of the target object. In other words, the decision unit 23 maintains the current control state of the light source unit 10. This avoids control being performed based on the result of false detection of the target object. In this case, as will be described in detail later, the learning unit 24 generates training data and retrains the AI ​​model. Furthermore, if the infrared sensor 12 does not detect anything and the AI ​​model also does not detect anything, the determination unit 23 may not immediately switch to turning off the light, but may gradually reduce the brightness of the light source unit 10. In other words, the determination unit 23 gradually reduces the brightness of the light source unit 10 as the number of times both the first detection result and the second detection result are not detected increases, and turns off the light when the number of such occurrences exceeds a threshold.

[0039] The detection process by the AI ​​model is initiated when the light switches from off to on, but it is not limited to this; for example, it could also be initiated when the infrared sensor 12 fails to detect anything while the light is on. In other words, the determination unit 23 initiates the detection process by the AI ​​model when the infrared sensor 12 fails to detect anything, and determines the control content to continue the on state if the target object is detected during the detection process, and to switch to the off state if the object is not detected during the detection process.

[0040] The detection process by the AI ​​model may start when the infrared sensor 12 fails to detect anything for the first time, or when it fails to detect anything for multiple times (for a predetermined period). In this way, by starting the detection process by the AI ​​model after the infrared sensor 12 fails to detect anything, the processing load on the control unit 20 due to the detection process can be reduced. In addition, if the detection process by the AI ​​model is started after the infrared sensor 12 fails to detect anything, the timing of the switch from on to off will be delayed, but since the switch to off occurs because there are no people in the area, there will be no inconvenience to people due to the delay in turning off the lights.

[0041] Alternatively, the detection process by the AI ​​model may be performed continuously regardless of the control state of the light source unit 10 or the detection state of the infrared sensor 12. In this case, the determination unit 23 switches from off to on based on the detection of a moving object by the infrared sensor 12 or the detection of a target object by the AI ​​model. The switching from on to off can be performed in the same manner as described above.

[0042] The determination unit 23 then controls the illumination of the light source unit 10 according to the determined control content. If the illumination control device is configured separately from the illumination device 1, the illumination control device generates and outputs a control signal to control the illumination device 1 according to the control content determined by the determination unit 23.

[0043] The learning unit 24 generates training data for the AI ​​model based on the combination of the first detection result and the second detection result.

[0044] Specifically, the learning unit 24 detects a moving object based on the first detection result. If the second detection result indicates no detection, it determines that the AI ​​model has not been able to detect the moving object as a target object and generates training data to detect the moving object as a target object. Specifically, the learning unit 24 generates training data that serves as training data, in which the image captured by the camera 11 at the time the moving object was detected based on the first detection result is given a correct label indicating that it is a moving object (target object), and retrains the AI ​​model. As a result, the AI ​​model can accurately detect the moving object detected by the infrared sensor 12 as a target object.

[0045] Furthermore, if the learning unit 24 detects a target object based on the second detection result, and if it has never detected a moving object at the location of the target object in the past first detection results, it determines that the AI ​​model is mistakenly detecting an object that is not moving as the target object, and generates training data to prevent the AI ​​model from detecting objects that are not moving as the target object. Specifically, the learning unit 24 generates training data by adding error labels to the image at the time the target object was detected, and retrains the AI ​​model. This enables the AI ​​model to prevent it from detecting objects that are not moving as the target object. As a method for adding correct or incorrect labels to an image, the correct or incorrect labels may be added to the image data itself, or they may be added to a specific location (region) on the image. In addition, the learning unit 24 does not have to automatically generate the training data with correct and incorrect labels; for example, it may generate data showing candidate correct labels and candidate incorrect labels, and the user can later check whether they are correct or incorrect, and the learning process may be conducted based on the results from the user.

[0046] Next, the processing steps of the lighting device 1 according to the embodiment will be described using Figures 3 and 4. Figures 3 and 4 are flowcharts showing the processing steps of the lighting device 1 according to the embodiment. Figure 3 shows the processing when the lighting device 1 is turned off, and Figure 4 shows the processing when the lighting device 1 is turned on.

[0047] First, the process when the lighting device 1 is turned off will be explained using Figure 3. As shown in Figure 3, the control unit 20 determines whether or not a moving object has been detected by the infrared sensor 12 (step S101).

[0048] If the infrared sensor 12 detects a moving object (step S101: Yes), the control unit 20 controls the light source unit 10, which is currently off, to be turned on (step S102), and then terminates the process. If the infrared sensor 12 does not detect a moving object (step S101: No), the control unit 20 keeps the light source unit 10 off (step S103), and then terminates the process.

[0049] Next, the process when the lighting device 1 is turned on will be explained using Figure 4. As shown in Figure 4, the control unit 20 determines whether or not the infrared sensor 12 has detected a moving object (step S201).

[0050] If the control unit 20 detects a moving object with the infrared sensor 12 (step S201: Yes), it continues to keep the light source unit 10 lit (step S202) and terminates the process.

[0051] On the other hand, if the infrared sensor 12 does not detect a moving object (step S201: No), the control unit 20 determines whether the state of non-detection by the infrared sensor 12 continues for a certain period of time or longer (step S203).

[0052] If the infrared sensor 12 remains undetected for a certain period of time or longer (step S203: Yes), the control unit 20 determines whether or not the AI ​​model has detected the target object using the image from the camera 11 (step S204). On the other hand, if the infrared sensor 12 does not remain undetected for a certain period of time or longer (step S203: No), the control unit 20 returns to step S201.

[0053] In step S204, if the control unit 20 detects a target object in the AI ​​model (step S204: Yes), it continues to keep the light source unit 10 lit (step S202) and terminates the process.

[0054] On the other hand, in step S204, if the AI ​​model does not detect the target object (step S204: No), the control unit 20 controls the lit light source unit 10 to an off state (step S205) and terminates the process.

[0055] As described above, the lighting device 1 (lighting control device) according to the embodiment comprises a first acquisition unit 21, a second acquisition unit 22, and a determination unit 23. The first acquisition unit 21 acquires a first detection result of a moving object based on the sensing results of a predetermined area. The second acquisition unit 22 acquires a second detection result of an object to be illuminated based on the sensing results. The determination unit 23 determines the control content of the lighting device 1 to illuminate the predetermined area according to the combination of the first and second detection results. This enables accurate lighting control in accordance with the movement of the moving object.

[0056] In the embodiment described above, an example was shown in which the lighting device 1 has an AI model, but the AI ​​model may be held by a separate device. For example, the lighting device 1 generates training data, and the AI ​​model is trained by a separate device (such as a server). The lighting device 1 may also acquire the generated AI model from the separate device, or it may transmit images from the camera 11 to the separate device, and the separate device may transmit the results of detecting target objects from the images using the AI ​​model back to the lighting device 1. This can reduce the processing load on the lighting device 1.

[0057] Furthermore, in the embodiment described above, a camera (a camera different from camera 11) may be used instead of the infrared sensor 12, and the detection result from the other camera may be used as the first detection result. In this case, the first acquisition unit 21 may use the camera input to acquire the pixel value difference between the currently acquired image and a previously acquired image, for example, one frame ago, and detect whether or not a moving object is present. The learning unit 24 may, if the position of the target object detected in the image by the second detection result is an image position where no moving object has been detected in the past using the first detection result with camera input, or if no moving object was detected in the first detection result for the same image as the image corresponding to the second detection result, generate training data with an error label assigned to the image position of the target object detected in the second detection result. If the learning unit 24 detects a moving object in the first detection result with camera input, and the second detection result does not detect anything in the same image, it generates training data with a correct label assigned to the image region where the first detection result detected a moving object in that image.

[0058] While embodiments of the present invention have been described, these embodiments are presented as examples only and are not intended to limit the scope of the invention. These embodiments can be carried out in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their variations are included in the scope and spirit of the invention, as well as in the claims and their equivalents. [Explanation of Symbols]

[0059] 1. Lighting device 10 Light source section 11 Cameras 12 Infrared Sensors 20 Control Unit 21 First acquisition part 22 Second acquisition part 23 Decision Section 24 Learning Department 30 Storage section

Claims

1. A first acquisition unit that acquires a first detection result of a moving object based on the sensing results of a predetermined area; A second acquisition unit that acquires a second detection result of the target object to be illuminated based on the aforementioned sensing result; A determination unit that determines the control content of the lighting device that illuminates the predetermined area according to the combination of the first detection result and the second detection result; A lighting control device equipped with the following:

2. The aforementioned determination unit, When the lighting device is lit, and the first detection result is "not detected" and the second detection result detects the target object, the control content is determined to continue the lighting state. The lighting control device according to claim 1.

3. The aforementioned determination unit, When the lighting device is lit, and the first detection result is not detected and the second detection result is not detected, the control content to change the lit state to an off state is determined. The lighting control device according to claim 1.

4. The object detection process for obtaining the second detection result is performed after the first detection result is determined to be undetectable. The lighting control device according to claim 1.

5. The object detection process for obtaining the second detection result is performed after a predetermined period of time has passed in which the first detection result indicates that no object has been detected. The lighting control device according to claim 1.

6. The aforementioned determination unit, If the position of the target object detected in the second detection result is a position where the moving object has never been detected in the first detection result in the past, the process of determining the control content is prohibited. The lighting control device according to claim 1.

7. The first detection result is the detection result of the moving object by the infrared sensor, The second detection result is the detection result of an AI model that detects the target object based on an image captured by a camera. The lighting control device according to claim 1.

8. The system further includes a learning unit that generates training data based on a combination of the first and second detection results and performs training on the AI ​​model. The lighting control device according to claim 7.

9. The aforementioned learning unit, If the position of the target object detected in the second detection result is a position where the moving object has not been detected in previous first detection results, training data is generated by assigning an error label of the target object to the image corresponding to the second detection result. The lighting control device according to claim 8.

10. The aforementioned learning unit, If the first detection result detects the moving object, and the second detection result indicates no detection, training data is generated by assigning the correct label of the target object to the image corresponding to the second detection result. The lighting control device according to claim 8.

Images