KR20200138465A - Smart cctv system for analysis of waste - Google Patents

Abstract

The present invention relates to an intelligent CCTV system having a waste analysis function. According to one embodiment of the present invention, the intelligent CCTV system having the waste analysis function comprises: one or more CCTV devices for photographing one or more collection boxes or spaces wherein waste is discharged and accommodated; and an image server configured to store information on images photographed by the one or more CCTV devices, wherein the CCTV device may comprise: a photographing unit configured to photograph the one or more collection boxes or spaces; and an operation unit configured to generate a database for extracting features of the waste by using images photographed by the photographing unit, train a model regarding the features of the waste by using the generated database, and calculate and output features of the waste accommodated in the one or more collection boxes or spaces according to a result of the training. According to the present invention, a type and amount of waste discharged to a collection box or space can be confirmed by using images by photographing the collection box or space for collecting waste by using a camera such as a CCTV and the like.

Description

Waste analysis intelligent CCTV system {SMART CCTV SYSTEM FOR ANALYSIS OF WASTE}

The present invention relates to a waste analysis intelligent CCTV system, and more particularly, to a waste analysis intelligent CCTV system capable of analyzing the type and amount of waste accommodated in the collection box for the collection of waste.

With increasing awareness of global warming and environmental destruction, efforts to recycle resources for reduction of carbon dioxide, resource conservation, and environmental protection are being made in various fields. In order to make efforts to recycle these resources more effectively, a policy is in place to classify and discharge wastes by type.

In general, to separate and dispose of waste, a separate collection box or space is provided according to the type of waste so that citizens can voluntarily separate and input the type of waste.

When such a waste collection box or space is filled over a certain amount, it must be transferred to a waste treatment facility using a vehicle to collect it, but it is difficult to analyze the amount of discharge by type of waste.

Conventionally, when waste is filled to a predetermined height or higher in a collection box or space using an ultrasonic sensor or the like, a technique has been applied to confirm that the waste is filled in the collection box or space. However, due to the nature of the ultrasonic sensor, if waste is accumulated at a predetermined height at a specific location, errors such as indicating that there is no more space to be filled with waste may occur in the collection box or the space even though there is a space to be filled with waste at another location. .

In addition, since the ultrasonic sensor does not recognize objects, there is a problem in that it cannot determine the type of waste and cannot check the cleanliness of a collection box or space.

Republic of Korea Patent Publication No. 10-2018-0086164 (July 30, 2018) Republic of Korea Patent Publication No. 10-2017-0017652 (2017.02.15.) Republic of Korea Patent Publication No. 10-2012-0118226 (2012.10.26.)

The problem to be solved by the present invention is to provide a waste analysis intelligent CCTV system that can check whether waste is normally separated and discharged in a waste collection box or space installed at a place for collecting waste, and also check the discharge amount of waste. .

Waste analysis intelligent CCTV system according to an embodiment of the present invention, one or more CCTV devices for photographing one or more collection boxes or spaces in which waste is discharged and accommodated; And an image server for storing information on images captured by the one or more CCTV devices, wherein the CCTV device includes: a photographing unit for photographing the one or more collection boxes or spaces; And generating a database for extracting characteristics of the waste using the image or image captured by the photographing unit, learning a model for the characteristics of the waste using the generated database, and learning the one It may include a calculation unit that calculates and outputs the characteristics of the waste accommodated in the above collection box or space.

The CCTV device may further include a storage unit for storing the characteristics of the waste fire calculated by the operation unit.

The characteristic of the waste calculated by the operation unit may be the type of waste or discharge amount of the waste accommodated in the one or more collection boxes or spaces.

The operation unit extracts category features for waste in the one or more collection boxes or spaces using the database, learns a model for waste characteristics in the one or more collection boxes or spaces using the extracted category features, and A model for waste characteristics can be used to compute results for the one or more waste characteristics.

The operation unit may learn a model for the characteristics of the waste based on a deep learning convolutional neural network (CNN) for image recognition.

On the other hand, the waste analysis intelligent CCTV system according to an embodiment of the present invention, one or more CCTV devices for photographing one or more collection boxes or spaces in which waste is discharged and accommodated; And an image server for storing information on images captured by the one or more CCTV devices, wherein the image server receives images or images captured by the one or more CCTV devices, and uses the received images or images. Generates a database for extracting characteristics of the waste, learns a model for the characteristics of the waste using the generated database, and calculates and outputs the characteristics of the waste accommodated in the one or more collection boxes or spaces according to the learned result It may include a server operation unit.

The characteristic of the waste calculated by the server operation unit may be the type of waste or discharge amount of waste accommodated in the one or more collection boxes or spaces.

The server operation unit extracts category features for waste in the one or more collection boxes or spaces using the database, and learns a model for the waste characteristics of the one or more collection boxes or spaces using the extracted category features, and learns The models for the waste features that have been generated can be used to compute results for the one or more waste features.

The server operation unit may learn a model for the characteristics of the waste based on a deep learning convolutional neural network (CNN) for image recognition.

The CCTV device, a photographing unit for photographing the one or more collection boxes or spaces; And generating a database for extracting characteristics of the waste using the image or image captured by the photographing unit, learning a model for the characteristics of the waste using the generated database, and learning the one And a calculation unit that calculates and outputs characteristics of wastes accommodated in the above collection box or space, wherein the image server receives the image or image captured by the photographing unit, and the server calculation unit includes the calculated result in the CCTV device. It can be transmitted to the calculation unit.

According to the present invention, by photographing a collection box or space for collecting waste by using a camera such as CCTV, there is an effect that the type of waste discharged to the collection box or space and the amount of waste discharged can be confirmed through an image.

In addition, by analyzing the image captured using CCTV, the type of waste can be recognized and the amount of waste discharged can be analyzed, so that the optimal route of vehicles for collecting waste can be calculated or the status of the discharge of waste can be analyzed by region. It works.

1 is a schematic diagram showing a waste analysis intelligent CCTV system according to an embodiment of the present invention.
2 is a block diagram showing an intelligent CCTV system for waste analysis according to an embodiment of the present invention.
3 is a view for explaining a concept for performing deep learning in the waste analysis intelligent CCTV system according to an embodiment of the present invention.
4 is a view showing an example of the analysis result of the collection box through the integrated terminal of the waste analysis intelligent CCTV system according to an embodiment of the present invention.
5 to 9 are views for explaining the result of analyzing the discharge amount of waste in the waste analysis intelligent CCTV system according to an embodiment of the present invention.
10 to 12 are views for explaining a result of analyzing the type of waste in the waste analysis intelligent CCTV system according to an embodiment of the present invention.
13 is a block diagram showing a waste analysis intelligent CCTV system according to another embodiment of the present invention.

A preferred embodiment of the present invention will be described in more detail with reference to the accompanying drawings.

1 is a schematic diagram showing a waste analysis intelligent CCTV system according to an embodiment of the present invention, Figure 2 is a block diagram showing a waste analysis intelligent CCTV system according to an embodiment of the present invention. 3 is a view for explaining a concept for performing deep learning in the waste analysis intelligent CCTV system according to an embodiment of the present invention.

1 and 2, the waste analysis intelligent CCTV system 100 according to an embodiment of the present invention includes a CCTV device 110, an image server 120, and a manager terminal 130.

CCTV device 110, as shown, may be provided with a plurality, each CCTV device 110, may be installed to photograph the inside of one collection box (WB). At this time, in this embodiment, it is described that one CCTV device 110 photographs one collection box (WB), but is not limited thereto, and one CCTV device 110 may photograph a space where waste is collected. have. That is, it may be a place where specific waste can be collected, not a specific box-shaped bin where waste is collected, such as a collection box (WB).

And in the present embodiment, when the collection box (WB) is used, a configuration capable of measuring the weight of the collection box (WB) may be included, including photographing in the CCTV device 110 as necessary. To measure the weight of the collection box WB, a device capable of measuring weight may be installed at each lower portion of the collection box WB. The weight of the collection box WB is measured, and information on the measured weight of the collection box WB may be transmitted to the manager terminal 130.

By measuring the weight of the collection box WB, the amount of waste discharged to the collection box WB can be measured. In this embodiment, even if the weighing device is optionally provided, it is possible to check the discharge amount of waste filling the collection bin (WB) by using the image captured by the CCTV device 110. To this end, in this embodiment, the CCTV device 110 includes a photographing unit 112, an image unit, and a storage unit 116.

The photographing unit 112 may photograph the inside of the collection box WB. At this time, one CCTV device 110 may be installed in one collection box WB, and the inside of the collection box WB may be photographed by the photographing unit 112 included in the CCTV device 110. In addition, in the present embodiment, the photographing unit 112 may photograph the inside of the collection box WB at daytime and at night, and may include a camera for daytime photographing and an infrared camera for night photographing.

Since the photographing unit 112 can photograph the inside of the collection box WB, in this embodiment, the CCTV device 110 may be fixedly installed on the upper part of the collection box WB, as shown in FIG. 1. Here, a plurality of CCTV devices 110 are provided to describe each photographing of a plurality of collection boxes (WB), but is not limited thereto. If necessary, one CCTV device 110 is provided, and one CCTV device 110 may be configured to photograph a plurality of collection boxes (WB). That is, one CCTV device 110 may be installed to move to a location where a plurality of collection boxes WB are arranged to sequentially photograph the interior of the plurality of collection boxes WB.

The operation unit 114 may analyze the image photographed by the photographing unit 112 to perform analysis on the type of waste and the capacity of the waste accommodated in the collection box WB. In this embodiment, the calculation unit 114 may analyze the type of waste accommodated in the collection box WB and the capacity of the waste through deep learning learning. In this case, a detailed description of the deep learning learning model will be described later, but the type of waste and the waste capacity in the collection bin (WB) can be analyzed through a model based on a deep learning convoloutional neural network (CNN).

To this end, the operation unit 114 may be configured with hardware including computing resources (CPU, GPU, RAM, storage device, etc.) so that analysis can be performed through a program based on a deep learning convolutional neural network. In addition, a model, analysis program, waste analysis information, and the like for the type of waste and the capacity of the collection box WB may be stored in the storage device included in the operation unit 114.

Referring to FIG. 3, first, the operation unit 114 builds a database based on input data on the type of waste and the amount of waste discharged at a specific place where waste is collected. And, from the database built for learning based on convolutional neural networks, category features for waste types and waste emissions are extracted (Convolutions). The type of waste and the amount of waste can be extracted, respectively, by using the analyzed model of the type of waste and the amount of waste that has been learned in this way. At this time, the learned model can secure an image or image of the waste type and waste discharge amount of the collection bin (WB).

Here, in the process of learning the analysis model of waste type and waste discharge, various images or images of various weather changes (e.g., snow or rain) or time changes (day and night changes) for the waste collection location And, it is possible to learn an analysis model for waste types and waste discharge through the secured images or images.

In addition, when extracting the learned waste type and waste discharge amount, analyze the actual type of waste and discharge amount by using images or images of the waste collection box (WB), and if an error occurs, the waste type and waste discharge amount It is possible to minimize the occurrence of errors by repeating the process of re-extracting the categories for each. Therefore, the calculation unit 114 may finally calculate the type of waste and information on the amount of waste discharged.

In this embodiment, the convolutional neural network is a statistical learning algorithm like a biological neural network in the field of machine learning and cognitive science. The convolutional neural network may refer to an overall model that has the ability to solve unrestrictedly as artificial neurons (nodes), which form a network through synaptic bonding, change the synaptic bonding strength through learning.

In this embodiment, the operation unit 114 included in the CCTV device 110 is a variety of neural networks used for image recognition and object recognition analysis using a convolutional neural network model (eg, LeNET, AlexNet, ZFNet, GooLeNet, ResNet, etc.) Through this, analysis models for waste types and waste emissions can be created.

That is, an image or an image captured by the photographing unit 112 may be input as input data shown in FIG. 3. In addition, by using the image or image input from the operation unit 114, the type of waste and the amount of waste received in the collection box WB may be analyzed, and the corresponding information may be output.

In this way, information on the type of waste and discharge amount of the waste calculated by the operation unit 114 may be transmitted to the image server 120.

The storage unit 116 stores the result calculated by the operation unit 114. At this time, the storage unit 116 may be provided in a configuration independent from the storage device included in the calculation unit 114, but is not limited thereto, and the storage device and the storage unit 116 included in the calculation unit 114 as necessary. May be provided in one configuration.

As described above, the image server 120 may receive calculation results for waste types and waste discharges calculated by the calculation unit 114 included in a plurality of CCTV devices 110, and collect the received results. It can be saved as information about (WB).

The manager terminal 130 is provided to manage the current state of waste discharge at the waste collection site. The manager terminal 130 may analyze and manage a waste collection box (WB) based on information transmitted from the CCTV device 110 and the image server 120. In this embodiment, the manager terminal 130 is the optimal movement path of the vehicle for collecting the waste contained in each collection box (WB) based on the information on the discharge amount according to the waste for each collection box (WB) of the waste collection location Can be set.

4 is a view showing an example of the analysis result of the collection box through the integrated terminal of the waste analysis intelligent CCTV system according to an embodiment of the present invention.

Referring to FIG. 4, it is confirmed that the capacity received in the collection box WB is analyzed using images for each waste in the collection box (WB) using the waste analysis intelligent CCTV system 100 according to an embodiment of the present invention. I can.

It can be seen that can-type waste is accommodated in the first collection bin (WB), and about 22% of the collection bin (WB) is filled with can-type waste. And it can be seen that plastic-type waste is accommodated in the second collection bin (WB), and about 83% of the collection bin (WB) is filled with plastic-type waste. In addition, it can be seen that paper-type waste is accommodated in the third collection bin (WB), and about 42% of the collection bin (WB) is filled with paper-type waste. And it can be seen that the fourth collection bin (WB) contains the waste of general trash.

Therefore, you can check how much recyclable waste is in the collection box (WB) and how much waste in the pay-as-you-go (general garbage) is in the collection box (WB), and how many of the loads of all collection boxes (WB) are I can confirm.

By checking how much waste is filled in each collection box (WB) by using the image captured in real time each collection box (WB) in this way, the manager terminal 130 collects the waste at the collection location where the corresponding collection box (WB) is placed. You can check in advance whether to do it.

5 to 9 are views for explaining the result of analyzing the discharge amount of waste in the waste analysis intelligent CCTV system according to an embodiment of the present invention.

5 to 9, about calculating the amount of waste discharged to the collection box (WB) using a deep learning convolutional neural network through the waste analysis intelligent CCTV system 100 according to an embodiment of the present invention. Explain. In this embodiment, it will be described that the amount of waste discharged to the collection box (WB) is divided into five stages of 20%, 40%, 60%, 80%, and 100% of the collection box (WB). However, the present invention is not limited thereto, and may be divided into more subdivided stages as needed, and may be divided into 1% stages.

Referring to FIG. 5, the probability for each step for the steps of 20%, 40%, 60%, 80% and 100% in the collection box (WB) of FIG. 5A is shown in FIG. 5B. As described above, it can be shown that it is 0.8725821 at 20%, 0.099532016 at 60%, 0.027615786 at 40%, 0.00014902806 at 80%, and 0.00012094527 at 100%. At this time, it can be seen that 0.492s was used for the time required.

Therefore, it is judged to be 20% with the highest probability, indicating that the waste has filled about 20% of the collection bin (WB).

Referring to FIG. 6, the probability for each step is shown in FIG. 6(b) for the steps of 20%, 40%, 60%, 80% and 100% in the collection box (WB) of FIG. 6(a). As described above, it can be shown to be 0.8305835 at 40%, 0.1431199 at 20%, 0.025084553 at 60%, 0.0010898137 at 80%, and 0.00012224758 at 100%. At this time, it can be seen that 0.373s was used for the time required.

Therefore, it is judged to be 40% with the highest probability, indicating that the waste has filled approximately 40% of the collection bin (WB).

Referring to FIG. 7, the probability for each step is shown in FIG. 7 (b) for the steps of 20%, 40%, 60%, 80% and 100% in the collection box (WB) of FIG. 7 (a). As described above, it can be shown to be 0.9658261 at 60%, 0.0234172 at 20%, 0.008906304 at 100%, 0.000983298 at 40%, and 0.0008670432 at 80%. At this time, it can be seen that 0.355s was used for the time required.

Therefore, it is judged to be 60% with the highest probability, which indicates that the waste has filled about 60% of the collection bin (WB).

Referring to FIG. 8, the probability for each step for the steps of 20%, 40%, 60%, 80% and 100% in the collection box (WB) of FIG. 8 (a) is shown in FIG. 8 (b). As described above, it can be shown to be 0.9572164 at 80%, 0.030484878 at 100%, 0.008193274 at 20%, 0.0028198282 at 60%, and 0.001285649 at 40%. At this time, it can be seen that 0.358s was used for the time required.

Therefore, it is judged to be 80% with the highest probability, indicating that the waste has filled about 80% of the collection bin (WB).

Referring to FIG. 9, the probability for each step for the steps of 20%, 40%, 60%, 80% and 100% in the collection box (WB) of FIG. 9(a) is shown in FIG. 9(b). As described above, it can be shown to be 0.9915324 at 100%, 0.005785733 at 60%, 0.0025920623 at 80%, 6.868307e-05 at 20%, and 2.1143418e-05 at 40%. At this time, it can be seen that 0.355s was used for the time required.

Therefore, it is judged to be 100% with the highest probability, which indicates that the waste has filled approximately 100% of the collection bin (WB).

10 to 12 are views for explaining a result of analyzing the type of waste in the waste analysis intelligent CCTV system according to an embodiment of the present invention.

With reference to FIGS. 10 to 12, a description will be given of calculating the type of waste discharged to the collection box (WB) using a deep learning convolutional neural network through the waste analysis intelligent CCTV according to an embodiment of the present invention. In this embodiment, a description will be given of separating general garbage, plastic and paper. However, the present invention is not limited thereto, and further subdivided types may be classified as necessary.

Referring to FIG. 10, with respect to the type of waste accommodated in the collection box (WB) of FIG. 10(a), as shown in FIG. 10(b), the probability of a general garbage waste is 0.7413017, and the probability of a plastic waste is It was found to be 0.2538774, and the probability of paper waste was 0.0048224716, and at this time, it could be confirmed that the spent time was 0.337s.

Accordingly, it can be determined that the general garbage waste is accommodated in the collection box WB shown in FIG. 10A.

Referring to FIG. 11, with respect to the type of waste accommodated in the collection box (WB) of FIG. 11(a), as shown in FIG. 11(b), the probability of paper waste is 0.996872, and the probability of plastic waste is 0.0030095605 It was found that the probability of being a general trash waste was 0.00011851702, and at this time, it was confirmed that the spent time was 0.344s.

Therefore, it can be determined that the paper waste is accommodated in the collection box WB shown in (a) of FIG. 11.

Referring to FIG. 12, for the type of waste accommodated in the collection box (WB) of FIG. 12 (a), as shown in FIG. 12 (b), the probability of plastic waste is 0.9160764, and the probability of paper waste is 0.08361702. It was found that the probability of being a general trash waste was 0.000306574007, and at this time, it was confirmed that the spent time was 0.356s.

Therefore, it can be determined that the plastic waste is accommodated in the collection box WB shown in FIG. 12A.

13 is a block diagram showing a waste analysis intelligent CCTV system according to another embodiment of the present invention.

Referring to FIG. 13, a waste analysis intelligent CCTV system 100 according to another embodiment of the present invention includes a CCTV device 110, an image server 120, and a manager terminal 130. While describing this embodiment, the same description as in the embodiment will be omitted.

In this embodiment, the CCTV device 110 may include only the photographing unit 112. Therefore, a number of CCTV devices 110 photograph each collection box (WB) or space for discharging waste.

In addition, the image captured by the CCTV device 110 may be transmitted to the image server 120. In this case, if necessary, a storage unit 116 for storing an image captured by the CCTV device 110 may be additionally included.

The image server 120 may include a server operation unit 122 and a server storage unit 124 to analyze an image transmitted from the CCTV device 110.

The server operation unit 122 may receive images taken of a collection box (WB) or a space from a plurality of CCTV devices 110 in real time, and calculate a waste type and waste discharge amount for each received image. In addition, the calculated result may be stored in the server storage unit 124 and transmitted to the manager terminal 130.

In this case, the image server 120 may perform optimal management based on the information on the amount of discharged by the type of waste, such as hourly, daily, monthly, and yearly calculated by the server operation unit 122.

In this embodiment, the server operation unit 122 continuously updates and updates information through a convolutional neural network by modeling the type of waste and the amount of waste received in each waste collection box (WB) disposed at the waste collection location based on deep learning. The most accurate model can be maintained according to each model.

The server operation unit 122 receives an image or image photographed by CCTV, and builds a database on the waste type and waste discharge amount from the received image or image. And, based on the convolutional neural network, category features for waste types and waste emissions are extracted from the database for learning. Waste types and waste discharges can be extracted by using the analyzed model for waste types and waste discharges learned in this way.

In addition, as the server operation unit 122 uses the image or image captured by the CCTV device 110 as it is, analysis of various images or images for various weather changes and time changes of the location placed in the collection box (WB) is made. I can.

In addition, when an error occurs when analyzing the type of waste and the amount of waste, the server operation unit 122 re-extracts the category characteristics for the type of waste and the amount of waste from the database for learning based on a convolutional neural network, and an analysis model The type of waste and the amount of waste can be calculated using.

As described above, in the present embodiment, an image or image captured through the CCTV device 110 may be analyzed by the server operation unit 122 included in the image server 120, but is not limited thereto. As in, the CCTV device 110 includes an operation unit 114, and, as in other embodiments, the image server 120 includes a server operation unit 122, so that the CCTV device 110 and the image server 120 All operations can be performed at

When the operation is performed in both the CCTV device 110 and the image server 120 in this way, the operation on the image or image captured by the CCTV device 110 may be more efficiently performed.

In this case, the image server 120 may receive an image or image captured by the CCTV device 110, and receive an operation result for the type of waste and the amount of waste analyzed by the operation unit 114 of the CCTV device 110. have. In addition, based on the analysis information of the type of waste and the amount of waste received by the operation unit 114 of the CCTV device 110, the discharge status of the waste may be analyzed. In addition, the server operation unit 122 of the image server 120 uses the image or image received from the photographing unit 112 of the CCTV device 110 and the analysis information received from the CCTV operation unit 114 to generate a convolutional neural network. Based on this, the analysis learning model of waste types and waste discharges can be updated.

In this way, the server operation unit 122 updates the analysis learning model and transmits the analysis result of the waste type and waste discharge to the CCTV device 110 and the manager terminal 130.

Alternatively, the server operation unit 122 generates an analysis learning model for waste types and waste discharge through associative analysis based on a convolutional neural network, and collects images or images captured by a plurality of CCTV devices 110. In addition, the type of waste and the amount of waste can be calculated using the collected images or images.

As described above, a detailed description of the present invention has been made by an embodiment with reference to the accompanying drawings, but the above-described embodiment has been described with reference to a preferred example of the present invention, so that the present invention is limited to the above embodiment. It should not be understood, and the scope of the present invention should be understood as the following claims and equivalent concepts.

100: intelligent CCTV system
110: CCTV device
112: photographing department
114: operation unit
116: storage unit
120: video server
122: server operation unit
124: server storage
130: administrator terminal
WB: collection box

Claims (10)

One or more CCTV devices for photographing one or more collection boxes or spaces in which waste is discharged and accommodated; And
Including an image server for storing information about the image captured by the one or more CCTV devices,
The CCTV device,
A photographing unit for photographing the at least one collection box or space; And
A database for extracting characteristics of the waste is created using the image or image captured by the photographing unit, and a model for the characteristics of the waste is learned using the generated database, and the at least one Waste analysis intelligent CCTV system including an operation unit that calculates and outputs the characteristics of the waste accommodated in the collection box or space.
The method according to claim 1,
The CCTV device, waste analysis intelligent CCTV system further comprises a storage unit for storing the characteristics of the waste calculated by the calculation unit.
The method according to claim 1,
The characteristic of the waste calculated by the operation unit is the type of waste or discharge amount of waste accommodated in the one or more collection boxes or spaces.
The method according to claim 1,
The operation unit extracts category features for waste in the one or more collection boxes or spaces using the database, learns a model for waste characteristics in the one or more collection boxes or spaces using the extracted category features, and Waste analysis intelligent CCTV system that calculates the results for the one or more waste characteristics using the model for the waste characteristics.
The method according to claim 1,
The operation unit is a waste analysis intelligent CCTV system that learns a model for the characteristics of the waste based on a deep learning convolutional neural network (CNN) for image recognition.
One or more CCTV devices for photographing one or more collection boxes or spaces in which waste is discharged and accommodated; And
Including an image server for storing information about the image captured by the one or more CCTV devices,
The image server receives images or images captured by the one or more CCTV devices, generates a database for extracting characteristics of the waste using the received image or image, and uses the generated database to A waste analysis intelligent CCTV system comprising a server operation unit that learns a model for a feature and calculates and outputs the feature of the waste accommodated in the one or more collection boxes or spaces according to the learned result.
The method of claim 6,
The characteristic of the waste calculated by the server operation unit is the type of waste or discharge amount of waste accommodated in the at least one collection box or space.
The method of claim 6,
The server operation unit extracts category features for waste in the one or more collection boxes or spaces using the database, and learns a model for the waste characteristics of the one or more collection boxes or spaces using the extracted category features, and learns Waste analysis intelligent CCTV system that calculates the results for the one or more waste characteristics using the model for the waste characteristics.
The method of claim 6,
The server operation unit is a waste analysis intelligent CCTV system that learns a model for the characteristics of the waste based on a deep learning convolutional neural network (CNN) for image recognition.
The method of claim 6,
The CCTV device,
A photographing unit for photographing the at least one collection box or space; And
A database for extracting characteristics of the waste is created using the image or image captured by the photographing unit, and a model for the characteristics of the waste is learned using the generated database, and the at least one It includes a calculation unit that calculates and outputs the characteristics of the waste accommodated in the collection box or space,
The image server receives the image or image captured by the photographing unit,
The server operation unit is a waste analysis intelligent CCTV system that transmits the calculated result to the operation unit included in the CCTV device.

Images