KR102695996B1 - Snow and ice melting device based on freezing prediction using artificial intelligence - Google Patents

Abstract

The present invention relates to a snow and ice melting device based on ice formation prediction using artificial intelligence, comprising: a processor including one or more cores; and a memory; wherein the processor comprises: a receiving unit which receives a road image captured by a monitoring camera and an image of a road surrounding environment, an atmospheric temperature and humidity measured by an atmospheric temperature and humidity measuring sensor, and a road surface temperature measured by a road surface temperature sensor; a first prediction unit which inputs the road surrounding environment image received by the receiving unit into a learned first prediction model and outputs a probability of rain, snowfall, fog, and clear weather; a second prediction unit which inputs the road image, atmospheric temperature and humidity, and road surface temperature received by the receiving unit into a learned second prediction model and outputs a first probability of ice formation; and a third prediction unit which multiplies the probability of rain, snowfall, fog, and clear weather output by the first prediction unit and the first probability of ice formation output by the second prediction unit by weights, and then adds them together to output a second probability of ice formation.

Description

{SNOW AND ICE MELTING DEVICE BASED ON FREEZING PREDICTION USING ARTIFICIAL INTELLIGENCE}

The present invention relates to a snow and ice melting device, and more particularly, to a snow and ice melting device based on ice prediction that uses artificial intelligence to predict the degree or presence of ice on a road and automatically sprays salt water.

In winter, when snow or rain accumulates on the road surface or the road surface freezes, vehicles may have difficulty controlling the vehicle or may slip, which may cause traffic accidents. In addition, slow-speed driving to prevent accidents may delay vehicle traffic. In such cases, when snow accumulates on the road or the road freezes, salt water can be sprayed to prevent or resolve road freezing.

However, due to the quantitative and distance limitations of snow removal equipment, if the road ice is not removed in a timely manner, the meaning of snow removal may fade as accidents due to ice cannot be prevented. Therefore, a method of installing and operating salt water spraying devices in vulnerable snow removal sections of the road has been recently adopted.

In this regard, Korean Patent No. 10-2547008 discloses an intelligent snowfall detection device and an automatic snow removal solution spraying system including the same. The above-mentioned prior art document combines a snowfall detection device capable of automatically detecting snowfall with an automatic snow removal solution spraying system, so that when snow begins to fall, the snow removal solution spraying system is automatically operated to spray snow removal solution on the road, thereby preventing traffic accidents caused by freezing. The snowfall detection device can not only immediately detect the snowfall condition as soon as it starts to snow, but can also detect whether snow is falling and piling up or whether rain is falling but not snowing and causing the road to freeze, and can even detect the formation of black ice in hazardous areas such as bridges.

As above, many inventions have been proposed to automatically detect whether there is freezing and spray salt water, but there is no device that sprays salt water in advance before snow accumulates or the road surface freezes, taking into account the probability of snowfall or rain and the probability of freezing.

Accordingly, there is a demand for an AI-based snow and ice melting device that can prevent freezing in advance by considering the probability of freezing on the road surface. In addition, there is a demand for an AI-based snow and ice melting system that can prevent freezing on a wide area of roads by automatically spraying salt water on roads adjacent to roads where freezing has occurred or is highly likely to occur.

Korean Patent Registration No. 10-2547008

The purpose of the present invention is to provide a snow and ice melting device based on artificial intelligence-based ice prediction, which sprays salt water in advance before snow accumulates or a road surface freezes, taking into account the probability of snowfall or rain, the probability of ice formation, etc.

In addition, the present invention has another purpose of providing an artificial intelligence-based snow and ice melting system that can prevent freezing on a wide area of roads by automatically spraying salt water on roads adjacent to roads where freezing occurs or is likely to occur.

In order to achieve the above object, the present invention is a snow and ice melting device based on ice formation prediction using artificial intelligence, comprising: a processor including one or more cores; and a memory; wherein the processor comprises: a receiving unit which receives a road image captured by a monitoring camera and an image of the road surrounding environment, an atmospheric temperature and humidity measured by an atmospheric temperature and humidity measuring sensor, and a road surface temperature measured by a road surface temperature sensor; a first prediction unit which inputs the road surrounding environment image received by the receiving unit into a learned first prediction model and outputs a probability of rain, snowfall, fog, and clear weather; a second prediction unit which inputs the road image, atmospheric temperature and humidity, and road surface temperature received by the receiving unit into a learned second prediction model and outputs a first probability of ice formation; and a third prediction unit which multiplies the probability of rain, snowfall, fog, and clear weather output by the first prediction unit and the first probability of ice formation output by the second prediction unit by weights, and then adds them together to output a second probability of ice formation.

Preferably, the processor may further include a control unit that controls the salt water pipe and the injection pump to spray salt water on the road when the second freezing probability output by the third prediction unit exceeds the first threshold value, and controls the salt water pipe and the injection pump to not spray salt water on the road when the second freezing probability output by the third prediction unit is less than or equal to the first threshold value.

Preferably, the receiving unit further receives a real-time weather forecast or information on whether an accident due to ice has occurred at the location where the road is located, and the control unit can adjust the first threshold value according to the real-time weather forecast received by the receiving unit or the information on whether an accident due to ice has occurred at the location where the road is located.

Preferably, the control unit controls the injection pump so that salt water is continuously sprayed onto the road when the second freezing probability output by the third prediction unit exceeds the second threshold value, and controls the brine pipe so that the time for which salt water is sprayed onto the road is reduced while maintaining the output of the injection pump when the second freezing probability output by the third prediction unit exceeds the first threshold value and is lower than or equal to the second threshold value.

Preferably, the receiving unit further receives information on the remaining amount of brine, and the control unit, if the remaining amount of brine is below a certain level, can control the brine pipe to increase the first threshold value or reduce the time for which brine is sprayed onto the road while maintaining the output of the spray pump.

Preferably, the receiving unit further receives a real-time weather forecast or information on whether an accident due to ice has occurred at a location where the road is located, and the processor may further include a weight adjusting unit that adjusts each of the weights according to the real-time weather forecast received by the receiving unit or the information on whether an accident due to ice has occurred at a location where the road is located.

Preferably, the first prediction model may be a deep learning model using compound scaling of MobileNetV2 and EfficientNet, and the second prediction model may be a linear regression model.

Preferably, the apparatus may further include: a communication unit; a power unit; an emergency power unit; a connection unit connected to a plurality of sensors to receive information from the plurality of sensors; a display unit configured as a touch panel to output information measured by the plurality of sensors, such as rain, snow, fog and clear probability, and first and second freezing probabilities, and to control the control unit; and a cover unit protecting the communication unit, the power unit, the emergency power unit, the display unit, the processor, and the memory.

In addition, the present invention provides an artificial intelligence-based snow and ice melting system based on ice formation prediction, which comprises a plurality of edge-based computers installed to control salt water spraying at each predetermined section of a road, and an integrated control unit controlling the plurality of edge-based computers, wherein the edge-based computer comprises a processor including one or more cores and a memory, and the processor comprises: a receiving unit which receives a road image captured by a monitoring camera, an image of the road surrounding environment, temperature and humidity of the air measured by an air temperature and humidity measuring sensor, and road temperature measured by a road surface temperature sensor; a first prediction unit which inputs the image of the road surrounding environment received by the receiving unit into a learned first prediction model and outputs a probability of rain, snowfall, fog, and clear weather; a second prediction unit which inputs the road image, temperature and humidity of the air, and road temperature received by the receiving unit into a learned second prediction model and outputs a first probability of ice formation; A third prediction unit multiplies the rainfall, snowfall, fog, and clear probability output by the first prediction unit, and the first freezing probability output by the second prediction unit by a weight, and then adds them to output a second freezing probability; and a control unit controls a salt water pipe and a spray pump to spray salt water onto a road if the second freezing probability output by the third prediction unit exceeds a first threshold value, and the integrated control unit is characterized in that if at least one third prediction unit among the plurality of edge-based computers outputs the second freezing probability exceeding the first threshold value, it controls an adjacent edge-based computer to spray salt water onto an adjacent road.

The present invention has the advantage of drastically reducing traffic accidents caused by freezing by preventing freezing in advance by considering the probability of freezing on the road surface.

In addition, the present invention has an advantage in that it can obtain a highly accurate freezing probability by extracting the freezing probability using artificial intelligence learned from road surrounding environment images, road images, air temperature, air humidity, and road surface temperature.

In addition, the present invention has an advantage in that it can prevent freezing on a wide area of roads by automatically spraying salt water on roads adjacent to roads where freezing occurs or is likely to occur.

Figure 1 shows a partial configuration diagram of a snow and ice melting device based on artificial intelligence-based freezing prediction according to an embodiment of the present invention.
Figure 2 shows a configuration diagram of a salt spraying device according to an embodiment of the present invention.
Figures 3 and 4 illustrate the overall configuration of a snow and ice melting device based on artificial intelligence-based freezing prediction according to an embodiment of the present invention.
Figure 5 shows a configuration diagram of a processor according to an embodiment of the present invention.
FIG. 6 is a flowchart illustrating the overall function of a processor according to an embodiment of the present invention.
FIG. 7 illustrates a flowchart for explaining freezing prediction of a processor according to an embodiment of the present invention.
FIG. 8 is a drawing for explaining the function of a control unit that controls salt water injection according to a first or second critical point according to an embodiment of the present invention.
Figure 9 shows a configuration diagram of a snow and ice melting system based on freezing prediction using artificial intelligence according to an embodiment of the present invention.

Hereinafter, the present invention will be described in detail with reference to the contents described in the attached drawings. However, the present invention is not limited or restricted by the exemplary embodiments. The same reference numerals presented in each drawing represent components that perform substantially the same function.

The purpose and effect of the present invention can be naturally understood or made clearer by the following description, and the purpose and effect of the present invention are not limited to the following description alone. In addition, when explaining the present invention, if it is judged that a detailed description of a known technology related to the present invention may unnecessarily obscure the gist of the present invention, the detailed description will be omitted.

The terminology used herein is only used to describe specific embodiments and is not intended to limit the present invention. The singular expression includes the plural expression unless the context clearly indicates otherwise. In this application, the terms "comprises" or "has" and the like are intended to specify the presence of a feature, number, step, operation, component, part or combination thereof described in the description of the invention, but should be understood to not exclude in advance the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts or combinations thereof.

The terms first, second, etc. may be used to describe various components, but the components should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be referred to as the second component, and similarly, the second component may also be referred to as the first component.

Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Terms defined in commonly used dictionaries, such as those defined in common dictionaries, should be interpreted as having a meaning consistent with the meaning they have in the context of the relevant art, and shall not be interpreted in an idealized or overly formal sense unless expressly defined herein.

In interpreting the components, even if there is no separate explicit description, it is interpreted as including the error range. In the case of a description of a temporal relationship, for example, if the temporal continuity is described as 'after', 'following', 'next to', 'before', etc., it also includes cases where it is not continuous unless 'right away' or 'directly' is used.

Hereinafter, the technical configuration of the present invention will be described in detail with reference to the attached drawings.

FIG. 1 shows a partial configuration diagram of a snow and ice melting device (100) based on ice formation prediction using artificial intelligence according to an embodiment of the present invention. Referring to FIG. 1, the snow and ice melting device (100) based on ice formation prediction may include a processor (110), a memory (120), and a network (130). The snow and ice melting device (100) based on ice formation prediction may collect data for predicting ice formation on a road through the network (130), and the collected data may be processed within the processor (110) to calculate an ice formation probability.

The snow and ice melting device (100) based on freezing prediction can drastically reduce traffic accidents caused by freezing by preventing freezing in advance by considering the probability of freezing on the road surface. The snow and ice melting device (100) based on freezing prediction can obtain a highly accurate freezing probability by extracting the freezing probability using artificial intelligence learned from road surrounding environment images, road images, air temperature, air humidity, and road surface temperature.

Fig. 2 shows a configuration diagram of a salt water spraying device according to an embodiment of the present invention. The salt water spraying device may be installed at each section of a road. Referring to Fig. 2, the salt water spraying device may include a monitoring camera, an air temperature and humidity measuring sensor, a road surface temperature sensor, a snow and ice melting device (100) based on freezing prediction, a salt water remaining amount confirmation sensor, salt water pipes (first salt water pipe, second salt water pipe), a spray pump, and a pressure measuring sensor.

The monitoring camera is composed of a PTZ dome camera, which can perform omnidirectional monitoring. The PTZ dome camera can mean an integrated camera that can be electrically rotated (PAN), tilted vertically, and zoomed using a fixed motor. The PTZ dome camera can be moved manually or automatically at high speed, and can view a wide area with a single camera, and can monitor all directions, minimizing blind spots.

The monitoring camera can monitor the road and the surrounding environment by shooting images in real time. The monitoring camera can perform smooth monitoring even at night by using an ultra-low light lens. The monitoring camera can shoot images of the road and the surrounding environment and transmit them to the first prediction unit or the second prediction unit. The images of the road and the surrounding environment shot by the monitoring camera can be used to predict weather such as rain, snow, fog, and clear weather, and whether the road is icy.

The atmospheric temperature and humidity measurement sensor can measure the temperature and humidity of the atmosphere and transmit them to the snow and ice melting device (100) based on the freezing prediction to provide information for freezing prediction. The atmospheric temperature and humidity measurement sensor is installed outdoors and is installed at a certain height from the ground so as to accurately measure the temperature and humidity of the atmosphere.

The road surface temperature sensor can be installed above or below the road surface to measure the temperature of the road. The road surface temperature sensor can measure the temperature of the road surface and transmit it to the ice prediction-based snow and ice melting device (100) to provide information for ice prediction.

The snow and ice melting device (100) based on freezing prediction, as described below, receives various data from installed sensors and inputs the received data into learned artificial intelligence to predict the probability of a road freezing. In addition, the snow and ice melting device (100) based on freezing prediction can remotely monitor data measured from various sensors, freezing probability, etc., and provide an environment in which various sensors and devices (injection pumps, salt water pipes, etc.) can be controlled.

The brine remaining amount confirmation sensor may be a non-contact water level gauge, and may measure the brine remaining amount in real time and generate an alarm and warning signal (electrical signal) when the water level falls below a certain level. The brine remaining amount confirmation sensor may provide the measured brine remaining amount (brine level) or warning signal (electrical signal) to the snow and ice melting device (100) based on the freezing prediction.

The brine pipe can be divided into a first brine pipe and a second brine pipe. The first brine pipe can be arranged between the brine tank and the injection pump, and the second brine pipe can be arranged between the injection pump and the injection nozzle. The brine pipe can be controlled by a snow and ice melting device (100) based on freezing prediction. The brine pipe can be controlled on/off, and the amount of brine spray can be adjusted by adjusting the on/off time.

The injection pump can generate pressure so that the salt water can be injected. The injection pump can be composed of an inverter. The injection pump can be controlled by a snow and ice melting device (100) based on freezing prediction. When the output of the inverter of the injection pump is adjusted to control the amount of salt water injection, if the output of the inverter is lowered, the amount of salt water injection is reduced, but the pressure generated is lowered, so the range of salt water injection can be reduced. Therefore, the present invention can control the on/off of the salt water pipe while maintaining the inverter output of the injection pump as it is in order to reduce the amount of salt water injection.

The pressure measurement sensor can measure the pressure when the brine is discharged through the injection nozzle. The pressure measurement sensor can detect malfunctions in each section, such as the injection nozzle/salt water pipe valve/check valve, depending on the discharge pressure. Therefore, the manager can check the measurement results of the pressure measurement sensor remotely and easily check whether a malfunction has occurred and the section where the malfunction occurred.

Figures 3 and 4 illustrate the overall configuration of a snow and ice melting device (100) based on artificial intelligence-based freezing prediction according to an embodiment of the present invention. The snow and ice melting device (100) based on freezing prediction may further include a communication unit (140), a power unit (150), an emergency power unit (160), a connection unit (170), a display unit (180), and a cover unit (190).

Figure 3 shows the front appearance of a snow and ice melting device (100) based on freezing prediction. A display unit (180) may be placed in the center of the front, a communication unit (140) may be placed in the lower left, a power unit (150) may be placed in the lower right, a connection unit (170) may be placed in the lower center, a power unit (150) may be placed in the upper right, and an emergency power unit (160) may be placed in the center right.

The communication unit (140) may be equipped with a communication port so that it can be connected to a network and a camera by wire or wirelessly. The communication unit (140) may be equipped with one or more communication ports, and may have a built-in switch so that each communication port can be switched.

The communication unit (140) can enable the snow and ice melting device (100) based on the freezing prediction to operate in a networked environment via logical connections via wired and/or wireless communications. The logical connections include wired/wireless connections to a local area network (LAN) and/or a larger network, for example, a wide area network (WAN). Such LAN and WAN networking environments are common in offices and companies and facilitate enterprise-wide computer networks, such as intranets, all of which can be connected to a worldwide computer network, for example, the Internet.

When used in a LAN networking environment, the communication unit (140) may connect the ice melting device (100) based on ice prediction to a local network via a wired and/or wireless communication network interface or adapter. The adapter may facilitate wired or wireless communications to the LAN, which may also include a wireless access point installed therein for communicating with the wireless adapter. When used in a WAN networking environment, the communication unit (140) may include a modem, be connected to a communication computing device on the WAN, or have other means for establishing communications over the WAN, such as via the Internet. The modem, which may be built-in or external and may be wired or wireless, is connected to the system bus via a serial port interface.

The communication unit (140) operates to communicate with any wireless device or object that is arranged and operates by wireless communication, such as a printer, a scanner, a sensor, a desktop and/or portable computer, a portable data assistant (PDA), a communication satellite, any equipment or location associated with a wireless detectable tag, and a telephone. This includes at least Wi-Fi and Bluetooth wireless technologies. Accordingly, the communication may be a predefined structure as in a conventional network, or may simply be an ad hoc communication between at least two devices.

Wi-Fi (Wireless Fidelity) allows you to connect to the Internet and other things without wires. Wi-Fi is a wireless technology that allows devices, such as computers, to send and receive data, indoors and outdoors, anywhere within the coverage area of a base station, similar to a cell phone. Wi-Fi networks use a wireless technology called IEEE 802.11 (a, b, g, etc.) to provide secure, reliable, and high-speed wireless connections. Wi-Fi can be used to connect computers to each other, to the Internet, and to wired networks (using IEEE 802.3 or Ethernet). Wi-Fi networks can operate in the unlicensed 2.4 and 5 GHz radio bands, at data rates of, for example, 11 Mbps (802.11a) or 54 Mbps (802.11b), or in products that include both bands (dual band).

The power supply unit (150) can be set to remote control, manual control, or stop. Here, stop acts as a power on/off function. When the switch of the power supply unit (150) is directed to stop, the power is off, and when the switch is directed to a location other than stop, the power is on. When the power supply unit (150) is on, it can be operated by either remote or manual control. Remote control means controlling from the outside through remote communication, and manual control means controlling through the touch panel of the display unit (180).

The emergency power supply (160) may be configured as an emergency stop switch, and can urgently turn off the power to the device when a problem occurs in the control of a device such as a salt water pipe or a spray pump.

The display unit (180) can be configured as a touch panel to output information measured from multiple sensors, such as rain, snow, fog and clear probability, and first and second freezing probabilities, and to control the control unit.

The connection unit (170) can be connected to a plurality of sensors to receive information from the plurality of sensors. That is, the connection unit (170) is a place where various sensors and the snow and ice melting device (100) based on freezing prediction are connected, and can be connected by wire through a connection port or connected wirelessly through the communication unit (140).

The cover part (190) can protect the communication part (140), the power part (150), the emergency power part (160), the connection part (170), the display part (180), the processor (110), and the memory (120). That is, the cover part (190) can serve as a housing. The cover part (190) can be formed by combining two parts divided into a front part and a back part, and the front part and the back part can be connected by one or more hinges (191). Therefore, the cover part (190) can be an upper and lower open type cover by the hinge (191).

Fig. 4 shows the inside of a snow and ice melting device (100) based on freezing prediction. A main control board having a processor (110), a memory (120), or a network (130) embedded therein may be placed in a section corresponding to the display section (180). An LTE router/shared device/switch, etc. constituting a communication section (140) may be included inside, and a network status check window may be placed on the left side of the display section (180) so that an administrator can easily check the network status. The connection section (170) may be equipped with an interface board, and DC power for relay coil operation may be directly supplied to the interface board, and I/O control is possible via RS232/485.

Fig. 5 shows a configuration diagram of a processor (110) according to an embodiment of the present invention. Referring to Fig. 5, the processor (110) may include a receiving unit (111), a first prediction unit (112), a second prediction unit (113), a third prediction unit (114), a control unit (115), and a weight adjustment unit (116).

FIG. 6 is a flowchart illustrating the entire function of the processor (110) according to an embodiment of the present invention. Referring to FIG. 6, the processor (110) can predict the probability of ice formation on the road using air temperature, air humidity, AI image analysis results, etc. The processor (110) can store information so that the captured real-time road image, road surrounding environment image, log record, and warning alarm (integrated control center or person in charge smartphone) can be remotely checked. The processor (110) can predict the probability of ice formation on the road and generate a control signal to spray salt water on the road.

FIG. 7 shows a flowchart for explaining the freezing prediction of the processor (110) according to an embodiment of the present invention. In the processor (110), an image of the surrounding environment of a road may be input into a virtual image classification model (a first prediction model) to output the probability of rain, snow, fog, and clear weather, and a preprocessed (normalized) road image, air temperature and humidity, and road surface temperature may be input into a linear regression model (a second prediction model) to output the first freezing probability. The processor (110) may weight-combine the outputs of the virtual image classification model and the linear regression model to output the final freezing probability.

The receiving unit (111) can receive road images and road surrounding environment images captured by the monitoring camera, air temperature and humidity measured by the air temperature and humidity measuring sensor, and road surface temperature measured by the road surface temperature sensor. The receiving unit (111) can collect data from various sensors through the connection unit (170).

The receiving unit (111) can further receive real-time weather forecast or information on whether an accident due to freezing has occurred at the location where the road is located. The receiving unit (111) can further receive information on the remaining amount of salt water. The receiving unit (111) can receive real-time weather forecast using an open API provided by the Korea Meteorological Administration. The receiving unit (111) can receive information such as GPS coordinates of accidents due to freezing provided by the Korea Meteorological Administration ( https://taas.koroad.or.kr/api/selectFreezingDataSet.do ).

The first prediction unit (112) can input the road surrounding environment image received by the receiver (111) into the learned first prediction model to output the probability of rain, snow, fog, and clear weather. For example, the first prediction unit (112) can output the probability of rain, snow, fog, and clear weather, such as a probability of rain of 12%, a probability of snow of 46%, a probability of fog of 23%, or a probability of clear weather of 7%. The first prediction unit (112) can output the probability of rain, snow, fog, and clear weather in real time by using the road surrounding environment image captured in real time and the learned first prediction model.

The first prediction model may be a weather image classification model that has advantages in real-time, accuracy, or lightness so that real-time analysis can be possible on edge-based hardware. Preferably, the first prediction model may be a deep learning model using compound scaling of MobileNetV2 and EfficientNet. Therefore, the first prediction model can minimize the amount of computation and enable efficient model expansion.

The first prediction model can be trained with weather images and rain, snow, fog, and clear probability information provided by the Korea Meteorological Administration. The first prediction model can be trained with weather images and weather situation information after a certain time provided by the Korea Meteorological Administration. The first prediction model can be trained with training data generated using a diffusion model. The first prediction model can be trained with images recorded at an actual site and weather situation information after a certain time.

The second prediction unit (113) can input the road image, the air temperature and humidity, and the road surface temperature received by the receiver (111) into the learned second prediction model and output the first freezing probability. For example, the second prediction unit (113) can input the road image in which the road is wet, the air temperature of -1 degree, the air humidity of 95%, and the road surface temperature of 0 degree into the second prediction model and output the first freezing probability of 78%. The second prediction model can be a linear regression model.

The second prediction model can be trained using information provided by the Korea Meteorological Administration (road images, atmospheric temperature and humidity, road surface temperature, etc.), training data generated using a diffusion model, or directly measured data.

The third prediction unit (114) can output the second freezing probability by multiplying the rainfall, snowfall, fog, and clear probability output by the first prediction unit (112) and the first freezing probability output by the second prediction unit (113) by their respective weights and then adding them up. For example, the third prediction unit (114) can calculate the second freezing probability using the following mathematical expression 4.

[Mathematical formula 1]

Here, w ₁ to w ₅ are weights, a ₁ is the probability of rain, a ₂ is the probability of snow, a ₃ is the probability of fog, a ₄ is the probability of clear weather, and b ₁ is the probability of first freezing.

The third prediction unit (114) can set the weights to preset fixed values or use an artificial intelligence model to track the optimal weights. The third prediction unit (114) can variably adjust the weights in consideration of weather forecasts or whether the area is prone to freezing. For example, if snowfall is forecast, the weights of w ₂ or w ₅ can be adjusted upward, and the weights of w ₁ , w ₃ , or w ₄ can be adjusted downward.

The control unit (115) may control the brine pipe and the injection pump to spray brine on the road if the second freezing probability output by the third prediction unit (114) exceeds the first threshold value, and may control the brine pipe and the injection pump to prevent salt water from being sprayed on the road if the second freezing probability output by the third prediction unit (114) is lower than or equal to the first threshold value. For example, the control unit (115) may open the brine pipe and operate the injection pump to spray salt water on the road if the first threshold value is 80% and the second freezing probability output by the third prediction unit (114) is 82%. Conversely, the control unit (115) may close the brine pipe and stop the injection pump to prevent salt water from being sprayed on the road if the first threshold value is 80% and the second freezing probability output by the third prediction unit (114) is 67%.

The control unit (115) can adjust the first threshold value according to the real-time weather forecast received by the receiver (111) or information on whether an accident due to freezing has occurred in the area where the road is located. For example, the control unit (115) can adjust the first threshold value upward when the real-time weather forecast predicts snowfall or an accident due to freezing has occurred in the area. Accordingly, the control unit (115) can comprehensively consider not only the second freezing probability output by the artificial intelligence but also the real-time weather forecast and whether an accident due to freezing has occurred when spraying salt water. Accordingly, the control unit (115) can save salt water by spraying salt water at the exact time when salt water spraying is necessary, and can minimize damage to the road and the underbody of the vehicle.

FIG. 8 is a diagram illustrating a function of a control unit (115) that controls salt water spraying according to a first or second threshold point according to an embodiment of the present invention. Referring to FIG. 8, if the second freezing probability output by the third prediction unit (114) exceeds the second threshold value, the control unit (115) controls the spray pump so that salt water is continuously sprayed onto the road, and if the second freezing probability output by the third prediction unit (114) exceeds the first threshold value and is lower than or equal to the second threshold value, the control unit (115) can control the salt water pipe so that the time for which salt water is sprayed onto the road is reduced while the output of the spray pump is maintained. Here, the second threshold value may be greater than the first threshold value.

The control unit (115) should set the salt water spray time to the maximum in the same way as in a freezing situation or a snowfall situation in order to prevent accidents caused by freezing in advance when the freezing probability is high. However, in the case where the freezing probability is not high, the salt water spray time should be adjusted between the minimum and maximum times to the extent that freezing can be prevented in order to save salt water. If the salt water spray time is set to the maximum even in the case where the freezing probability is not high, there may be insufficient salt water to be used when actual snowfall occurs or the temperature becomes very low and the freezing condition worsens. In particular, in the case where freezing occurs simultaneously in a midwinter snowfall situation, the amount of salt water used increases and it may take a long time to replenish it, so the strategic use of salt water is a very important issue. Therefore, the control unit (115) can adjust the salt water spray time to the minimum and maximum times between the first and second threshold values, which are not high freezing probability, and maintain the salt water spray time at the maximum in the case where the freezing probability exceeds the second threshold value.

The control unit (115) can use various methods to control the amount of salt water sprayed, and representative methods include a method of controlling the inverter output of the spray pump and a method of controlling the on/off of the salt water pipe. The method of controlling the inverter output of the spray pump while maintaining the on state of the salt water pipe has a problem in that when the inverter output goes down, the radius of the salt water spray decreases, and the range in which freezing is prevented or relieved by the salt water is reduced. Therefore, the control unit (115) according to the embodiment of the present invention can control the on/off of the salt water pipe so that the time for which salt water is sprayed on the road is reduced while maintaining the output of the spray pump in controlling the amount of salt water sprayed.

The control unit (115) can control the brine pipe to increase the first threshold value if the remaining amount of brine is below a certain level, or to reduce the time for which brine is sprayed on the road while the output of the injection pump is maintained. If the remaining amount of brine is insufficient, the control unit (115) needs to adjust the amount of brine sprayed until the brine is stably replenished. Therefore, the control unit (115) can increase the first threshold value so that brine is sprayed only in situations where the probability of freezing is high. Alternatively, the control unit (115) can control the brine pipe to reduce the time for which brine is sprayed on the road while the output of the injection pump is maintained, thereby saving brine by reducing the amount of brine sprayed.

The weight adjustment unit (116) can adjust each weight according to the real-time weather forecast received by the receiver (111) or information on whether an accident due to freezing has occurred at the location of the road. The five factors that affect the weights are the probability of rainfall, the probability of snowfall, the probability of fog, the probability of clear weather, and the first probability of freezing. If the probability of rainfall or snowfall is high in the real-time weather forecast, the weights of the probability of rainfall, snowfall, and/or the first probability of freezing can be increased. If the probability of clear weather is high in the real-time weather forecast, the probability of rainfall, snowfall, and/or the first probability of freezing can be lowered or the weight of the probability of clear weather can be increased. In this way, information from the real-time weather forecast of the Korea Meteorological Administration can also be used to predict the probability of the second probability of freezing.

In addition, the weight adjustment unit (116) can increase the weight of the first freezing probability if the road is a road where an accident due to freezing has occurred based on information on whether an accident due to freezing has occurred at the location of the road. Through this, information on whether an accident due to freezing has occurred at the location of the road can also be used to predict the second freezing probability.

Figure 9 shows a configuration diagram of a snow and ice melting system (1000) based on ice prediction using artificial intelligence according to another embodiment of the present invention. Referring to Figure 9, the snow and ice melting system (100) based on ice prediction may be equipped with a plurality of edge-based computers (100) installed to control salt water spraying at each section of a road, and an integrated control unit (200) that controls the plurality of edge-based computers (100).

The edge-based computer (100) refers to the aforementioned ice melting device (100) based on freezing prediction. That is, the edge-based computer (100) may include a processor (110) including one or more cores and a memory (120). The processor (110) of the edge-based computer (100) includes a receiving unit (111) that receives the road image captured by the monitoring camera and the road surrounding environment image, the atmospheric temperature and humidity measured by the atmospheric temperature and humidity measuring sensor, and the road surface temperature measured by the road surface temperature sensor, a first prediction unit (112) that inputs the road surrounding environment image received by the receiving unit (111) into a learned first prediction model and outputs the probability of rain, snow, fog, and clear weather, a second prediction unit (113) that inputs the road image, atmospheric temperature and humidity, and road surface temperature received by the receiving unit (111) into a learned second prediction model and outputs the first freezing probability, a third prediction unit (114) that multiplies the rain, snow, fog, and clear weather probability output by the first prediction unit (112) and the first freezing probability output by the second prediction unit (113) by their respective weights and then adds them to output the second freezing probability, and a third prediction unit (114) that outputs the second freezing probability. If the second freezing probability output by the prediction unit (114) exceeds the first threshold value, a control unit (115) that controls the salt water pipe and spray pump so that salt water is sprayed on the road may be included.

The integrated control unit (200) can control the adjacent edge-based computer to spray salt water on the adjacent road when at least one third prediction unit among the plurality of edge-based computers outputs a second freezing probability exceeding a first threshold value. When the probability of freezing on a specific road is output as high, the integrated control unit (200) can control the control unit of the adjacent edge-based computer to spray salt water not only on the road in question but also on the adjacent road. For example, when the edge-based computer (110c) of FIG. 9 outputs a second freezing probability exceeding the first threshold value, the integrated control unit (200) can control the control units of the adjacent edge-based computers 100c and 100d or 100a, 100b, 100d, and 100e to spray salt water. Through this, when freezing occurs on a specific road, the probability of freezing occurring on the adjacent road is very high, so that accidents due to road freezing can be effectively prevented.

The integrated control unit (200) can control the control unit of the edge-based computer that outputs the second freezing probability exceeding the first threshold value to set the salt water spray time to the maximum, and can control the control unit of the adjacent edge-based computer to adjust the salt water spray time between the minimum and maximum times.

Although the present invention has been described in detail through representative examples above, those skilled in the art will understand that various modifications can be made to the above-described embodiments without departing from the scope of the present invention. Therefore, the scope of the rights of the present invention should not be limited to the described embodiments, but should be determined by all changes or modifications derived from the claims and equivalent concepts as well as the claims.

100: Snow and ice melting device based on freezing prediction 110: Processor
111: Receiver 112: First Prediction Unit
113: 2nd Prediction Department 114: 3rd Prediction Department
115: Control unit 116: Weight adjustment unit
120: Memory 130: Network
140: Communication section 150: Power section
160: Emergency power supply 170: Connection section
180: Display part 190: Cover part
191: Hinge 200: Integrated Control Unit
1000: Snow and ice melting system based on freezing prediction

Claims (9)

A snow and ice melting device based on artificial intelligence-based freezing prediction.
a processor comprising one or more cores; and
memory;
Including,
The above processor,
A receiving unit that receives road images and road surrounding environment images captured by a monitoring camera, air temperature and humidity measured by an air temperature and humidity measuring sensor, and road surface temperature measured by a road surface temperature sensor;
A first prediction unit that inputs the road surrounding environment image received by the above receiver into a learned first prediction model and outputs the probability of rain, snow, fog, and clear weather;
A second prediction unit that inputs the road image, air temperature and humidity, and road surface temperature received by the above receiver into a learned second prediction model and outputs a first freezing probability; and
A third prediction unit that multiplies the rainfall, snowfall, fog and clear probability output by the first prediction unit and the first freezing probability output by the second prediction unit by their respective weights and then adds them to output the second freezing probability;
The above processor,
A control unit further includes: a control unit that controls the salt water pipe and the injection pump so that salt water is sprayed on the road when the second freezing probability output by the third prediction unit exceeds the first threshold value, and controls the salt water pipe and the injection pump so that salt water is not sprayed on the road when the second freezing probability output by the third prediction unit is less than or equal to the first threshold value;
The above receiver,
Receive real-time weather forecasts or additional information on whether or not there have been any accidents due to ice in the area where the above road is located;
The above control unit,
The first threshold value is adjusted based on the real-time weather forecast received by the above receiver or information on whether an accident due to ice has occurred at the location of the above road,
The above control unit,
If the second freezing probability output by the third prediction unit exceeds the second threshold value, the injection pump is controlled so that salt water is continuously sprayed on the road.
If the second freezing probability output by the third prediction unit exceeds the first threshold value and is lower than or equal to the second threshold value, the salt water pipe is controlled so that the time for which salt water is sprayed on the road is reduced while the output of the injection pump is maintained.
The above receiver,
Receive more brine balance information,
The above control unit,
If the remaining amount of brine is below a certain level, the first threshold value is increased, or the brine pipe is controlled so that the time for which brine is sprayed onto the road is reduced while the output of the spray pump is maintained.
The third prediction section above,
An artificial intelligence-based snow and ice melting device based on ice prediction, characterized in that the second freezing probability is calculated using the following mathematical expression 1, and if snowfall is forecasted by considering the weather forecast or whether it is an area prone to ice formation, the weight of w ₂ or w ₅ is upwardly adjusted or the weights of w ₁ , w ₃ , or w ₄ are downwardly adjusted, and if the road is a road where an ice formation accident has occurred based on information on whether an ice formation accident has occurred at the location of the road, the weight of the first ice formation probability is upwardly adjusted.
[Mathematical formula 1]

(Here, w ₁ to w ₅ are weights, a ₁ is the probability of rain, a ₂ is the probability of snow, a ₃ is the probability of fog, a ₄ is the probability of clear weather, and b ₁ is the probability of first freezing.) delete delete delete In the first paragraph,
The above receiver,
Receive more brine balance information,
The above control unit,
An artificial intelligence-based snow and ice melting device based on freezing prediction, characterized in that when the amount of brine remaining is below a certain level, the first threshold value is increased or the brine pipe is controlled so that the time for which brine is sprayed onto the road is reduced while the output of the spray pump is maintained. In the first paragraph,
The above receiver,
Receive real-time weather forecasts or additional information on whether or not there have been any accidents due to ice in the area where the above road is located;
The above processor,
A snow and ice melting device based on ice prediction using artificial intelligence, characterized in that it further includes a weight adjustment unit that adjusts each of the weights according to real-time weather forecast received by the receiving unit or information on whether an accident due to ice has occurred at the location of the road. In the first paragraph,
The above first prediction model is,
It is a deep learning model that uses compound scaling of MobileNetV2 and EfficientNet.
The above second prediction model is,
A snow and ice melting device based on artificial intelligence-based freezing prediction characterized by a linear regression model. In the first paragraph,
Department of Communications;
Power supply;
emergency power supply;
A connection unit connected to a plurality of sensors to receive information from the plurality of sensors;
A display unit configured as a touch panel for outputting information measured from the plurality of sensors, such as rain, snow, fog and clear probability, and first and second freezing probabilities, and for controlling the control unit; and
A cover section protecting the communication section, the power section, the emergency power section, the display section, the processor, and the memory;
A snow and ice melting device based on artificial intelligence-based freezing prediction, characterized by further including: A snow and ice melting system based on artificial intelligence-based ice formation prediction, comprising a plurality of edge-based computers installed to control salt spraying at each section of a road and an integrated control unit controlling the plurality of edge-based computers,
The above edge-based computer,
A processor comprising one or more cores and memory,
The above processor,
A system for controlling a road, comprising: a receiving unit for receiving a road image captured by a monitoring camera, a road surrounding environment image, an atmospheric temperature and humidity measured by an atmospheric temperature and humidity measuring sensor, and a road surface temperature measured by a road surface temperature sensor; a first prediction unit for inputting the road surrounding environment image received by the receiving unit into a learned first prediction model and outputting a probability of rain, snowfall, fog, and clear weather; a second prediction unit for inputting the road image, atmospheric temperature and humidity, and road surface temperature received by the receiving unit into a learned second prediction model and outputting a first freezing probability; a third prediction unit for multiplying the rain, snowfall, fog, and clear weather probabilities output by the first prediction unit and the first freezing probability output by the second prediction unit by their respective weights and then adding them up to output a second freezing probability; and a control unit for controlling a salt water pipe and a spray pump so that salt water is sprayed onto the road when the second freezing probability output by the third prediction unit exceeds a first threshold value.
The above integrated control unit,
If at least one third prediction unit among the plurality of edge-based computers outputs a second freezing probability exceeding the first threshold, control an adjacent edge-based computer to spray salt water on an adjacent road;
The above processor,
A control unit further includes: a control unit that controls the salt water pipe and the injection pump so that salt water is sprayed on the road when the second freezing probability output by the third prediction unit exceeds the first threshold value, and controls the salt water pipe and the injection pump so that salt water is not sprayed on the road when the second freezing probability output by the third prediction unit is less than or equal to the first threshold value;
The above receiver,
Receive real-time weather forecasts or additional information on whether or not there have been any accidents due to ice in the area where the above road is located;
The above control unit,
The first threshold value is adjusted based on the real-time weather forecast received by the above receiver or information on whether an accident due to ice has occurred at the location of the above road,
The above control unit,
If the second freezing probability output by the third prediction unit exceeds the second threshold value, the injection pump is controlled so that salt water is continuously sprayed on the road.
If the second freezing probability output by the third prediction unit exceeds the first threshold value and is lower than or equal to the second threshold value, the salt water pipe is controlled so that the time for which salt water is sprayed on the road is reduced while the output of the injection pump is maintained.
The above receiver,
Receive more brine balance information,
The above control unit,
If the remaining amount of brine is below a certain level, the first threshold value is increased, or the brine pipe is controlled so that the time for which brine is sprayed onto the road is reduced while the output of the injection pump is maintained.
The third prediction section above,
A snow and ice melting system based on artificial intelligence for predicting freezing, characterized in that the second freezing probability is calculated using the following mathematical expression 1, and if snowfall is forecasted by considering the weather forecast or whether the area is an area prone to freezing, the weight of w2 or w5 is increased or the weights of _w1 , _w3 , or _w4 are decreased, and if the road is a road where an ice-related accident has occurred based on information on whether an ice-related accident has occurred at the location of the road, the weight of the first ice-related probability is increased.
[Mathematical formula 1]

(Here, w ₁ to w ₅ are weights, a ₁ is the probability of rain, a ₂ is the probability of snow, a ₃ is the probability of fog, a ₄ is the probability of clear weather, and b ₁ is the probability of first freezing.)