JP7249589B2 - Antifreezing agent spraying control method and stationary antifreezing agent spraying device - Google Patents

Description

INDUSTRIAL APPLICABILITY The present invention is used by installing it on the side of a road or the like to spray an antifreezing agent, sand or other spraying agent (collectively referred to as an "antifreezing agent") to prevent the road surface from freezing. It relates to a stationary antifreezing agent spraying device, and more particularly to a method for controlling the spraying of the antifreezing agent.

A stationary antifreezing agent spraying device is proposed in Patent Documents 1 and 2. The stationary antifreezing agent spraying device described in Patent Document 1 is configured to automatically spray an antifreezing agent when the outside temperature is low during a time period when freezing and snow accumulation on the road surface are expected. The antifreezing agent spraying system described in Patent Document 2 is configured so that an antifreezing agent spraying device installed in a remote location can be activated by a remote control from an administrator and can be sprayed when necessary. It is

Since the stationary anti-freezing agent spraying system is premised on unmanned automatic operation, the situation in which there is a risk of road surface freezing is set in advance based on the time of day, outside temperature, etc., and if such a situation occurs, the anti-icing agent will be applied. are disseminated. For example, the antifreezing agent automatic supply device described in Patent Document 3, when the outside temperature is low and the snowfall detector detects snowfall, the amount calculated based on the information from the road surface temperature detector and the road surface condition detector. Antifreeze is sprayed on the road surface. Also, based on weather information, when the road surface is expected to freeze, an anti-freezing agent is sprayed on the road surface.

In the road surface freeze prevention method described in Patent Document 4, a road surface condition determination unit in a control unit of a road surface information movement and collection system determines whether the road surface is dry, wet, frozen, or compacted snow based on information collected by a road surface condition sensor. , the road surface temperature comparison unit compares the road surface temperature measured by the road surface thermometer with 0°C, and the freeze prediction determination unit compares the salt concentration from the salinity meter with the road surface temperature stored in the memory. and the salinity concentration of the road surface, it is predicted whether or not the road surface will freeze (paragraph 0015).

Japanese Patent No. 3105771 Japanese Unexamined Patent Application Publication No. 2001-279637 JP-A-9-143948 JP-A-11-229311

In spraying the antifreezing agent, it is necessary to grasp the state of the road surface and, based on this, spray the antifreezing agent on the road surface that is frozen or is expected to freeze at an appropriate timing. The road surface sensor described in Patent Document 4 has a configuration in which a laser beam having a deflection characteristic is emitted toward the road surface by a projector, and the light reflected from the road surface is received by a light receiver installed facing the projector. ing. The changes in the intensity and deflection characteristics of reflected light due to road surface conditions are used to determine whether the road surface is dry, wet, compacted, or icy. However, the current situation is that a device for judging the road surface condition with practical accuracy and spraying the antifreezing agent at an appropriate time and timing based on the judgment has not been put into practical use.

In recent years, image recognition of human faces and the like has been performed using machine learning algorithms such as neural networks. In particular, a learning model constructed by deep learning such as a convolutional neural network (CNN) is used to recognize images such as face images. For example, a model called VGG16 is known, and this VGG16 is a typical trained convolutional neural network that classifies 1,200,000 images trained on a large-scale image data set called "ImageNet" into 1000 categories. is a model. Also, if the VGG16 model is fine-tuned and trained, it can classify things other than those 1000 categories.

The object of the present invention is to apply image classification by deep learning to determine the road surface condition for the necessity of spraying anti-freezing agent, determine the road surface condition with practical accuracy, and To propose a spraying control method capable of appropriately spraying an antifreeze agent on the ground. It is also an object of the present invention to provide a stationary antifreeze sprayer incorporating a control mechanism according to this method.

In the antifreezing agent spray control method of the present invention, using a trained model constructed by deep learning using a convolutional neural network or the like, image classification of road surface captured images is performed to determine the road surface state, and the determination result and Based on the road surface temperature, it is determined whether or not the road surface is in a frozen state requiring the application of an antifreezing agent.

The present inventors used Keras, one of the open source neural network libraries, to create a convolutional neural network using road surface images in four categories: dry, wet or frozen, snow-covered, and other states. Image classification of road surface conditions was performed using a trained model for road surface image classification by , and a trained model constructed by fine-tuning of a convolutional neural network. As a result, it was confirmed that it is possible to classify the road surface condition into the above four with sufficient accuracy for practical use. By installing an image classifier equipped with a trained model for road surface image classification built by deep learning in the control computer of the stationary anti-freezing agent spraying device, the anti-freezing agent can be sprayed appropriately according to the road surface condition. A possible stationary antifreeze sprayer can be realized.

Here, at night or in a state of low illuminance, there is a possibility that a clear road surface photographed image cannot be acquired to the extent that the image can be recognized. Therefore, both a visible light image and an infrared image are acquired as the road surface photographed image. It is desirable to determine the road surface condition by In this way, the road surface condition can always be determined with a predetermined accuracy regardless of whether it is day or night or whether the ambient illumination is high. For this purpose, in the present invention, a visible light image and an infrared image are acquired as captured images of the road surface, and image classification is performed on both the visible light image and the infrared image.

Instead of using both images, it is also possible to select the road surface image used for image classification based on the average brightness of the visible light image and the infrared image. In this case, the average brightness of the visible light image is calculated, and if the average brightness is higher than a certain value, the visible light image is used to determine the road surface condition by image classification. is used to determine the road surface condition.

Here, when there are obstacles such as people and cars on the road surface, the antifreezing agent cannot be sprayed. Generally, a dedicated obstacle sensor such as a distance sensor is arranged to detect the presence or absence of an obstacle. If the presence or absence of an obstacle can be detected from the road surface photographed image, a dedicated obstacle sensor can be omitted. In the present invention, an image of a line for specifying an extraction range attached to the road surface, or an image of a center line or a roadside strip line attached to the road surface is recognized from the photographed image, and the line image is recognized from the image of the line. If all can be recognized, it is determined that there are no obstacles within the spray area. Moreover, even when the road surface is covered with snow, all or part of the line cannot be recognized from the photographed road surface image. Of course, it is also possible to double check that there are no obstacles on the road surface by arranging a dedicated obstacle sensor such as a distance sensor.

Furthermore, lines attached to the road surface may be used to cut out images used for image classification from the road surface photographed images. In this case, a closed-shaped line mark such as a circle, an oval, or a polygon is attached to the road surface as a line for specifying the extraction range, and the portion within the range of the line mark shown in the road surface photographed image should be cut out.

The stationary antifreezing agent spraying apparatus of the present invention sprays the antifreezing agent on the road surface when it is determined that there are no obstacles in the road surface and the spraying area is in a frozen state. Since the action is performed, the spraying action can be avoided in road surface conditions in which spraying is unnecessary or inappropriate.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is an explanatory view showing a configuration example of an antifreezing agent spraying system equipped with a stationary antifreezing agent spraying device to which the present invention is applied; It is an explanatory view showing a spraying mechanism and a control system of a stationary antifreezing agent spraying device. It is explanatory drawing which shows the dispersion|spreading area|region etc. of a stationary type antifreezing agent spraying apparatus. Fig. 3 is a functional block diagram showing main functions of a sprayer controller of the stationary antifreezing agent spraying device; It is a schematic flowchart which shows an example of antifreezing agent spraying operation|movement.

An embodiment of an antifreezing agent spraying system equipped with a stationary antifreezing agent spraying device to which the present invention is applied will be described below with reference to the drawings. In addition, the following embodiment shows an example of the present invention, and is not intended to limit the present invention to the configuration of the embodiment.

FIG. 1 is an explanatory diagram showing the overall configuration of the antifreezing agent spraying system. An antifreezing agent spraying system 1 includes a management server 2 and a large number of stationary antifreezing agent spraying devices 3(1), 3(2), 3(3), . . . and a communication network 10 for communication connection between the management server 2 and the stationary antifreezing agent spraying devices 3(1), 3(2), 3(3) . . . In addition, if necessary, one or more communication terminals 9 (1) such as personal computers equipped with communication functions and Internet browser functions, etc., one or more mobile communication terminals 9 such as smartphones and tablet terminals (2) is also provided. The communication network 10 can use a communication line network such as a mobile phone network and an optical communication line.

The management server 2 can be constructed as a cloud server installed on the Internet, for example. The management server 2 communicates with the stationary antifreezing agent spraying devices 3 set in various places, can check their operation history, etc., can remotely control their spraying operations, Conditions can be set or changed. Using the communication terminal 9 (1) and the communication portable terminal 9 (2), for example, communication with the management server 2 is performed, and the operation history etc. of the stationary antifreezing agent spraying device 3 installed in various places can be confirmed. Alternatively, it is possible to directly communicate with each stationary antifreezing agent spraying device 3 and check the operation history and the like. In addition, for example, one of the plurality of communication portable terminals 9 (2) has administrative authority to check the operation of each stationary antifreeze agent spraying device 3 and to set or change operating conditions such as spraying conditions. can also be given. By logging in to the management server 2 from a smartphone or tablet terminal using the registered ID and password, these terminals can be provided with such a browsing function for management and other functions. Each stationary antifreezing agent spraying device 3 (1), 3 (2), 3 (3) . and When the monitoring communication unit 8 is externally attached to the spreader 4 via a cable or the like, the monitoring communication unit 8 may be built in and integrated with the spreader 4 . In the following, when there is no need to distinguish between the stationary antifreeze agent spraying devices 3 (1), 3 (2), 3 (3), etc., these are referred to as "stationary antifreeze agent spraying device 3”.

FIG. 2(a) is an explanatory view showing a configuration example of the mechanical portion of the sprayer 4 of the stationary antifreeze agent spraying device, and FIG. 2(b) is an explanatory view showing a configuration example of the control system. As shown in FIG. 2( a ), the spreading mechanism 40 of the spreader 4 includes a body case 41 , a warning light 42 is arranged on the upper front surface of the body case 41 , and a warning light 42 is arranged on the lower front portion of the body case 41 A horizontally elongated antifreezing agent sprinkling port 43 is opened, and the sprinkling port 43 is opened and closed by an opening/closing door 44 a of an opening/closing door mechanism 44 . A hopper 45 filled with an antifreezing agent is provided in the upper end portion of the inside of the main body case 41, and a replenishment port (not shown) for replenishing the antifreezing agent is arranged on the side surface of the main body case 41. be. Inside the main body case 41 , the antifreezing agent that drops from the lower end opening of the hopper 45 is supplied through the supply port 46 to the disk rotary spraying unit 47 on the lower side. The supply port 46 is opened and closed by a shutter plate 48 a of a shutter mechanism 48 . A sprinkling port 43 formed in the main body case 41 is positioned on the front side of the disc rotary sprinkling unit 47 . A driving battery power supply 49 is arranged below the disk rotary type spreading unit 47 .

As shown in FIG. 2(b), the control system of the spreader 4 includes a detection section 50 that detects the operating state of each section and the surrounding situation, and based on the detection signal from the detection section 50, satisfies preset spreading conditions. and a spreader control unit 60 for driving the spreading mechanism 40 to perform the spreading operation of the antifreezing agent. The detection unit 50 includes an outside air temperature sensor 51a that detects the ambient temperature, a road surface temperature sensor 51b that detects the road surface temperature, and an obstacle that detects whether or not there are obstacles such as people and vehicles in the antifreezing agent spraying area. An object sensor 52 is included. The detector 50 also includes a door sensor 53 for detecting opening/closing of the opening/closing door 44a and a shutter opening sensor 54 for detecting the opening of the shutter plate 48a. A spraying agent remaining amount sensor 55 for detecting the remaining amount of the antifreezing agent remaining in the hopper 45 may be arranged. The dusting agent remaining amount sensor 55 detects the load or volume of the antifreezing agent. Further, the detection unit 50 of this example is provided with cameras 56 and 57 for photographing the road surface. A camera 56 acquires a visible light image as a road surface photographed image, and a camera 57 acquires an infrared image as a road surface photographed image.

The spreader control unit 60 can be configured, for example, mainly by a microcomputer. an input interface 64 for receiving input signals from the outside, an output interface 65 for outputting control signals and the like to the outside, an input/output interface 68 for communication, and data and address It basically consists of a bus 66 for transmitting and receiving.

The spreader control unit 60 is connected via an input interface 64 to an input unit 67 in which various operation buttons including a replenishment button 67a, operation switches, etc. are arranged. Detection signals from 51a, road surface temperature sensor 51b, obstacle sensor 52, door sensor 53, shutter opening sensor 54, spraying agent remaining amount sensor 55, etc., and road surface photographed images from cameras 56 and 57 are input. The spreader controller 60 is connected to each part of the spreader mechanism 40 via an output interface 65 . For example, the drive motor 44b of the opening/closing door mechanism 44, the drive motor 48b of the shutter mechanism 48, the drive motor 47b of the disk rotary spreading unit 47, and the like are driven and controlled via a motor driver. Also, the cameras 56 and 57 are activated to photograph the road surface at a predetermined cycle.

The spreader control unit 60 activates the spreading mechanism 40 to perform the spreading operation of the antifreezing agent when preset spreading conditions are satisfied. As described later, when the outside temperature is below the set temperature and the road surface is likely to freeze or is in a frozen state, the antifreeze agent is applied at the set time on the condition that no obstacles such as people or cars are detected. disperse throughout. Various conditions can be set as the spraying conditions. For example, spraying can be performed at a set time, such as at night, on the condition that there are no obstacles such as people or vehicles.

FIG. 3 is an explanatory diagram showing the road surface on which the antifreezing agent is sprayed by the sprayer 4. As shown in FIG. The spreader 4 spreads an antifreezing agent, for example, on the asphalt-paved road surface 81 of the road 80 . The cameras 56 and 57 are attached, for example, to posts or the like set up on the side of the spreader 4, and photograph all or part of the antifreezing agent spread area 82 on the road surface 81 by these cameras. A plurality of marks for extracting the specified range are attached on the road surface within the photographing range 83 of the scattering area 82 . For example, between the center line 84 of the white line and the lines 85 and 86 of the side strips on both sides, a plurality of line marks 87 are attached at regular intervals. The line mark 87 is, for example, a square white outline line mark or a square outline line mark drawn on the road surface using fluorescent paint.

FIG. 4 is a functional block diagram showing main functions of the spreader controller 60 of the spreader 4 related to the spreader operation control. The spreader control unit 60 functions as the following units by executing a spread control program. The spreader control unit 60 includes an image processing unit 70 that processes the road surface photographed images (56a, 57a) acquired from the cameras 56, 57, a road surface freeze determination unit 71 that determines whether the road surface is frozen, and a spread start determination/instruction unit. It functions as part 72 . The image processing unit 70 has the functions of a road surface confirmation unit 701 , an image clipping unit 702 , an image selection unit 703 and a road surface condition determination unit 704 .

A road surface confirmation unit 701 determines whether or not an obstacle such as a person or a vehicle exists on a road surface 81 based on a visible light image 56a and an infrared image 57a, which are captured images of the road surface acquired from the cameras 56 and 57 at regular sampling intervals. to confirm. For example, an image of a line mark 87 attached to the road surface for specifying a cutout range, or an image of a center line 84 attached to the road surface 81 or lines 85 and 86 of the roadside strip is extracted from the captured road surface image. If all of the line marks 87 can be recognized from the image of the line marks 87, it can be determined that there are no obstacles on the road within the scattering area 82 and that the road surface 81 is not covered with snow. When an obstacle such as a car exists, part of the line mark 87 is hidden and cannot be captured in the image. Or, it can be determined that there is snow on the road surface.

The image clipping unit 702 uses the line marks 87 attached to the road surface 81 to perform clipping processing of the captured road surface image. In this example, a line mark 87 attached to the road surface 81 is used as a line for specifying the extraction range. An image of the area surrounded by the line marks 87 is extracted as a cropped image. It is also possible to cut out a preset designated area from the road surface photographed image without using the line mark 87 . Also, when the line mark 87 is used as a line mark for specifying the extraction range, the shape of the line mark 87 is a closed shape such as a circle, an ellipse, or a polygon.

The image selection unit 703 calculates the average brightness of the clipped images 56b and 57b of the visible light image 56a and the infrared image 57a. If the calculated average brightness is equal to or greater than a preset value, the visible light image 56a is selected as the road surface condition determination image. If the average brightness is lower than a preset value, it is determined that the visible light image 56a is unclear, and the cutout image 57b of the infrared image 57a is selected as the cutout image used for determining the road surface condition.

The road surface condition determination unit 704 classifies the selected clipped image (56b or 57b) using an image classifier 705 equipped with a trained model constructed using deep learning, in this example, a convolutional neural network. , the spray surface condition of the spray area 82 is classified into four conditions: dry, wet or frozen, snow-covered, and other conditions. Here, the image classifier 705 is equipped with a visible light image model and an infrared image model as trained models for image classification, and can classify visible light images and infrared images.

In this example, image recognition is performed on one of the selected clipped images, but it is also possible to perform image recognition on both clipped images. In this case, if either the visible light image or the infrared image determines that the spray surface is wet or frozen, it is determined that the road surface is wet or frozen.

When the image processing unit 70 determines that the road surface is wet or frozen, the road surface temperature sensor 51b detects the road surface temperature sensor 51b in the road surface temperature sensor 51b. When the temperature is lower than the freezing determination temperature threshold, it is determined that the road surface is in a frozen state requiring the application of an antifreezing agent.

When it is determined that the road surface is frozen, the spraying start determination/instruction unit 72 determines whether or not to start spraying the antifreezing agent. issues a spraying start command to the spraying mechanism 40 to carry out the spraying operation. The spraying start determination/instruction unit 72, for example, determines whether or not a specified time has passed since the previous spraying of the antifreezing agent to the spraying area 82. If the specified time has passed, the antifreezing is immediately started. When the specified time has not passed, the start of the spraying operation is instructed after the specified time has elapsed. Information such as antifreezing agent spraying history (spraying date and time, etc.), temperature threshold for judging road surface freezing, specified time, etc. is stored in the storage unit 73 formed in the ROM 62 or RAM 63, and is read from there.

Alternatively, the freeze effect monitoring section may be arranged in the spreader control section 60 . The antifreezing effect monitoring unit monitors changes in the road surface temperature in the spraying area 82 since the last spraying operation of the antifreezing agent was completed. If the value indicating the spraying effect becomes equal to or less than the set value, the start of the spraying operation of the antifreezing agent is instructed. It is also possible to monitor changes in the relationship between the road surface temperature and the outside air temperature in the spraying area 82 and to determine whether the spraying effect is continuing or not based on these changes.

In addition, the spraying start determination/instruction unit 72 uses the obstacle sensor 52 to confirm again that there are no obstacles such as people or vehicles in the spraying area 82, and if there are no obstacles, the spraying start instruction is issued. You can put it out. In this case, the detection range of the obstacle sensor 52 may be set to a range different from the photographing range 83, or to a range that only partially overlaps. By double checking in this way, it is possible to reliably avoid the harmful effects of spraying the antifreezing agent on people, cars, and the like.

FIG. 5 is a schematic flow chart collectively showing an example of the antifreezing agent spraying control operation of the stationary antifreezing agent spraying device 3 of the present embodiment. An example of the spraying operation will be described with reference to this figure. First, the cameras 56 and 57 of the sprayer 4 are driven at a constant sampling cycle, and a photographed road surface image is acquired by the sprayer controller 60 (step ST1). : Street photography). Further, the road surface temperature and the outside air temperature of the dispersion area 82 at the time of photographing the road surface 81 detected by the road surface temperature sensor are acquired (step ST2: temperature detection).

An image of a line mark attached to the road surface for designating an extraction range, a center line 84 attached to the road surface 81, and images of lines 85 and 86 of the roadside strip are extracted from the photographed road surface image. When the line mark 87, the center line 84, and the roadside strip lines 85 and 86 can all be recognized, it is determined that there are no obstacles within the scattering area 82 and that the road surface 81 is not covered with snow (step ST3). : check the road surface). When an obstacle or the like is confirmed, the spraying operation is not started.

If there is no obstacle or the like, the image portion of the specified range of the scattering area 82 is cut out from the acquired road surface photographed image (step ST4: image cutout). A road surface photographed image is cut out along the line marks 87 attached to the scattering area 82 to obtain a cut-out image 56b of the visible light image 56a and a cut-out image 57b of the infrared image 57a.

Next, the average brightness is calculated for the clipped image 56b of the visible light image 56a. If the average brightness is equal to or higher than a certain value, the clipped image of the visible light image is selected. An infrared image is selected as a clipped image (step ST5: image selection).

Image classification is performed on the selected clipped image using an image classifier 705 equipped with a trained model generated by deep learning. It is determined which one of the four states of the state and the other state (step ST6: road surface state determination).

If the road surface temperature in the spray area 82 is lower than the set temperature threshold when the spray surface condition is determined to be wet or frozen, the spray surface condition is determined to require antifreeze spraying. It is determined that the road surface is frozen (step ST7: Determination of frozen road surface).

Determining whether or not to start the spraying operation of spraying the antifreezing agent on the spraying area 82 of the road surface 81 when it is determined that there are no obstacles in the spraying area 82 of the road surface and the spraying area 82 is in a frozen state. Judgment (step ST8: spraying start judgment). In determining the start of spraying, it is determined whether or not a specified time has elapsed since the previous spraying of the antifreezing agent to the spraying area 82, and if the specified time has passed, the antifreezing agent is sprayed immediately. The start of the operation is instructed (step ST9). If the specified time has not elapsed, the start of the spraying operation is instructed after the specified time has elapsed. When the start of the spraying operation is instructed, the spraying mechanism 40 is driven to perform the spraying operation of the antifreezing agent to the spraying area of the road surface.

(Other embodiments)
(1) A monitoring communication unit 8 is attached to the stationary antifreezing agent spraying device 3 of this embodiment. Data is transmitted and received to and from the management server 2 via this monitoring communication unit 8 . In addition to this, the monitoring communication unit 8, for example, periodically acquires weather radar data at a set position, and controls the weather conditions in the spraying area 82 in the sprayer controller 60 based on the acquired weather data. , for example, it is also possible to determine rain or snow conditions. It is also possible to set a freezing condition for determining a frozen state from a combination of acquired weather information and road surface temperature, and drive the spraying mechanism 40 to spray when the freezing condition is satisfied.

(2) In addition, it is also possible to perform data linkage with the nearby stationary anti-freezing agent spraying device 3 via the monitoring communication unit 8 regarding road surface conditions, weather information, road surface temperature, spraying history, etc. is.

1 Antifreezing agent spraying system 2 Management servers 3, 3 (1), 3 (2), 3 (3) Stationary antifreezing agent spraying device 4 Spreader 8 Monitoring communication unit 9 (1) Communication terminal 9 (2) Portable communication terminal 10 Communication network 40 Sprinkling mechanism 41 Main body case 42 Warning light 43 Sprinkling port 44 Open/close door mechanism 44a Open/close door 44b Drive motor 45 Hopper 46 Supply port 47 Disk rotating sprinkling unit 47b Drive motor 48 Shutter mechanism 48a Shutter plate 48b Drive motor 49 Drive battery power supply 50 Detector 51a Outside air temperature sensor 51b Road surface temperature sensor 52 Obstacle sensor 53 Door sensor 54 Shutter opening sensor 55 Spray agent remaining amount sensor 56, 57 Camera 56a Visible light image 56b, 57b Cut image 57a Infrared Image 60 Spreader control unit 61 CPU
62 ROMs
63 RAM
64 Input interface 65 Output interface 66 Bus 67 Input unit 67a Replenishment button 68 Communication input/output interface 70 Image processing unit 71 Road surface freeze determination unit 72 Spray start determination/instruction unit 73 Storage unit 80 Road 81 Road surface 82 Spray area 83 Shooting range 84 Center lines 85, 86 Line 87 Line mark 701 Road surface confirmation unit 702 Image clipping unit 703 Image selection unit 704 Road surface condition determination unit 705 Image classifier

Claims (14)

A road surface photographing step of photographing an image of the road surface including the sprayed area where the antifreezing agent is sprayed;
a road surface temperature measuring step of measuring the road surface temperature of the dispersion area when photographing the road surface;
an image clipping step of clipping an image portion of a specified range within the scatter area from the acquired photographed image;
Image classification is performed on the clipped image clipped from the captured image using an image classifier equipped with a trained model constructed by deep learning, and the road surface state of the spray area is dry, wet or frozen. a road surface state determination step of determining which of the four states of the snow cover state and other states;
When the road surface condition is determined to be the wet or frozen condition and the road surface temperature in the spraying area is lower than a set threshold value, the antifreeze agent is not sprayed. A frozen road surface determination step for determining that the frozen state is required;
and
In the road surface photographing step, a visible light image and an infrared image are obtained as the photographed image,
In the image clipping step, images are clipped from each of the visible light image and the infrared image,
In the road surface condition determination step, the image classifier in which the trained model for visible light images and the trained model for infrared images are incorporated as the trained models are used to classify each image of the visible light image and the infrared image. do the classification,
In the frozen road surface determination step, when the road surface state is determined to be the wet or frozen state in either the visible light image or the infrared image, whether the road surface is in the frozen state is determined based on the road surface temperature. make a decision whether or not
A method for controlling the spraying of an antifreeze agent. In the road surface condition determination step, the average luminance of the visible light image is calculated, and if the average luminance is equal to or higher than a predetermined value, the visible light image is classified. Perform image classification of infrared images
The method for controlling the spraying of the antifreezing agent according to claim 1. performing a road surface confirmation step between the road surface photographing step and the image cutout step;
In this road surface confirmation process,
recognizing from the photographed image an image of a line mark attached to the road surface for specifying an extraction range, or a center line or roadside strip line attached to the road surface;
If all of the line marks, the center line, or the roadside strip can be recognized from the photographed image, it is determined that no obstacle exists within the spray area.
The method for controlling the spraying of the antifreezing agent according to claim 1 or 2. A road surface photographing step of photographing an image of the road surface including the sprayed area where the antifreezing agent is sprayed;
a road surface temperature measuring step of measuring the road surface temperature of the dispersion area when photographing the road surface;
an image clipping step of clipping an image portion of a specified range within the scatter area from the acquired photographed image;
Image classification is performed on the clipped image clipped from the captured image using an image classifier equipped with a trained model constructed by deep learning, and the road surface state of the spray area is dry, wet or frozen. a road surface state determination step of determining which of the four states of the snow cover state and other states;
When the road surface condition is determined to be the wet or frozen condition and the road surface temperature in the spraying area is lower than a set threshold value, the antifreeze agent is not sprayed. A frozen road surface determination step for determining that the frozen state is required;
and
Furthermore,
performing a road surface confirmation step between the road surface photographing step and the image cutout step;
In this road surface confirmation process,
recognizing from the photographed image an image of a line mark attached to the road surface for specifying an extraction range, or a center line or roadside strip line attached to the road surface;
If all of the line marks, the center line, or the roadside strip can be recognized from the photographed image, it is determined that no obstacle exists within the spray area.
A method for controlling the spraying of an antifreezing agent , characterized by : 5. The antifreezing agent spraying control method according to claim 3, wherein in said road surface confirmation step, along with image recognition of said line mark, a distance sensor is used to confirm that said obstacle does not exist within said spraying area. The shape of the line mark for specifying the extraction range is circular, elliptical or polygonal,
5. The antifreezing agent dispersion control method according to claim 3 , wherein in the image clipping step, an image of a region surrounded by the line marks is extracted as a clipped image. A spraying step of spraying the antifreezing agent on the spraying area of the road surface when it is determined that there is no obstacle in the spraying area of the road surface and the spraying area is in the frozen state. 7. The method for controlling the spraying of the antifreeze agent according to any one of items 1 to 6. A road surface photographing step of photographing an image of the road surface including the sprayed area where the antifreezing agent is sprayed;
a road surface temperature measuring step of measuring the road surface temperature of the dispersion area when photographing the road surface;
an image clipping step of clipping an image portion of a specified range within the scatter area from the acquired photographed image;
Image classification is performed on the clipped image clipped from the captured image using an image classifier equipped with a trained model constructed by deep learning, and the road surface state of the spray area is dry, wet or frozen. a road surface state determination step of determining which of the four states of the snow cover state and other states;
When the road surface condition is determined to be the wet or frozen condition and the road surface temperature in the spraying area is lower than a set threshold value, the antifreeze agent is not sprayed. A frozen road surface determination step for determining that the frozen state is required;
a spraying step of spraying the antifreezing agent on the spraying area of the road surface when it is determined that there is no obstacle in the spraying area of the road surface and the spraying area is in the frozen state;
and
Furthermore,
a spraying start determination step that is performed when the frozen road surface is determined to be in the frozen state in the road surface freeze determination step;
In this spraying start determination step,
Determining whether a specified time has passed since the previous antifreezing agent was sprayed to the spraying area,
instructing to start spraying the antifreezing agent immediately when the specified time has passed;
In the spraying step, the antifreeze agent is sprayed in response to an instruction to start the spraying operation.
A method for controlling the spraying of an antifreezing agent , characterized by : A road surface photographing step of photographing an image of the road surface including the sprayed area where the antifreezing agent is sprayed;
a road surface temperature measuring step of measuring the road surface temperature of the dispersion area when photographing the road surface;
an image clipping step of clipping an image portion of a specified range within the scatter area from the acquired photographed image;
Image classification is performed on the clipped image clipped from the captured image using an image classifier equipped with a trained model constructed by deep learning, and the road surface state of the spray area is dry, wet or frozen. a road surface state determination step of determining which of the four states of the snow cover state and other states;
When the road surface condition is determined to be the wet or frozen condition and the road surface temperature in the spraying area is lower than a set threshold value, the antifreeze agent is not sprayed. A frozen road surface determination step for determining that the frozen state is required;
a spraying step of spraying the antifreezing agent on the spraying area of the road surface when it is determined that there is no obstacle in the spraying area of the road surface and the spraying area is in the frozen state;
and
Furthermore,
a spraying start determination step that is performed when the frozen road surface is determined to be in the frozen state in the road surface freeze determination step;
From the end of the antifreezing agent spraying operation in the spraying step, the change in the road surface temperature in the spraying area or the change in the relationship between the road surface temperature and the outside air temperature in the spraying area is monitored, and based on the change a spraying effect monitoring step of determining whether the spraying effect of the antifreezing agent sprayed in the spraying area is sustained;
and
In the spraying start determination step, when the value indicating the spraying effect is equal to or less than a set value, instructing the start of spraying operation of the antifreeze agent;
In the spraying step, the antifreeze agent is sprayed in response to an instruction to start the spraying operation.
A method for controlling the spraying of an antifreezing agent , characterized by : a camera that takes an image of the road surface including the application area where the antifreeze is applied;
a road surface temperature sensor that measures the road surface temperature of the dispersion area when the road surface is photographed;
a spreader controller comprising a computer;
and
The spreader control unit is
an image processing unit that determines the road surface condition of the scattering area from the image captured by the camera;
a frozen road surface determination unit that determines whether the road surface is in a frozen state based on the determination result of the image processing unit and the road surface temperature;
and
The image processing unit
an image clipping unit for clipping an image portion of a specified range including the scatter area;
Image classification is performed on the clipped image clipped from the captured image using an image classifier equipped with a trained model constructed by deep learning, and the spray surface state of the spray area is a dry state, a wet state, or a frozen state. , a road surface state determination unit that determines which of four states, snow cover state and other states,
and
If the road surface temperature is lower than a preset threshold value when the road surface condition is determined to be the wet or frozen condition, the road surface freeze determination unit determines the road surface condition as the antifreeze condition. Judging that the frozen state requires spraying of the agent ,
The camera acquires a visible light image and an infrared image as the captured image,
The image clipping unit clips an image from each of the visible light image and the infrared image,
The road surface condition determination unit uses the image classifier in which a trained model for visible light images and a trained model for infrared images are incorporated as the trained models, and classifies each image of the visible light image and the infrared image. do the classification,
The frozen road surface discrimination unit determines that the sprayed surface state is the wet or frozen state in either the visible light image or the infrared image, and the frozen state is determined based on the road surface temperature. determine whether or not
A stationary antifreezing agent spraying device characterized by: The camera acquires a visible light image and an infrared image as the captured image,
The image clipping unit clips an image from each of the visible light image and the infrared image,
The image processing unit calculates the average brightness of the visible light image, selects the visible light image when the average brightness is equal to or higher than a certain value, and selects the infrared image when the average brightness is lower than the certain value. and an image selection unit that
The road surface condition determination unit performs image classification on an image selected from the visible light image and the infrared image.
11. The stationary antifreeze agent spraying device according to claim 10. A line mark for specifying a cutout range is provided on the road surface of the scattering area,
The spreader control unit includes a road surface confirmation unit that extracts the line marks from the photographed image and determines whether or not an obstacle exists on the road surface within the dispersion area based on the extraction result,
The stationary antifreezing agent spraying device according to claim 10 or 11 , wherein the image clipping section extracts an image of the area surrounded by the line marks as a clipped image . The sprayer control unit includes a spraying start determination unit that determines whether to start spraying the antifreezing agent,
The spraying start determination unit
When it is determined that there are no obstacles on the road surface of the dispersion area and the road surface state is the frozen state,
Determining whether a specified time has passed since the previous antifreezing agent was sprayed to the spraying area,
13. The stationary antifreezing agent spraying device according to claim 12 , wherein when the specified time has passed, the instruction to start the spraying operation of the antifreezing agent is given immediately . Having an outside temperature sensor for measuring the outside temperature around the road surface,
The spreader control unit is
Changes in the road surface temperature in the spraying area or changes in the relationship between the road surface temperature and the outside air temperature are monitored, and whether the spraying effect of the antifreezing agent sprayed in the spraying area is sustained based on the change a spraying effect monitoring unit for determining whether or not
a spraying start determination unit that determines whether to start spraying the antifreezing agent;
and
When it is determined that there are no obstacles on the road surface in the spraying area and the road surface condition is the frozen condition, the spraying start determining unit determines that when the value indicating the spraying effect becomes equal to or less than a set value, 14. The stationary antifreeze agent spraying device according to claim 13, which instructs the start of the antifreeze agent spraying operation .

Images