Anti-dazzling screen lighting system using artificial intelligence to track objects

Abstract

Embodiments provide an anti-dazzling screen lighting system that tracks an object using artificial intelligence. A visor illumination system according to an embodiment includes: an image capturing device that captures an image of a preset area on which visor illumination is projected; a control device connected to the imaging device, recognizing an object from the image, and generating first illumination setting information and second illumination setting information based on information about the recognized object; a first pattern illumination device that directly projects the first pattern image to the object on the basis of first illumination setting information including a projection range of the first pattern image, a light amount of the first pattern image, and a color of the first pattern image; and a second pattern illumination device projecting the second pattern image on the floor of the preset area based on second illumination setting information including a projection range of the second pattern image, a light amount of the second pattern image, and a color of the second pattern image.

Description

Technical Field

[0001] Embodiments of the present disclosure relate to a light-shielding sheet illumination system that uses artificial intelligence to track an object. Background Art

[0002] A pattern lamp is a device that combines disks with shapes, patterns, or symbols thereon into an illumination device, and can selectively transmit light according to the shapes, patterns, or symbols formed on the disks.

[0003] In the past, pattern sheet illumination was mainly used to produce special effects such as performances and broadcasts, but recently, images for advertising or promoting a company or product are projected onto the floor of a road or the wall of a building, and it is also widely used for advertising purposes.

[0004] At the same time, as the light sources used in street lamps installed on farm roads are replaced by high-brightness LEDs, the damage caused by light pollution to crops is increasing day by day.

[0005] In particular, it has a negative impact on the light cycle of crops. If outdoor lighting is not used to prevent the damage of light pollution, the impact on crops will disappear, but the original purposes of night lighting, such as safety and security, will disappear and be lost, leading to another problem.

[0006] In addition, the projection range of existing street lamps is limited by a screen to prevent light pollution to crops, but it is difficult to apply it to various environments by installing a screen, and there is a risk of injury caused by the screen falling off.

[0007] Therefore, there is a need for a farm street lamp using a light-shielding plate illumination module, so that a light-shielding plate image formed in the form of a farm road can be projected onto farmland, and the projection of light on crops can be restricted without installing a separate screen.

[0008] At the same time, control refers to managing and controlling people or things for the safety and comfort of society, or the maintenance and preservation of order. Generally speaking, an integrated control system is a system operated by a national or local government to prevent or promptly resolve traffic accidents, crimes, disaster situations, or various illegal acts.

[0009] Existing camera-based management systems use CCTV (Closed Circuit Television) and the Internet (Internet Protocol Network) to obtain information, and judge whether an event has occurred based on the obtained information, and play a role in immediately providing event information. This has also happened in relevant organizations such as the National Emergency Management Agency and the National Police Agency.

[0010] However, existing such comprehensive control systems generally require multiple CCTVs and a large bandwidth to transmit the video information collected by multiple CCTVs in real time to a control room, and a large number of personnel are needed to watch the multiple videos received from the CCTVs in real time to determine the time when an event occurs. In other words, existing comprehensive control systems are only systems that can be used by organizations such as national or local governments with a large amount of resources and manpower.

[0011] Therefore, a solution is needed that allows individuals or small organizations to conduct control within a narrow geographical area, such as controlling illegal waste dumping in a residential area or smoking in a non-smoking area within a building, which is happening.

[0012] In addition, existing control systems adopt non-face-to-face deep learning-based systems, but there are insufficient measures to directly intervene at the scene. Generally, sound devices are used for remote intervention at the scene, but this is less effective for people wearing sound output devices (such as headphones), and the residents around the control system may be affected by the noise.

[0013] Therefore, a light-shielding plate lighting that can ensure visibility is used, but based on the image obtained through the imaging device, the information of the object in the image is analyzed by artificial intelligence, and based on the information of the object, the object is directly contacted. A system is needed to project a pattern piece lighting onto the floor to share the current situation with the people around. Summary of the Invention

[0014] Problems to be Solved by the Invention

[0015] Embodiments of the present disclosure may provide a light-shielding plate lighting system that uses artificial intelligence to track an object.

[0016] The technical challenges to be achieved by the embodiments are not limited to the above, and those skilled in the art can consider other technical challenges not mentioned from the various embodiments described below.

[0017] Methods for Solving the Problems

[0018] A shutter illumination system for tracking an object using artificial intelligence according to an embodiment includes: a camera device that captures an image of a preset area where shutter illumination is projected; a control device connected to the camera device, which identifies an object from the image and generates first illumination setting information and second illumination setting information based on information about the identified object; a first pattern illumination device that directly projects a first pattern image onto the object based on the first illumination setting information including the projection range of the first pattern image, the light amount of the first pattern image, and the color of the first pattern image; and a second pattern illumination device that projects a second pattern image onto the floor of the preset area based on the second illumination setting information including the projection range of the second pattern image, the light amount of the second pattern image, and the color of the second pattern image.

[0019] The control device uses an object extraction model of a first neural network based on images to determine the type, size, and position of the object in real time, as well as the image, the type, and the position of the object. Based on the position of the object, a behavior analysis model of a second neural network is used to determine the behavior type of the object. The information of the object includes the type of the object, the size of the object, and the behavior type of the object.

[0020] The first shutter illumination device can project a first shutter image while tracking the position of the object in real time. The second shutter illumination device can project a second shutter image at the projection position of the second shutter image.

[0021] The projection position of the second shutter image can be determined based on the position of the object and the behavior type of the object. The second pattern image can include a message determined based on the type of the object and the behavior type of the object.

[0022] According to an embodiment, the object extraction model can use an algorithm that simultaneously performs bounding box coordination and classification.

[0023] Multiple image vectors can be generated by performing first data preprocessing on the image. The image vectors can include pixel values of static images of the video.

[0024] Advantages of the Invention

[0025] As described above, the present invention has the following advantages:

[0026] According to an embodiment, the shutter illumination system determines the type and size of the object in real time by using an object extraction model of a first neural network based on images, so as to determine the type and size of the object, and thus can prevent the movement of the object by tracking the position of the object and directly projecting an appropriate first pattern image onto the object.

[0027] According to an embodiment, the light-shielding sheet illumination system determines the behavior type of an object by using a behavior analysis model of a second neural network based on an image, the type of the object, the size of the object, and the position of the object, and determines the behavior type. The projection position is determined according to the position of the object and the behavior type of the object, and the second pattern image includes a message determined according to the type of the object and the behavior type of the object, thereby forming guidance or warning information for the object. Two pattern images can be effectively projected.

[0028] The effects that can be obtained from the embodiments are not limited to the above effects, and other effects not mentioned can be clearly derived and understood by those skilled in the art based on the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 FIG. is a diagram of a rural lighting system including an agricultural street lamp using a light-shielding sheet lamp module according to an embodiment.

[0030] Figure 2 FIG. is a diagram of a farm street lamp using a light-shielding sheet illumination module according to an embodiment.

[0031] Figure 3 FIG. shows an example of a lens unit of a light-shielding plate illumination module according to an embodiment. DETAILED DESCRIPTION

[0032] Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings so that those skilled in the art can easily implement the present invention.

[0033] When describing the present embodiment, descriptions of content well known in the technical field to which the present invention belongs and technical content not directly related to the present invention will be omitted. This is to more clearly convey the gist of the present invention without confusing the gist of the present invention by omitting unnecessary descriptions.

[0034] For the same reason, some components in the drawings will be shown in an exaggerated or omitted or simplified manner.

[0035] Figure 1 FIG. is a diagram of a rural lighting system including an agricultural street lamp using a light-shielding sheet lamp module according to an embodiment. Figure 1 One embodiment of can be combined with various embodiments of the present disclosure.

[0036] Referring to Figure 1 , the agricultural road lighting system may include a rural street lamp 100 and a street lamp management server 200 using a light-shielding plate illumination module.

[0037] The farm street lamp 100 can be a street lamp that uses a light-shielding plate illumination module to project a light-shielding plate image onto the ground of the farmland. At this time, the farm street lamp 100 can project light only onto the farm road without a light shield by using a light-shielding plate image that limits the projection range to the area where crops are planted in the farmland. For example, the farmland road street lamp 100 can project a light-shielding plate image in the form of a farmland road formed on the farmland (hereinafter referred to as a farmland road light-shielding plate image) onto the bottom surface of the farmland.

[0038] The farm street lamp 100 can communicate with an external device through a wireless communication network. For example, the agricultural street lamp 100 can be pre-connected to the street lamp management server 200 through a wireless communication network.

[0039] The agricultural street lamp 100 can send the surrounding status information sensed by the sensor module set in the agricultural street lamp 100 to the street lamp management server 200. The agricultural street lamp 100 can receive setting information from the street lamp management server 200 through a wireless communication network. The farmland road street lamp 100 can adjust the angle of the pattern piece illumination module according to the setting and project a farmland road pattern image onto the bottom surface of the farmland, the size of which is the same as the size of the farmland road formed in the farmland.

[0040] The agricultural street lamp 100 can adjust the light amount for projecting the agricultural pattern piece image based on the surrounding status information and the setting information.

[0041] The street lamp management server 200 can be a server that manages the agricultural street lamp 100. The street lamp management server 200 can be pre-connected to the farm street lamp 100 through a wireless communication network. The street lamp management server 200 can receive the surrounding status information from the agricultural street lamp 100 through a wireless communication network and determine the setting information of the agricultural street lamp 100 based on the surrounding status information. The street lamp management server 200 can send the setting information to the farm street lamp 100 through a wireless communication network. For example, the street lamp management server 200 can remotely control the farm street lamp 100 through a wireless communication network.

[0042] For example, the farm street lamp 100 can include a processor, a memory, and a communication module.

[0043] The processor may, for example, execute software to control at least one other component (e.g., a hardware or software component) of a device coupled to the processor and may perform various data processing or operations. According to one embodiment, as at least part of the data processing or computing, the processor stores commands or data received from other components (e.g., a sensor module or a communication module) in a volatile memory and stores the commands or data stored in the volatile memory that can be processed and the resulting data can be stored in a non-volatile memory. According to one embodiment, the processor is a main processor (e.g., a central processing unit or an application processor) or an auxiliary processor (e.g., a graphics processing unit, a neural processing unit (NPU), an image signal processor, a sensor hub processor, or a communication processor) that can operate independently or together.

[0044] An artificial intelligence model may be created through machine learning. For example, such learning may be performed in the device itself (e.g., the agricultural street lamp 100) that executes the artificial intelligence model, or may be performed by a separate server (e.g., the street lamp management server 200). The learning algorithm may include, for example but not limited to, supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning. The artificial intelligence model may include multiple artificial neural network layers. The artificial neural network includes a deep neural network (DNN), a convolutional neural network (CNN), a recurrent neural network (RNN), a restricted Boltzmann machine (RBM), a deep belief network (DBN), a bidirectional recurrent deep neural network (BRDNN), etc. It may be one of the deep Q networks or a combination of two or more of the above networks, but is not limited to the above examples. In addition to the hardware structure, the artificial intelligence model may additionally or alternatively include a software structure.

[0045] The memory may store various data used by at least one component (e.g., the processor) of the device. The data may include, for example, input data or output data of software and instructions related thereto. The memory may include a volatile memory or a non-volatile memory.

[0046] The communication module can support establishing a direct (e.g., wired) or wireless communication channel between the device and an external device and performing communication through the established communication channel. The communication module operates independently of the processor and can include one or more communication processors that support direct (e.g., wired) communication or wireless communication. According to one embodiment, the communication module is a wireless communication module (e.g., a cellular communication module, a short-range wireless communication module, or a GNSS (Global Navigation Satellite System) communication module) or a wired communication module (e.g., a wireless communication module). LAN (Local Area Network) communication module), or a power line communication module). Among these communication modules, the corresponding communication module is a first network (e.g., a short-range communication network such as Bluetooth, WiFi (Wireless Fidelity) Direct, or IrDA (Infrared Data Association)) or a second network (e.g., a traditional network). A cellular network, a 5G network, which can communicate with an external device through a telecommunications network such as a next-generation communication network, the Internet, or a computer network (e.g., LAN or WAN). These various types of communication modules can be integrated into one component (e.g., a single chip) or can be implemented as multiple separate components (e.g., multiple chips).

[0047] Throughout this specification, neural network, neural network, and network function may be used with the same meaning. A neural network may consist of a set of interconnected computing units, commonly referred to as "nodes". These "nodes" may also be called "neurons". A neural network consists of at least two or more nodes. The nodes (or neurons) that make up a neural network can be interconnected by one or more "links".

[0048] In a neural network, two or more nodes connected by a link can relatively form a relationship of an input node and an output node. The concepts of input node and output node are relative, and any node in an output node relationship with one node may be in an input node relationship with another node, and vice versa. As described above, the relationship of input node to output node can be created around the link. One or more output nodes can be connected to an input node by a link, and vice versa.

[0049] In the relationship between an input node and an output node connected by a link, the value of the output node can be determined based on the data input to the input node. Here, the node connecting the input node and the output node can have a weight. The weight can be variable and can be changed by a user or an algorithm so that the neural network performs the desired function. For example, when one or more input nodes are connected to an output node through their respective links, the output node is set to the value of the input nodes connected to the output node and the corresponding links for each input node. The node value can be determined according to the weight.

[0050] As described above, in a neural network, two or more nodes are interconnected by one or more links to form an input node and output node relationship within the neural network. The characteristics of a neural network can be determined based on the number of nodes and links in the neural network, the correlation between the nodes and links, and the weight values assigned to each link. For example, if two neural networks have the same number of nodes and links and different weight values between the links, the two neural networks may be recognized as different from each other.

[0051] Figure 2 is a diagram of a farm street lamp using a light-shielding plate lighting module according to an embodiment. Figure 3 An example of a lens unit of a light-shielding plate lighting module according to an embodiment is shown.

[0052] Referring to Figure 2 , the farm street lamp 100 may include a light-shielding plate lighting module 110, a solar power supply module 120, a sensor module 130, a communication module 140, and a control module 150.

[0053] The light-shielding plate lighting module 110 may project a light-shielding plate image onto the ground of the farmland. For example, the light-shielding plate lighting module 110 may change the angle at which the light-shielding plate image is projected according to a control signal received from the control module 150. For example, the light-shielding plate lighting module 110 may change the amount of light used to project the light-shielding plate image according to a control signal received from the control module 150.

[0054] The solar power supply module 120 may generate electricity using sunlight and supply power to the farm street lamp. The solar power supply module 120 may convert sunlight into electrical energy. For example, the solar power supply module 120 may change the angle at which sunlight is absorbed according to a control signal received from the control module 150.

[0055] The sensor module 130 detects the operating state (e.g., light amount or angle) of the farmland street lamp 100 or the external environment state (e.g., temperature or illuminance), and generates an electrical signal or data value corresponding to the detected state. is created. For example, the sensor module 130 may include a plurality of sensors and a camera unit. The sensor module 130 may sense the surrounding state of the agricultural street lamp 100 through a plurality of sensors. For example, the plurality of sensors may include an infrared sensor, a temperature sensor, and an illuminance sensor. The sensor module 130 may capture static images and images around the farm street lamp 100 through the camera unit. For example, the camera unit may include one or more lenses, an image sensor, an image signal processor, or a flash.

[0056] The communication module 140 may support the establishment of a direct (e.g., wired) or wireless communication channel between the farm street lamp 100 and an external device, and communication through the established communication channel.

[0057] The control module 150 may include a processor and a memory. For example, the control module 150 may control the shutter illumination module 110, the solar power supply module 120, the sensor module 130, and the communication module 140 through the processor and the memory.

[0058] The control module 150 may obtain surrounding state information through the sensor module 130. For example, the surrounding state information may include an image captured by a camera unit provided in the sensing module 130 and sensing information. The sensing information may include an illuminance value for each time period, a temperature value for each time period, and an infrared-related value for each time period.

[0059] The control module 150 may determine the angle of the pattern illumination module 110 based on the setting information. The control module 150 may control the angle of the pattern illumination module 110 to project a farm road pattern image onto the bottom of the farmland, the size of which is the same as that of the farm road formed on the farmland.

[0060] The control module 150 may adjust the amount of light for projecting the agricultural pattern piece image based on the surrounding state information and the setting information.

[0061] The shutter illumination module 110 includes an LED (light-emitting diode) light source unit 111, a first shutter image filter unit 112, a second shutter image filter unit 113, a lens unit 114, and an illumination driver unit 115.

[0062] The LED light source unit 111 may include a plurality of LEDs and a heat sink. The LED light source unit 111 may output monochromatic light or white light having an arbitrary wavelength through the plurality of LEDs. For example, the plurality of LEDs may be installed inside the rear housing of the shutter illumination module 110 and output light at an arbitrary light distribution angle.

[0063] At this time, the heat sink may dissipate the heat generated by the plurality of LEDs. That is, the plurality of LEDs may be installed on the heat sink or the heat sink and configured to dissipate the heat generated during the light-emitting operation to the outside. For example, the heat sink may be made of a copper or aluminum material having a high thermal conductivity, but is not limited thereto. Additionally, for example, the heat sink may be combined with a separate cooling pen.

[0064] The first light-shielding image filter unit 112 can selectively transmit the light output from the LED light source unit 111 according to a pre-formed shape. For example, the first light-shielding image filter unit 112 can include an aluminized mirror that blocks visible light by depositing aluminum on one surface of glass. The first light-shielding image filter unit 112 can be manufactured to block visible light by depositing aluminum on one surface of glass. In this case, the side where aluminum is deposited reflects visible light, while the other side can absorb visible light. That is, the aluminum deposited on the first light-shielding image filter unit 112 is selectively removed using a laser so that a predetermined shape can be imprinted on the first light-shielding image filter unit 112. For example, the first light-shielding plate image filtering unit 112 can include a first area in which the shape of a farm road formed on farmland is imprinted. The first area can be formed by selectively removing the aluminum deposited on the first light-shielding image filter unit 112 using a laser.

[0065] For example, the first light-shielding image filter unit 112 can be disposed between a first lens group and a second lens group that constitute a lens unit 114 in a disk form. At this time, the first lens group can refract light so that the light output from the LED light source unit 111 is evenly projected onto the first light-shielding image filter unit 112. The second lens group can refract the light passing through the first light-shielding plate image filter unit 112 to form an image.

[0066] Additionally, for example, the first light-shielding image filter unit 112 can be manufactured in the form of a turntable including a plurality of pre-formed areas. The first light-shielding image filter unit 112 is formed with a plurality of openings, and the first light-shielding image filters can be detachably mounted in each opening. Here, the first light-shielding image filter can be a aluminized mirror, where aluminum is deposited on one surface of the glass and then laser-engraved into a specific shape. The first light-shielding image filter unit 112 can also include a motor connected to a rotating shaft to rotate the first light-shielding image filter unit 112. In this case, the first light-shielding image filter unit 112 can rotate according to the rotation angle or driving force of the motor. For example, the first light-shielding image filter unit 112 can rotate such that a specific first light-shielding image filter is aligned with the first lens group. For example, the motor can rotate the first light-shielding image filter detachably mounted on the first light-shielding image filter unit 112 provided in front of the first lens group according to a control signal from the control module 150. At this time, the motor can be connected to the central axis of the first light-shielding image filter unit 112. That is, a plurality of first light-shielding image filters can be respectively mounted in the plurality of openings of the first light-shielding image filter unit 112, and any one of the plurality of first light-shielding image filters can be selected by the rotation of the motor. Thus, a specific preset pattern image can be selected or changed by the operation of the motor.

[0067] The second light-shielding image filter unit 113 can selectively transmit wavelengths of a specific color in the light output from the LED light source unit 111. For example, the second light-shielding image filter unit 113 can include a dichroic filter that transmits only a specific color by coating a dichroic material on one surface of the glass. The second light-shielding image filter unit 113 can be manufactured by coating a multi-layer film of a dichroic material such as metal oxide, sulfide, or fluoride on one surface of the glass in a high vacuum to transmit only a specific color. For example, by selectively removing the coating of the dichroic filter from the second light-shielding image filter unit 113 using a laser, the dichroic filter having wavelengths within a preset wavelength range can be retained in the second light-shielding image filter unit 113. The wing filter area is transparent. For example, the second light-shielding image filter unit 113 can include a second area engraved such that wavelengths within a preset wavelength range are not transmitted. The preset wavelength range can include the wavelength ranges that interfere with the growth of crops planted in the farmland. For example, the preset wavelength range can include the red light wavelength range and the blue light wavelength range. The preset wavelength range can be determined differently according to the type of crops. For example, the detailed wavelength range of red light can be determined within the wavelength range of 620 to 750 nanometers of red light and the wavelength range of 450 to 495 nanometers of blue light according to the type of crop. The blue light can be determined according to the type of the relevant crop.

[0068] In other words, plants use photosensitive proteins such as phytochrome and cryptochrome to sense seasonal changes and photoperiods based on the length of the night, and phytochrome and cryptochrome detect red light and blue light respectively. Therefore, since the intermediate green light does not affect the photoperiod of crops and is not used for photosynthesis, the farm street lamp 100 does not transmit red light and chord light through the second light-shielding plate image filtering unit 113. By doing so, only green light can be projected onto the bottom surface of the farmland. Thus, the farm street lamp 100 can project light without disturbing the growth of crops and can ensure visibility by utilizing the sensitivity of the human eye to green.

[0069] For example, the second light-shielding plate image filter unit 113 can be arranged between the first lens group and the second lens group that form a disk-shaped lens unit 114. At this time, the first lens group can refract light so that the light output from the LED light source unit 111 is evenly projected onto the second light-shielding plate image filter unit 113. The second lens group can refract the light passing through the second light-shielding plate image filter unit 113 to form an image.

[0070] For example, the first light-shielding plate image filtering unit 112 and the second light-shielding plate image filtering unit 113 can be arranged between the first lens group and the second lens group.

[0071] At this time, the first light-shielding plate image filter unit 112 can be adjacent to the first lens group, and the second light-shielding plate image filter unit 113 can be adjacent to the second lens group.

[0072] Alternatively, the first light-shielding plate image filter unit 112 can be adjacent to the second lens group, and the second light-shielding plate image filter unit 113 can be adjacent to the first lens group.

[0073] In addition, for example, the second light-shielding image filter unit 113 may be manufactured in the form of a turntable including a plurality of pre-formed regions. The second light-shielding image filter unit 112 is formed with a plurality of openings, and the second light-shielding image filters may be detachably installed in each opening. Here, the second light-shielding image filter may be a dichroic filter, in which a dichroic material is deposited on one surface of glass and then laser-engraved into a specific shape. The second light-shielding image filter unit 113 may further include a motor connected to a rotating shaft to rotate the second light-shielding image filter unit 113. In this case, the second light-shielding image filter unit 113 may rotate according to the rotation angle or driving force of the motor. For example, the second light-shielding image filter unit 113 may rotate such that a specific second light-shielding image filter is aligned with the first lens group. For example, the motor may rotate the second light-shielding image filter detachably installed on the second light-shielding image filter unit 113 provided in front of the first lens group according to a control signal from the control module 150. At this time, the motor may be coupled to the central axis of the second light-shielding image filter unit 113. That is, a plurality of second light-shielding image filters may be respectively installed in the plurality of openings of the second light-shielding image filter unit 113, and any one of the plurality of second light-shielding image filters may be selected by the rotation of the motor. Thus, a specific preset wavelength range may be selected or changed by the operation of the motor.

[0074] The lens unit 114 is composed of a plurality of lenses and may refract light such that the light output from the LED light source unit 111 is projected onto the first light-shielding image filter unit 112 and the second light-shielding image filter unit 113. The lens unit 114 may refract the light projected onto the first light-shielding image filter unit 112 and the second light-shielding image filter unit 113 to form an image.

[0075] For example, the lens unit 114 may include a first lens group and a second lens group.

[0076] For example, the first lens group may be composed of three lenses arranged in sequence from the LED light source unit 111 to the first light-shielding image filter unit 112 and the second light-shielding image filter unit 113. At this time, among the lenses included in the first lens group, the first lens arranged closest to the LED light source unit 111 may have a positive refractive power. Among the lenses included in the first lens group, the third lens arranged closest to the first light-shielding image filter unit 112 and the second light-shielding image filter unit 113 may have a positive refractive power. The second lens arranged between the first lens and the third lens may have a positive refractive power.

[0077] For example, the surface of the first lens facing the image direction is convex towards the image direction, the surface facing the LED light source unit 111 is concave towards the image direction, and the effective light diameter of the first lens where the surface facing the image is concave (the light passing aperture) can be more than twice the effective light diameter on the side facing the LED light source unit 111.

[0078] For example, the first lens can have a positive radius of curvature on both the side facing the image and the side facing the LED light source unit 111. The second lens can have a positive radius of curvature on both the side facing the image and the side facing the LED light source unit 111. The surface of the third lens facing the image can have a positive radius of curvature, but the surface facing the LED light source unit 111 can have a negative radius of curvature.

[0079] That is to say, in order to further increase the maximum light amount, average light amount, and light distribution of the pattern illumination module 110 compared with the existing pattern illumination device, the number of lenses is changed, and the side of the first lens facing the LED light source unit 111 is changed and designed to have a concave shape towards the image direction.

[0080] For example, the second lens can refract the light passing through the first lens towards the optical axis. For this purpose, the side of the second lens facing the image can have a convex shape towards the image direction, and at the same time, the side facing the LED light source unit 111 can have a concave shape image towards the image direction.

[0081] For example, the effective optical diameter of the second lens on the side facing the light source can be 80% of the effective optical diameter on the side facing the image.

[0082] For example, the third lens can refract the light passing through the second lens towards the optical axis. For this purpose, the surface of the third lens facing the image can have a convex shape towards the image direction, and the surface facing the LED light source unit 111 can have a convex shape towards the direction of the LED light source unit 111.

[0083] For example, the third lens can have the same effective diameter on the side facing the image and the same effective diameter on the side facing the light source.

[0084] For example, the second lens group can be composed of four lenses arranged in sequence in the image direction in the first light-shielding image filter unit 112 and the second light-shielding image filter unit 113.

[0085] For example, among the lenses included in the second lens group, the fourth lens arranged closest to the first light-shielding image filter unit 112 and the second light-shielding image filter unit 113 may have a positive refractive power. The fifth lens arranged adjacent to the fourth lens in the image direction may have a negative refractive power. The sixth lens arranged adjacent to the fifth lens in the image direction may have a positive refractive power. Among the lenses included in the second lens group, the seventh lens arranged closest to the image may have a positive refractive power.

[0086] For example, the surface of the fourth lens facing the image may have a positive radius of curvature, but the surface facing the LED light source unit 111 may have a negative radius of curvature. The surface of the fifth lens facing the image may have a negative radius of curvature, but the surface facing the LED light source unit 111 may have a positive radius of curvature. The surface of the sixth lens facing the image may have a negative radius of curvature, but the surface facing the LED light source unit 111 may have a positive radius of curvature. The surface of the seventh lens facing the image may have a positive radius of curvature, but the surface facing the LED light source unit 111 may have a negative radius of curvature.

[0087] For example, the fourth lens may refract the light passing through the third lens of the first lens group to direct it toward the optical axis. For this purpose, the surface of the fourth lens facing the image may have a convex shape toward the image direction, and the surface facing the LED light source unit 111 may have a convex shape toward the direction of the LED light source unit 111.

[0088] For example, the fifth lens may refract the light passing through the fourth lens to direct it in the opposite direction of the optical axis. For this purpose, the surface of the fifth lens facing the image may have a concave shape toward the image direction, and the side facing the LED light source unit 111 may have a concave shape toward the direction of the LED light source unit 111.

[0089] For example, the sixth lens may refract the light passing through the fifth lens toward the optical axis. For this purpose, the surface of the sixth lens facing the image may have a concave shape toward the image direction, and the side facing the LED light source unit 111 may have a concave shape toward the direction of the LED light source unit 111.

[0090] For example, the seventh lens may refract the light passing through the sixth lens toward the optical axis. For this purpose, the surface of the seventh lens facing the image may have a convex shape toward the image direction, and the surface facing the LED light source unit 111 may have a convex shape toward the direction of the LED light source unit 111.

[0091] For example, the surface of the sixth lens facing the image and the surface of the seventh lens facing the LED light source unit 111 may be adhered to each other.

[0092] For example, the fourth lens to the sixth lens may have the same effective optical diameter on the side facing the LED light source unit 111 and the same effective optical diameter on the side facing the image.

[0093] For example, the seventh lens may have the following dimensions: the effective optical diameter on the side facing the LED light source unit 111 is 75% of the effective optical diameter on the side facing the image.

[0094] The first lens group G1 may include a first lens L1, a second lens L2, and a third lens L3. The first light-shielding sheet image filter unit 112 and the second light-shielding sheet image filter unit 113 may be disposed between the first lens group G1 and the second lens group G2. The second lens group G2 may include a fourth lens L4, a fifth lens L5, a sixth lens L6, and a seventh lens L7.

[0095] The illumination driver 115 may adjust the angle of the shutter illumination module 110. For example, the illumination driver 115 may include a first stepping motor for adjusting the horizontal angle of the pattern illumination module 110 and a second stepping motor for adjusting the vertical angle. In a stepping motor, each step represents a certain angle. For example, in a 1.8-degree stepping motor, 1 step may mean a rotation of 1.8 degrees. For example, the illumination driver 115 may receive a first control signal for the horizontal angle and a second control signal for the vertical angle from the control module 150. The illumination driver 115 may adjust the angle of the pattern illumination module 110 by operating the first stepping motor and the second stepping motor according to the number of the first control signal and the second control signal.

[0096] For example, the farmland street lamp 100 transmits the light output from the LED light source unit 111 to the first area and the second area to project a farmland road pattern image in the form of a farmland road formed on the farmland onto the bottom surface of the farmland. At this time, the farmland road lamp 100 may adjust the angle of the pattern illumination module 110 according to the setting information and project the farmland road pattern image onto the bottom surface of the farmland having the same size as the farmland road formed on the farmland.

[0097] The solar power supply module 120 may include a solar panel 121 and a panel driver 122.

[0098] The solar panel 121 is installed on the top of the farmhouse street lamp 100 and can convert sunlight into electric energy. The solar panel 121 may include a plurality of solar cells that absorb sunlight, convert it into electric energy and store it, and then send the stored electric energy to the agricultural street lamp 100 when the illuminance is lower than a preset reference illuminance value.

[0099] The panel driver 122 can adjust the angle of the solar panel 121. For example, the panel driver 122 can include a third stepper motor for adjusting the rotation angle of the solar panel 121 and a fourth stepper motor for adjusting the vertical angle. For example, the panel driver 122 can receive a third control signal for the rotation angle and a fourth control signal for the vertical angle from the control module 150. The panel driver 122 can adjust the angle of the solar panel 121 by operating the third stepper motor and the fourth stepper motor according to the amounts of the third control signal and the fourth control signal.

[0100] For example, the control module 150 determines in real time the direction with the highest illuminance through the sensor module 140 and sends a control signal to adjust the angle of the solar panel 121 to the direction with the highest illuminance to the drive unit 122.

[0101] According to an embodiment, the farmland street lamp 100 can project a farmland shading plate image corresponding to an area with a diameter of 8 m onto the farmland through the shading plate lighting module 110 installed at a high position, using a lens with a field of view angle of 53 degrees. There is a 5-meter one.

[0102] According to an embodiment, the farmland street lamp 100 can project a farmland shading plate image corresponding to an area with a diameter of 15 m onto the farmland through the shading plate lighting module 110 installed at a high position, using a lens with a 90-degree viewing angle. There is a 5-meter one.

[0103] According to an embodiment, the farm road street lamp 100 or the street lamp management server 200 can provide information about the farm road area and the crop planting area based on the image captured by the camera unit, by using an object search model of the first neural network. Information about the crop area can be determined.

[0104] For example, the agricultural street lamp 100 can take a photo of the surrounding environment of the agricultural street building 100 through the camera unit before operating the shading plate lighting module 110.

[0105] For example, the agricultural street lamp 100 can send the surrounding state information to the street lamp management server 200. The surrounding state information can include the image captured by the camera unit and the sensing information. The sensing information can include the illuminance value for each time period, the temperature value for each time period, and the value related to infrared rays for each time period. The illuminance value is the illuminance value around the agricultural street lamp 100 and can be in lux. The temperature value is the temperature value around the farm street lamp 100 and can be in degrees Celsius. The value related to infrared rays includes the temperature change value and the value of the coordinates where the temperature change is detected for each time period around the farm street lamp 100, and can be in degrees Celsius.

[0106] For example, the agricultural street lamp 100 can periodically send the surrounding state information to the street lamp management server 200.

[0107] For example, the agricultural road lamp 100 can determine information about the agricultural road area and information about the crop area through an object search model using a first neural network based on the images captured by the camera unit.

[0108] Alternatively, for example, the agricultural street lamp 100 can send the images captured by the camera unit to the street lamp management server 200. The street lamp management server 200 can determine information about the agricultural road area and information about the crop area through an object search model using a first neural network based on the images captured by the camera unit.

[0109] For example, the street lamp management server 200 can send setting information including information about the agricultural road area and second area information to the agricultural road lamp 100.

[0110] The setting information can include information about the agricultural road area and second area information. The information about the agricultural road area can include coordinate information of the agricultural road area in the area where the agricultural road lamp 100 will project the agricultural road pattern image. The information about the crop area can include coordinate information of the crop area in the area where the farm road lamp 100 will project the farm road light-shielding sheet image.

[0111] For example, the farm road lamp 100 can determine the angle of the light-shielding plate lighting module 110 based on the information about the farm road area and the information about the crop area. For example, the control module 150 can set the horizontal angle and the vertical angle, and can send a control signal corresponding to the horizontal angle and the vertical angle of the pattern lighting module 110 to the lighting driver 115.

[0112] For example, the horizontal angle and the vertical angle of the pattern lighting module 110 are the first center of the agricultural pattern image projected at the horizontal angle and the vertical angle of the pattern lighting module 110 from the current position of the pattern lighting module 110. The xy values of the coordinates can be determined as the angles compensating for the difference between the horizontal angle and the vertical angle to match the xy values of the second center coordinates determined based on the information about the agricultural road area and the information about the crop area. For example, the information about the farm road area and the information about the crop area can be used to determine the center coordinates where the light-shielding plate lighting module 110 can be projected onto the farmland. For example, the control module 150 operating the light-shielding plate lighting module (110) can determine the projectable center coordinates.

[0113] In addition, for example, even if the xy values of the first center coordinates at the current position of the pattern illumination module 110 match the xy values of the second center coordinates, in the farm pattern image projected onto the bottom surface of the farmland, if the width of the farmland road does not match the actual width of the farmland road, the illumination driver 115 sets the pattern illumination module 110 to a height that matches the actual width of the farmland road in the farmland road pattern image projected onto the bottom surface of the farmland. The farmland can be controlled by an actuator.

[0114] For example, an image captured by the camera unit can be converted into an image vector including a plurality of pixel values through data preprocessing. The image vector can be input into the object detection model. For example, a plurality of image vectors can be generated by performing data preprocessing on the image captured by the camera unit, and the plurality of image vectors can be input into the object search model.

[0115] A system that uses a light-shielding sheet illumination to track an object using artificial intelligence may include a camera device, a control device, a first light-shielding sheet illumination device, and a second light-shielding sheet illumination device.

[0116] The pattern illumination system projects a first pattern image directly onto an object according to the type, size, and behavior type of the object within a preset area, and projects a second pattern image to form warning or guiding information in the viewing direction or orientation of the object. This is a system that projects onto the floor according to the viewing direction of an observer (such as a pedestrian).

[0117] The camera device can capture an image of the preset area where the pattern light is projected and provide the image to the control device. For example, the camera device can capture an image by horizontally or vertically tilting the installation position. Alternatively, the camera device can be fixed at the installation position to capture an image. The camera device converts the light incident through the lens into an electrical signal through a CCD (Charge-Coupled Device), converts the electrical signal into a digital signal using an ADC (Analog-to-Digital Converter), and converts the converted digital signal into a set of digital signal formats. You can compress the video according to the format for saving. The camera device operates permanently (permanently activated), and the captured images can be transmitted to the control device. For example, in response to a signal received from the control device, the camera device can capture images only at specific time points and send them to the control device. For example, the area where the first light-shielding sheet illumination device and the second light-shielding sheet illumination device can project the light-shielding sheet illumination and the area to be photographed by the camera device can be the same or can include each other. The camera device and the control device can be directly connected through an optical cable. The control device can receive video data from the camera device through UDP (User Datagram Protocol) or IPX (Internetwork Packet Exchange). For example, the camera device may include one or more lenses, an image sensor, an image signal processor, or a flash.

[0118] The control device may be a device that controls a camera device, a first shutter illumination device, and a second shutter illumination device. For example, the control device may identify an object from an image received from a combined camera device, and generate first illumination setting information and second illumination setting information based on information about the identified object. Here, the first illumination setting information is information about the illumination setting of the first pattern illumination device, and may include the projection range of the first pattern image, the light amount of the first pattern image, and the color of the first pattern image. The projection range of the first pattern image may include the radius within which the first pattern image is projected. The light amount of the first pattern image is the illuminance of the light source of the first pattern illumination device for projecting the first pattern image, and its unit may be lux. The color of the first pattern image may be represented as an RGB value. The second illumination setting information is information about the illumination setting of the second pattern illumination device, and may include the projection range of the second pattern image, the light amount of the second pattern image, and the color of the second pattern image. The projection range of the second shutter image may include the radius within which the second shutter image is projected. The light amount of the second shutter image is the illuminance of the light source of the second shutter illumination device for projecting the second shutter image, and the unit may be lux. The color of the second pattern image may be represented as an RGB value.

[0119] For example, the control device may use an object extraction model based on a first neural network of images to determine the type, size, and location of the object in real time.

[0120] For example, the control device may determine the type of behavior of the object based on the image, the type of the object, the size of the object, and the location of the object, by using a behavior analysis model of a second neural network.

[0121] The information about the object may include at least one of the type of the object, the size of the object, the location of the object, or the type of behavior of the object.

[0122] The type of the object may include a person, an animal, and a vehicle, and the animal may include types of animals found in urban areas, such as a dog, a cat, or a mouse. The vehicle may include a car, a motorcycle, a bicycle, and a floating board.

[0123] The behavior type of the object can be any one of a variety of behavior types. The multiple behavior types can include behavior types related to criminal activities, behavior types related to violation of public order, behavior types related to emergencies, and other behavior types here. The behavior types related to criminal activities may include attacking others and stealing other people's items. The behavior types related to violation of public order may include illegal littering, smoking, and urinating in public. The third type of actions related to emergencies can include a person suddenly falling down and a person leaning backward. In addition to humans, other behavior types can also include the behaviors of animals and vehicles. That is, other action types can be the action types selected when the object is an animal or a vehicle rather than a person.

[0124] For example, the projection range, light amount, and color of the first pattern image can be determined differently according to the type of the object, the size of the object, and the behavior of the object. Each of the projection range, light amount, and color of the second light-shielding image can be determined differently according to the type of the object, the size of the object, and the type of the behavior.

[0125] For example, the projection range, light amount, and color of the first pattern image according to the type of the object, the size of the object, and the behavior type of the object can be preset.

[0126] For example, according to the type of the object, the size of the object, and the behavior type of the object, the projection range, light amount, and color of the second pattern image can be preset.

[0127] For example, multiple messages formed by the second pattern image can be matched according to the type of the object and the type of the behavior of the object and pre-stored in the control device.

[0128] The first light-shielding illumination device can directly project the first light-shielding image onto the object based on the first illumination setting information. For example, the first light-shielding illumination device can project the first light-shielding image while tracking the position of the object in real time. Here, the first light-shielding image can be a light-shielding image used to emphasize the illumination of the object. The first pattern illumination device is a device equipped with a film or lens of a specific color and projects it onto a wall or floor to ensure high visibility. At this time, multiple films or lenses are provided so that the film or lens installed on the first light-shielding illumination device can be automatically changed according to the type of the object, the size of the object, and the behavior type of the object. For example, the first light-shielding illumination device can include a rotating disk including multiple films or lenses, and by rotating the rotating disk, the film or lens installed on the first light-shielding illumination device can be automatically changed.

[0129] The second light-shielding sheet lighting device can project a second light-shielding sheet image onto the floor of a preset area based on second lighting setting information. For example, the second light-shielding sheet lighting device can project the second light-shielding sheet image at the projection position of the second light-shielding sheet image. The projection position of the second light-shielding sheet image can be determined based on the position of the object and the type of the object's behavior. The second pattern image can include a message determined based on the type of the object and the type of the object's behavior. That is to say, the second light-shielding sheet image can be formed with different messages according to the type of the object and the type of the object's behavior. The second light-shielding sheet lighting device is a device equipped with a film or lens printed with a specific combination of characters or pictures and projects it onto a wall or the floor to ensure high visibility. At this time, a plurality of films or lenses are provided so that the film or lens mounted on the second light-shielding sheet lighting device can be automatically changed according to the type of the object and the type of the object's behavior. For example, the second light-shielding sheet lighting device can include a rotating disk including a plurality of thin films or lenses, and by rotating the rotating disk, the thin film or lens mounted on the second light-shielding sheet lighting device can be automatically changed.

[0130] For example, the pattern lighting system can be configured to further include a user terminal and a server.

[0131] The user terminal can receive information about the object or image information captured by the camera device from the control device through a wireless network. The user terminal can send control information about the camera device, the first pattern lighting device, and the second pattern lighting device to the control device through a wireless network. Through this, the user terminal can remotely control the camera device, the first pattern lighting device, and the second pattern lighting device.

[0132] The server can receive map information around the preset area, setting information of the camera device, and setting information of the first light-shielding sheet lighting device and the second light-shielding sheet lighting device through a wireless network. For example, the server can receive an image captured by the camera device from the control device, determine information about the object, and send the information about the determined object to the control device.

[0133] For example, the control device, the user terminal, and the server can include a processor, a memory, and a communication module.

[0134] The processor may, for example, execute software to control at least one other component (e.g., a hardware or software component) of a device coupled to the processor and may perform various data processing or operations. According to one embodiment, as at least a part of data processing or computing, the processor stores commands or data received from another component (e.g., a communication module) in volatile memory and processes the commands or data stored in the volatile memory. The resultant data may be stored in non-volatile memory. According to one embodiment, the processor is a main processor (e.g., a central processing unit or an application processor) or an auxiliary processor (e.g., a graphics processing unit, a neural processing unit (NPU), an image signal processor, a sensor hub processor, or a communication processor) that may operate independently or together.

[0135] An artificial intelligence model may be created through machine learning. For example, the learning may be performed in the device itself (e.g., a control device) that executes the artificial intelligence model or may be performed by a separate server (e.g., a server). The learning algorithm may include, for example but not limited to, supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning. The artificial intelligence model may include multiple artificial neural network layers. The artificial neural network includes a deep neural network (DNN), a convolutional neural network (CNN), a recurrent neural network (RNN), a restricted Boltzmann machine (RBM), a deep belief network (DBN), a bidirectional recurrent deep neural network (BRDNN), etc. It may be one of the deep Q networks or a combination of two or more of the above networks, but is not limited to the above examples. In addition to the hardware structure, the artificial intelligence model may additionally or alternatively include a software structure.

[0136] The memory may store various data used by at least one component (e.g., a processor) of the device. The data may include, for example, input data or output data of software and instructions related thereto. The memory may include volatile memory or non-volatile memory. For example, the memory may be at least one type of storage medium such as a flash type, a hard disk type, a multimedia card micro type, a card type memory (e.g., SD or XD memory, etc.), a RAM (random access memory), an SRAM (static random access memory), a ROM (read only memory), an EEPROM (electrically erasable programmable read only memory), a PROM (programmable read only memory), a magnetic memory, a magnetic disk, and an optical disk.

[0137] The communication module can support establishing a direct (e.g., wired) or wireless communication channel between the device and an external device and performing communication through the established communication channel. The communication module operates independently of the processor and can include one or more communication processors that support direct (e.g., wired) communication or wireless communication. According to one embodiment, the communication module is a wireless communication module (e.g., a cellular communication module, a short-range wireless communication module, or a GNSS (Global Navigation Satellite System) communication module) or a wired communication module (e.g., a wireless communication module). LAN (Local Area Network) communication module), or a power line communication module). Among these communication modules, the corresponding communication module is a first network (e.g., a short-range communication network such as Bluetooth, WiFi (Wireless Fidelity) Direct or IrDA (Infrared Data Association)) or a second network (e.g., a traditional network). A cellular network, a 5G network, which can communicate with an external device through a telecommunications network such as a next-generation communication network, the Internet, or a computer network (e.g., LAN or WAN). These various types of communication modules can be integrated into one component (e.g., a single chip) or can be implemented as multiple separate components (e.g., multiple chips).

[0138] The object navigation model can include a YOLO (You Only Look Once)-based model. The object extraction model can use an algorithm that simultaneously performs bounding box coordinates and classification. At this time, the first neural network used in the object search model includes a confidence loss function indicating the probability of object existence, a localization loss function for adjusting the position of the bounding box, and a localization loss function for predicting the object class. The class loss function can be learned by minimizing it.

[0139] The control device can generate a plurality of image vectors by performing first data preprocessing on the images captured from the camera device. The image vectors may contain pixel values of video still images. The plurality of image vectors can be the image vectors of each still image for each time period of the video.

[0140] For example, the control device can determine the type, size, and position of the object by inputting the plurality of image vectors into the object search model.

[0141] The object extraction model consists of a backbone network and a detection network. The backbone network may include different numbers of convolutional layers and fully connected (FC) layers depending on the model size. At this time, the backbone network can perform efficient feature extraction and gradient propagation by introducing CSP (Cross Stage Partial Connections). The model sizes of the backbone network can include small, medium, large, and extra-large. Depending on the model size, the backbone network may include multiple convolutional layers and multiple FC layers. For example, the larger the model size, the more convolutional layers and FC layers can be used. The convolutional layers at this time are the intersection of 1X1 convolutional layers that control the interaction between channels and 3X3 convolutional layers that learn spatial patterns, which can reduce the model complexity and improve performance.

[0142] For example, the model size of the backbone network including the number of convolutional layers and the number of FC layers can be determined based on the total number of image vectors input to the backbone network and the resolution of each image vector.

[0143] The backbone network can learn the shape, structure, and main features of an object through the process of extracting low-level visual features from each image vector and converting the visual features into high-level semantic information. In addition, the backbone network can detect various sizes and positions of objects by extracting features of various sizes and levels through multiple layers. In addition, the backbone network can partially exchange features by dividing the input data into two parts using a CSP block, processing them, performing other operations in the middle, and then combining the two parts again.

[0144] For example, when the type of the object is a preset type, the control device estimates the skeleton of the object by preprocessing the data and the position of the object on the image, and provides a skeleton vector that can create distribution values including each time interval of the values of multiple two-dimensional coordinates and the angles between multiple two-dimensional coordinates. The values of multiple two-dimensional coordinates for each time interval for estimating the object skeleton may include vertices that form the edges of the object in each image based on the position of the object in each image. The distribution values of the angles between multiple two-dimensional coordinates for each time period of multiple images may only include the distribution of vertices in which the angles between two line segments adjacent to the vertex in each of the multiple images are greater than or equal to a critical angle.

[0145] For example, if the category value corresponds to a preset category value, the control device can generate a skeleton vector by preprocessing the data of the image and the position of the object. Here, the preset category value can be a value for an individual.

[0146] For example, if the category value does not correspond to a preset category value, the control device can determine that the behavior type of the object is another behavior type.

[0147] For example, the control device may generate an object vector including an object size value. The value of the size of the object may be a value proportional to the area occupied by the object in the image. For example, the value of the size of the object may be a value estimated by applying a preset ratio to the area occupied by the object in the image. The preset ratio may be the ratio between the image and a preset area.

[0148] For example, the control device may determine the behavior type of an object by inputting the skeleton vector and the object vector into a behavior analysis model. The behavior analysis model may be learned based on multiple skeleton vectors, multiple object vectors, multiple reference skeleton vectors, and multiple correct behavior types.

Claims

1. A pattern lighting system, wherein, the pattern lighting system comprises: a camera device for capturing an image of a preset area onto which pattern light is projected; a control device connected to the camera device, identifying an object from the image, and generating first lighting setting information and second lighting setting information based on information about the identified object; a first pattern lighting device for directly projecting a first pattern image onto an object based on the first lighting setting information including the projection range of the first pattern image, the light amount of the first pattern image, and the color of the first pattern image; and a second pattern lighting device for projecting a second pattern image onto the floor of the preset area based on the second lighting setting information including the projection range of the second pattern image, the light amount of the second pattern image, and the color of the second pattern image, the control device determines the type, size, and position of the object in real time based on the image by using an object extraction model of a first neural network, uses a second neural network to determine the behavior type of the object through a behavior analysis model based on the image, the type of the object, the size of the object, and the position of the object, the information of the object includes the type of the object, the size of the object, and the behavior type of the object, the first pattern lighting device projects the first pattern image while tracking the position of the object in real time, the second pattern lighting device projects the second pattern image at the projection position of the second pattern image, the projection position of the second pattern image is determined according to the position of the object and the behavior type of the object, the second pattern image includes a message determined based on the type of the object and the behavior type of the object.

Images