KR102468117B1 - A surveillance camera with a sun shield installed and control method thereof - Google Patents

Abstract

The present invention relates to a surveillance camera device with a sunshield mounted and operating, and a control method thereof. More specifically, through a sunshield which can be moved in the forward and backward directions, the field of view of the surveillance camera is secured in sunlight and rain, in the surveillance camera device with a sunshield mounted and operating, and the control method thereof.

Description

Surveillance camera device with a sun shield installed and its control method {A surveillance camera with a sun shield installed and control method thereof}

The present invention relates to a surveillance camera device equipped with a sunshield and a method for controlling the same, and more particularly, to a sunshield capable of moving forwards and backwards to secure a view of the surveillance camera in sunlight and rain. It relates to a surveillance camera device mounted and operated and a control method thereof.

A sunshield of a surveillance camera is a part that is attached to protect the image capturing part of the surveillance camera from rain, light, etc. so that the image can be captured clearly.

In general, surveillance cameras installed indoors use a dome-type type that does not have a sunshield attached because there is little change in the surrounding environment. Use a camera with a sunshield attached to prevent deterioration of the sharpness of the image due to snow or rain on the street.

Since the camera to which the sunshield is attached is made in a manual operation method in which a person must directly adjust the sunshield, it is difficult to improve video control efficiency because the sunshield cannot be continuously adjusted according to changes in the surrounding environment. In particular, AI-based video analysis technology, which has recently increased in importance following the 4th industrial revolution, has a clear difference in accuracy depending on the quality of the video being filmed.

In addition, CCTV has a problem in that when strong sunlight is incident, the contrast difference becomes extreme according to the amount of incident sunlight, and the quality of the image is deteriorated.

Therefore, surveillance cameras require a technology to prevent rain and strong sunlight from entering the lens in order to improve the quality of captured images.

As a prior art, Republic of Korea Patent Registration No. 10-1576772 "CCTV camera control method" is described.

Therefore, the present invention has been proposed to improve such conventional problems, and a surveillance camera that operates with a sunshield that secures the field of view of the surveillance camera device in sunlight and rain through a sunshield that can move in the forward and backward directions. It is an object to provide an apparatus and a control method thereof.

In order to solve the above problems, in a surveillance camera device and a control method for which a sunshield capable of moving back and forth is mounted on one side of a camera unit and operating thereof, control of the surveillance camera device in which the sunshield is mounted and operated according to an embodiment of the present invention The method analyzes an image captured through the camera unit to distinguish a dynamic area and a static area, and does not limit the moving distance of the sunshield even if a blind spot occurs in the static area, and the blind spot is determined by the camera unit. It is characterized in that the part where the sunshield is exposed to the lens and the sunshield is exposed to the image.

In addition, the control method of the surveillance camera device operating with the sunshield installed includes an information generation step of generating rainfall information by the camera unit capturing and analyzing the image; A moving step of moving the sunshield forward and backward according to the pre-stored sunrise or sunset time information and the rainfall information, and calculating the limited distance at which the blind spot does not occur according to the relationship between the zoom value during the movement of the sunshield, A limiting step of limiting the moving distance of the shield may be included.

In addition, the information generating step may include a basic information collection step of collecting basic rainfall information to be used for deep learning learning; an image collection step of acquiring and collecting the image captured by the camera unit; It may include a learning step of conducting deep learning based on the basic rainfall information and the image, and a rainfall information generating step of generating the rainfall information by analyzing the image based on deep learning.

In addition, the learning step may include a rainfall feature extraction step of extracting the basic rainfall information and the rainfall features of the image through CNN (Convolutional Neural Network) and the rainfall feature extracted from the CNN (Deep Convolutional Neural Network) DNN (Deep Convolutional Neural Network). A rainfall feature learning step of learning through a neural network) may be included.

In addition, the rain information generation step may include a pre-processing step of pre-processing the image and a rainfall determination step of generating the rainfall information by learning the pre-processed image and determining it as 'rainfall' or 'non-rainfall'.

In addition, the moving step may include a rainfall occurrence information generating step of generating rainfall occurrence information according to the rainfall information generated from the information generating step; A sunlight information calculation step of calculating sunlight angle information incident on the lens of the camera part through pre-stored sunrise or sunset time information and GPS information of the camera part, and one or more of the rainfall occurrence information and the sunlight angle information A movement control step of moving the sunshield in the forward and backward directions may be included.

In addition, the control method of the monitoring camera may further include an area determination step of determining the blind spot as a static area or a dynamic area when the blind spot occurs after the limiting step is finished.

In addition, in the area determination step, if the blind spot is determined to be the static area, a static determination step of eliminating the blind spot through zoom and an error determination step of generating information again if the blind spot is determined to be the dynamic area. It may include an ongoing dynamic judgment step.

A surveillance camera device in which a sunshield is installed and operated according to an embodiment of the present invention includes a camera unit and a sunshield coupled to an upper portion of the camera unit, movable in a forward and backward direction, and formed in an arch shape, and the sunshield includes: It is formed on the front surface and may include a bent portion for preventing rain water from flowing down when sunlight interferes with the lens of the camera unit as the camera unit moves in the forward and backward directions.

In addition, the sunshield may further include an anti-reflective sheet on a lower surface of the sunshield to prevent sunlight reflected from the sunshield from entering the lens of the camera unit.

A surveillance camera device with a sunshield installed and operating and a method for controlling the same according to an embodiment of the present invention can prevent rainwater from flowing to a lens part by moving the sunshield forward and backward.

In addition, it is possible to prevent sunlight from being incident on the lens portion by moving the sunshield according to the angle of sunlight.

In addition, it is possible to prevent a blind spot from occurring by limiting the moving distance of the sunshield according to the zoom value.

In addition, the effects according to the embodiments of the present invention mentioned above are not limited to the described contents, and may further include all effects predictable from the specification and drawings.

1 is a perspective view of a surveillance camera device in which a sunshield is mounted and operated according to an embodiment of the present invention.
Figure 2 is a side view of Figure 1;
Figure 3 is a front view of Figure 1;
Figure 4 is a bottom view of Figure 1
5 is a flowchart sequentially illustrating a control method of a surveillance camera device in which a sunshield is mounted and operated according to an embodiment of the present invention.
6 is a flowchart schematically illustrating steps that may be performed in step S1 of FIG. 5;
7 is a flowchart schematically illustrating steps that may be performed in step S12 of FIG. 6;
8 is a flowchart schematically illustrating steps that may be performed in step S13 of FIG. 6;
9 is a flowchart schematically illustrating steps that may be performed in step S2 of FIG. 5;

Hereinafter, the description of the present invention with reference to the drawings is not limited to specific embodiments, and various transformations may be applied and various embodiments may be applied. In addition, the content described below should be understood to include all conversions, equivalents, or substitutes included in the spirit and scope of the present invention.

In the following description, terms such as first and second are terms used to describe various components, and are not limited in meaning per se, and are used only for the purpose of distinguishing one component from another.

Like reference numbers used throughout this specification indicate like elements.

Singular expressions used in the present invention include plural expressions unless the context clearly dictates otherwise. In addition, terms such as "include", "include" or "have" described below are intended to designate that features, numbers, steps, operations, components, parts, or combinations thereof described in the specification exist. should be construed, and understood not to preclude the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

Hereinafter, an embodiment of the present invention will be described in detail with reference to FIGS. 1 to 9 attached.

In addition, in the following description, unless otherwise specified, a direction toward the front with respect to FIG. 1 is referred to as 'front', and an opposite direction is referred to as 'rear'. Accordingly, terms indicating directions such as 'left', 'right', 'upper' and 'downward' are defined as indicating respective directions based on the forward and backward directions.

1 is a perspective view of a monitoring camera device in which a sunshield is mounted and operated according to an embodiment of the present invention, FIG. 2 is a side view of FIG. 1 , FIG. 3 is a front view of FIG. 1 , and FIG. 4 is a bottom view of FIG. 1 to be.

First, a surveillance camera device in which a sunshield is mounted and operated according to an embodiment of the present invention will be described. Referring to FIGS. 1 to 4 , a surveillance camera device in which a sunshield is mounted and operated according to an embodiment of the present invention may include a camera unit 10 and a sunshield 20 .

The camera unit 10 may include a photographing unit 11 , a fixing unit 12 and an angle adjustment unit 13 .

The photographing unit 11 captures an image of a front subject, and may include a lens, a LIDAR sensor or an infrared sensor, and a control module, and may include a housing accommodating the configuration of the photographing unit 11 . The control module may control the operation and movement of the sunshield 20 as well as the photographing unit 11 . However, it is not limited thereto and a separate sunshield control unit may be provided.

The fixing part 12 may be configured to fix the photographing part 11 to the installation surface. At this time, the fixing part 12 may be provided with a mount cover made of an elastic material having flexibility that contracts or expands according to the deformation of the fixing part 12 on one side thereof, so that deformation autonomy may be imparted.

Specifically, the mount cover can attenuate vibration while the fixing part 12 contracts or expands in accordance with deformation due to vibration or shaking at one side of the fixing part 12, and contracts or expands according to deformation due to thermal expansion and contraction. While being expanded, stress generated in the fixing part 12 can be minimized.

The angle adjusting unit 13 may be provided between the capturing unit 11 and the fixing unit 12 to adjust a rolling angle and a pitching angle of the capturing unit 11 .

On the other hand, since the described configuration of the camera unit 10 is only a preferred example and is not necessarily limited, all configurations of cameras commonly used in the technical field of the present invention can be used.

The sunshield 20 is a solar blocking film configured to increase the clarity of an image, and may be coupled to the camera unit 10 to be movable in the forward and backward directions.

In addition, the sunshield 20 is coupled to the upper part of the camera unit 10, and when moving forward, the front side of the camera unit 10 is covered by the shadow of the sunshield 20 so that the front side of the camera unit 10 is longer than the front side of the camera unit 10. It is preferably formed to be movable to a protruding position, but is not limited thereto. Accordingly, when the sunshield 20 forms a shadow to cover the camera unit 10, interference from sunlight is blocked so that a clear image can be obtained.

In addition, the sunshield 20 is coupled to the upper portion of the camera unit 10 and is formed in an arcuate shape to cover portions of the upper portion and both side surfaces of the camera unit 10, so that interference of sunlight can be blocked as wide as possible.

In addition, the sunshield 20 may include a bent portion 21 . The curved portion 21 may be formed on the front surface of the sunshield 20 to prevent rainwater from flowing down on the lens of the camera unit 10 .

The bent portion 21 may include a vertical portion formed to protrude vertically from the front surface of the sunshield and a horizontal portion formed to extend horizontally from a lower end of the vertical portion. The bent portion 21 is composed of a vertical portion and a horizontal portion and may be formed in a 'b' shape.

Here, the vertical portion has a height protruding from the upper surface of the sunshield 20 to form a step with the upper surface of the sunshield 20, and the horizontal portion can form a length at the lower end of the vertical portion in the direction of facing away from the camera unit 10. have. As a result, rainwater is blocked on the vertical portion and is induced to flow in both directions of the sunshield 20 , and even if it overflows the vertical portion, it may fall at a position farther from the camera unit 10 due to the horizontal portion.

In addition, since the configuration of the described bent portion 21 is only a preferred example and is not necessarily limited, it may be formed in various shapes.

In addition, as shown in FIG. 4 , the sunshield 20 may further include a lower prevention layer 30 .

Specifically, the lower prevention film 30 may be configured to protrude downward from the lower surface of the sunshield 20 . At this time, the lower prevention film 30 may be provided alone, but may be provided in plurality, and in the case where a plurality of the lower prevention film 30 is provided, the length of the lower prevention film 30 closer to the camera unit 10 may be gradually increased. In the present invention, the inflow of rainwater into the camera unit 10 is secondarily prevented together with the curved portion 21 through the lower prevention film 30 as described above, thereby doubling the effect of preventing the inflow of rainwater.

In addition, the sunshield 20 may further include an anti-reflective sheet 40 .

The anti-reflective sheet 40 is attached to the lower surface of the sunshield 20 to prevent external light from being reflected from the sunshield 20 and entering the lens of the camera unit 10 . Here, the external light is light incident on the lower end of the sunshield 20 after sunlight is reflected from the ground or a surrounding object or light incident on the lower end of the sunshield 20 from a surrounding object such as a street lamp. ) is made of a material that absorbs light to prevent light from being incident on the lens of the camera unit 10 and to increase the sharpness of shooting of the camera unit 10.

At this time, the anti-reflective sheet 40 is attached to the lower surface of the sunshield 20, and is configured to be exposed only when the sunshield 20 moves forward, and when the sunshield 20 moves forward, the upper and lower surfaces All of them may be configured to prevent direct sunlight or reflected light, such as sunlight or street lights, from entering the lens of the camera unit 10 .

That is, according to the present invention, by preventing the inflow of rainwater into the camera unit 10 through the above-described structure, failure of the camera unit 10 may be prevented, and furthermore, the sharpness of an image may be increased by preventing diffuse reflection.

6 is a flowchart schematically showing steps that may be performed in step S1 of FIG. 5, FIG. 7 is a flowchart schematically showing steps that may be performed in step S12 of FIG. 6, and FIG. 8 is a flowchart showing steps that may be performed in step S13 of FIG. It is a flowchart schematically showing possible steps, and FIG. 9 is a flowchart schematically showing steps that can be performed in step S2 of FIG. 5 .

First, in the control method of a surveillance camera device in which a sunshield is mounted and operated according to an embodiment of the present invention, an image captured through the camera unit 10 is analyzed to distinguish a dynamic area from a static area, and a blind spot in the static area. may occur, the movement distance of the sunshield 20 may not be limited.

Here, the static area may be referred to as an area in which there are static subjects that do not move in the image, and the dynamic area may be referred to as an area in which motion or shape-changing subjects are located in the image.

In addition, the blind spot may be a portion where the sunshield 20 is exposed to the lens of the camera unit 10 and the sunshield is exposed to the image. In the monitoring camera of the present invention, since the sunshield 20 is installed above the camera unit 10, a blind spot can be formed at the upper corner of the image.

Referring to FIGS. 5 to 9 , the method for controlling a surveillance camera device in which a sunshield is mounted and operated according to an embodiment of the present invention includes an information generation step (S1), a movement step (S2), and a limiting step (S3). can do.

First, in the information generation step (S1), the camera unit 10 may capture and analyze an image to generate rainfall information. To this end, the information generation step (S1) may include a basic information collection step (S10), an image collection step (S11), a learning step (S12) and a rainfall information generation step (S13).

The basic information collection step (S10) may be a step of collecting basic rainfall information to be used for deep learning learning. In the basic information collection step (S10), rainfall photos and weather information collected on the Internet in real time or collected from national organizations such as meteorological companies and the Korea Meteorological Administration may be received and collected as basic rainfall information. In the basic information collection step (S10), the collected basic rainfall information may be stored in a database provided in the camera unit 10 or in a separate server. The weather information may be weather information according to the time of the area where the monitoring camera is installed.

Here, deep learning is an artificial neural network model that mimics the nervous system of a living thing. If the existing artificial neural network model consists of connections between thin layers of neuron models, the deep learning technique builds up deep layers of neuron models. This is a technique that applies a model that increases the learning ability of a neural network.

In the image collection step ( S11 ), images captured by the camera unit 10 may be acquired and collected. In the image collection step (S11), the collected image may be stored in a database provided in the camera unit 10 or a separate server, similarly to basic rainfall information.

Here, it is preferable that the basic rainfall information and the image are matched by time and stored as one data set in a database or a separate server, but is not limited thereto.

In the learning step (S12), deep learning learning may be performed based on basic rainfall information and images. The learning step (S12) may include a rainfall feature extraction step (S120) and a rainfall feature learning step (S121), and binary classification that derives 'rainfall' or 'non-rainfall' as a result value through basic rainfall information and images. A model (Binary Classification Model) can be created.

In the rainfall feature extraction step (S120), basic rainfall information and rainfall features of the image may be extracted through a convolutional neural network (CNN). CNN consists of a convolutional layer that extracts rainfall features from basic rainfall information and images, which are input data, and a plurality of fully-connected layers that analyze the extracted features. Accordingly, the CNN can classify which category the input data belongs to according to the rainfall characteristics.

Here, CNN (Convolutional Neural Network) simulates the human brain function based on the assumption that when a person recognizes an object, the basic features of the object are extracted, and then the object is recognized based on the result of complex calculation in the brain. is a model CNN preferably uses any one of a 2D convolutional neural network, a 3D convolutional neural network, and a virtual 3D convolutional neural network, but is not limited thereto.

In addition, the convolutional layer may execute various configurations and functions including, but not limited to, any one of AlexNet, VGGNet, LeNet, ResNet, and Relu.

In addition, the rain feature is a feature extracted for learning from CNN, and may include one or more of wave patterns formed on the ground, rain drops, rain sounds, rain streaks, umbrellas, and rain drop traces. The umbrella may be in an open state, and the traces of raindrops may be the surface of a subject whose color is darkened by raindrops, but is not limited thereto.

Specifically, rain streaks have different visibility depending on the intensity of rainfall, and umbrellas can be confused with parasols used in bright weather, so it is preferable to extract them together with other rainfall features.

In the rainfall feature learning step (S121), rainfall features extracted from a convolutional neural network (CNN) may be learned through a deep neural network (DNN). In the rainfall feature learning step (S121), the rain type is determined using the rainfall feature through the DNN, and the extracted rainfall feature and the determined rain type are learned to output a result value of 'rainfall' or 'rainfall'. .

Here, the type of rain may be classified as 1 to 4 mm, 5 to 9 mm, 10 mm, 20 mm, 50 mm, 100 mm, etc. according to the amount of precipitation by the Korea Meteorological Administration, and the numerical value of the amount of precipitation may be changed according to the user's request. . In the rainfall feature learning step (S121), one of rain types may be determined according to the rainfall feature through the DNN, but is not limited thereto.

Here, a deep neural network (DNN) is an artificial neural network consisting of a plurality of hidden layers between an input layer and an output layer.

The number of nodes in the input layer of the DNN of the present invention may be the same as the number of rain features, which are input data, and the number of nodes in the output layer may be divided into two because 'rainfall' or 'non-rainfall' is classified. The hidden layer may consist of 1 to 4 layers, but is not limited thereto.

In the rainfall information generating step (S13), rainfall information may be generated by analyzing the image based on deep learning. The rainfall information generating step (S13) may include a preprocessing step (S130) and a rainfall determination step (S131).

The pre-processing step ( S130 ) may be a step of pre-processing an image to accurately generate rainfall information. In the preprocessing step (S130), size adjustment, cropping, enlargement, brightness adjustment, saturation adjustment, and color adjustment may be used, and image preprocessing may be performed by selectively applying one or more depending on the situation.

Here, resizing is a method of converting the size of a photo to a desired size, and cropping is a method of cutting out leaving only a desired portion, which may include free area cropping and cropping by various ratios. Magnification is a method of converting a desired part to appear larger. Brightness control is a method of converting a photo to desired brightness, saturation control is a method of maintaining color balance by adjusting the sharpness of colors, and color control may be a method of restoring colors distorted by light.

In the rain determination step (S131), rainfall information may be generated by learning the preprocessed image and determining it as 'rainfall' or 'non-rainfall'. Specifically, the image may be classified as 'rainfall' or 'nonrainfall' through a binary classification model. Here, the classification process is the same as the learning step (S12), so the description will be omitted.

In the moving step (S2), the sunshield 20 may be moved forward and backward according to pre-stored sunrise or sunset time information and rainfall information. To this end, the moving step (S2) may include a rainfall occurrence information generating step (S20), sunlight information calculating step (S21), and movement control step (S22).

In the rain occurrence information generating step (S20), rainfall occurrence information may be generated according to the rainfall information generated in the information generating step (S1). In the rainfall occurrence information generating step (S20), when 'rainfall' is classified among the rainfall information, rainfall occurrence information may be generated.

In the sunlight information calculation step S21 , angle information of sunlight incident on the lens of the camera unit 10 may be calculated through pre-stored sunrise or sunset time information and GPS information of the camera unit 10 .

Here, the sunrise time information and the sunset time information may be weather information collected from national organizations such as meteorological companies and the Korea Meteorological Administration, and most preferably may be weather information from a weather center in the area where the surveillance camera is installed, and the camera unit ( 10) can be stored in the database or a separate server.

In addition, the GPS information is location information of the monitoring camera device, and through this, the altitude of the sun according to time can be calculated, so that the angle of sunlight incident on the lens of the camera unit 10 can be calculated. At this time, it is preferable that the GPS of the camera unit 10 is provided close to the lens to minimize errors in sunlight angle information, but is not limited thereto.

In the movement control step (S22), the sunshield 20 may be moved forward and backward by calculating a movement distance through at least one of rainfall occurrence information and sunlight angle information. In the movement control step (S22), the lens of the camera unit 10 can be moved in the forward and backward directions so that rainwater does not flow or sunlight does not enter the lens, and the sunshield is moved through a control module provided in the camera unit 10 or a separate control means. (20) can be controlled.

In addition, in the movement control step (S22), when the sunshield 20 is moved based on both the rainfall occurrence information and the sunlight angle information, it can be moved according to the larger value among the respective movement distances.

In the limiting step (S3), the moving distance of the sunshield 20 may be limited by calculating a limiting distance at which a blind spot does not occur according to a relationship between zoom values when the sunshield 20 is moved.

In the limiting step (S3), only the movement distance calculated according to one or more of the rainfall occurrence information and sunlight angle information in the movement control step (S22) is considered, and when the sunshield 20 moves, the blind spot is corrected. It may be a step to

Specifically, in the limiting step (S3), a limiting distance at which blind spots do not occur may be calculated by using the zoom of the camera unit 10, and the limiting distance may be shorter than the moving distance.

In the limiting step (S3), it is preferable to calculate the limiting distance so that an area image of 90% or more of the image before magnification can be obtained even if the zoom value is applied, but it is not limited thereto.

In this case, zoom may be magnification of an image by operation of a lens of the camera unit 10 .

In addition, the control method of the surveillance camera operating with the sunshield installed according to the embodiment of the present invention may further include a region determination step (S4) after the limiting step (S3).

In the area determination step S4, when a blind spot occurs after the limiting step S3 is finished, the blind spot may be determined as a static area or a dynamic area. To this end, the region determination step S4 may include a static determination step and a dynamic determination step.

In the static judgment step, if the blind spot is determined to be a static area, the blind spot can be eliminated through zoom. The static judgment step may allow the user to secure an image without blind spots.

In the dynamic determination step, when the blind spot is determined to be a dynamic region, it may be determined as an error and the information generation step (S1) may be performed again. In the dynamic determination step, when it is difficult to secure an image because the dynamic region corresponding to the motion of the subject is covered by a blind spot, the information generation step (S1) can be performed again so that an accurate image can be captured.

In this way, through the above steps, the method for controlling a surveillance camera with a sunshield installed and operating according to an embodiment of the present invention moves the sunshield 20 forward and backward to prevent rainwater from flowing to the lens part, It is possible to prevent sunlight from entering the lens part by moving the sunshield 20 according to the angle of sunlight, and to prevent blind spots from occurring by limiting the moving distance of the sunshield 20 according to the zoom value. can

Although the embodiments of the present invention have been described with reference to the accompanying drawings, those skilled in the art can implement them in other specific forms without changing the technical spirit or essential features of the present invention. you will be able to understand Therefore, the embodiments described above are illustrative in all respects and are not restrictive.

10: camera unit
11: shooting department
12: fixed part
13: angle adjustment unit
20 : Sunshield
21: bend
30: lower barrier
40: anti-glare sheet

Claims (10)

In the control method of a monitoring camera device in which a front and rear movable sunshield is mounted on one side of the camera unit and operates,
By analyzing the image captured through the camera unit, a dynamic area and a static area are distinguished, and even if a blind spot occurs in the static area, the moving distance of the sunshield is not limited.
The blind spot is
Characterized in that the sunshield is exposed to the lens of the camera unit and the sunshield is exposed to the image,
an information generating step of capturing and analyzing the image by the camera unit to generate rainfall information;
A moving step of moving the sunshield forward and backward according to the pre-stored sunrise or sunset time information and the rainfall information; and
A limiting step of limiting the movement distance of the sunshield by calculating a limit distance at which the blind spot does not occur according to a relationship between zoom values when the sunshield is moved,
The control method of the surveillance camera,
Further comprising an area determination step of determining the blind spot as a static area or a dynamic area when the blind spot occurs after the limiting step is finished,
In the area determination step,
A static determination step of eliminating the blind spot through zoom when the blind spot is determined to be the static area, and
A control method for a surveillance camera device operating with a sunshield mounted thereon, comprising a dynamic determination step of determining that the blind spot is an error and proceeding with an information generation step again when the blind spot is determined to be the dynamic region.
delete According to claim 1,
In the information generation step,
Basic information collection step of collecting basic rainfall information to be used for deep learning learning;
an image collection step of acquiring and collecting the image captured by the camera unit;
A learning step of conducting deep learning learning based on the basic rainfall information and the image; and
A method for controlling a surveillance camera device operating with a sunshield mounted thereon, comprising a rainfall information generation step of generating the rainfall information by analyzing the image based on deep learning.
According to claim 3,
In the learning phase,
A rainfall feature extraction step of extracting the basic rainfall information and the rainfall features of the image through a convolutional neural network (CNN); and
A method for controlling a surveillance camera device operating with a sunshield mounted thereon, comprising a rainfall feature learning step of learning the rainfall feature extracted from the CNN (convolutional neural network) through a DNN (deep neural network).
According to claim 3,
The rainfall information generating step,
A pre-processing step of pre-processing the image; and
A method for controlling a surveillance camera device operating with a sunshield mounted thereon, comprising a rain determination step of learning the preprocessed image and determining it as 'rainfall' or 'non-rainfall' to generate the rainfall information.
According to claim 1,
The moving step is
a rainfall occurrence information generating step of generating rainfall occurrence information according to the rainfall information generated from the information generating step;
A sunlight information calculation step of calculating sunlight angle information incident on a lens of the camera unit through pre-stored sunrise or sunset time information and GPS information of the camera unit; and
and a movement control step of moving the sunshield forward and backward based on at least one of the rainfall occurrence information and the sunlight angle information.

delete delete delete delete

Images