CN111646045A

CN111646045A - Four-classification garbage can for intelligently identifying and automatically classifying garbage

Info

Publication number: CN111646045A
Application number: CN202010652282.9A
Authority: CN
Inventors: 赵林; 江明翰; 李希; 易嘉闻; 吴健辉; 胡文静; 张国云
Original assignee: Guangzhou Zhongsheng Intelligent Technology Co ltd; Hunan Institute of Science and Technology
Current assignee: Guangzhou Zhongsheng Intelligent Technology Co ltd; Hunan Institute of Science and Technology
Priority date: 2020-07-08
Filing date: 2020-07-08
Publication date: 2020-09-11

Abstract

The invention belongs to the technical field of garbage cans, and particularly relates to a four-classification garbage can for intelligently identifying and automatically classifying garbage, which comprises a can body, a driving structure and a controller; a partition plate is arranged in the barrel body; the number of the partition plates is three, the partition plates divide the interior of the barrel body into four cavities in the length direction, and a driving structure is arranged at an upper port of the barrel body; the driving structure comprises a first motor, a second motor, a fixing plate, a screw rod, a first sliding block, a second sliding block, a sliding rod, a receiving groove and a touch switch; through the cooperation between a motor and No. two motors, a motor removes the receiving tank and rotates in the corresponding cavity that belongs to rubbish and hold with the rubbish of receiving tank to accomplish waste classification, long service life, the consumption is little, and simple structure realizes easily, and categorised accuracy and classification efficiency are high when guaranteeing the cost.

Description

Four-classification garbage can for intelligently identifying and automatically classifying garbage

Technical Field

The invention belongs to the technical field of garbage cans, and particularly relates to a four-classification garbage can for intelligently identifying and automatically classifying garbage.

Background

In production and life, large machine production and automation appliances greatly improve the production level and the quality of life of people, good instruments also need a high-efficiency mechanical structure while requiring excellent power, at present, along with the continuous and serious garbage problem, a garbage can is used as a first station for garbage treatment, the position is very important, all the more important, if the garbage can be put into the can (classified), the garbage problem is solved at the source, which is regarded as important for the country and people, and simultaneously, more and more inventers and traders are brought to the eyes, various intelligent classification garbage cans are in endless, the current classification mechanical device with comparatively hot fire is mainly provided with double-layer partition plates, rotating discs and the like, the double-layer partition plate device is simple and easy to realize, and consists of two vertically placed partition plates, the rotating directions of the two partition plates are different, a south-north direction and an east-west direction are combined, four classification is completed through four combined rotary dumping modes after recognition is completed by recognition equipment, but because huge waste exists in space utilization, a separating device occupies nearly one half of the volume of a garbage can, if the storage capacity of the garbage can is guaranteed, the garbage can looks too big by the device, the practicability is not high, a disc with a notch is adopted by a rotary disc structure, the notch is usually one half at present, after the recognition is completed by the recognition equipment, a baffle is rotated to a specified position, the garbage is swept down by the cooperation of the rotation and the baffle, and classification is completed, the main defects of the structure have two aspects, one is that the baffle and the disc are coaxial, the structure is easy to age and influence each other, and the classification effect of elastic garbage or garbage which is easy to roll by one disc is always bad, on the other hand, the distance between the baffle and the disc is a big difficulty, when the baffle is too close to the disc (directly attached to the disc), the motor with too large friction is likely to be difficult to operate or generate great sound after the baffle is used for a period of time, and when the baffle is too far away from the disc (a gap is formed between the baffle and the disc), the practical value of the baffle is greatly reduced, if the baffle is tightly attached to the disc or the garbage with thinner thickness cannot be swept down, the baffle and the garbage can be swept down together with other non-similar garbage, the classification is not accurate, and the defects of various structures are overcome.

The invention provides a novel intelligent classification garbage can adopting a screw transmission and motor rotation dumping structure, which uses a screw long rod, a rotating motor of a garbage receiving groove machine is fixed on a screw rod together, the other side of a receiving groove is placed on a slide way, the garbage receiving groove is transmitted to the upper part of a garbage can belonging to the classification through screw transmission after the identification part completes the identification, and the garbage in the receiving groove is dumped to the corresponding garbage can through the rotation of the motor, thereby completing the garbage classification, having long service life, low power consumption, simple and easy structure, accurate classification and high classification efficiency while ensuring the cost.

Disclosure of Invention

In order to make up for the defects of the prior art and solve the problems described in the background art, the invention provides a four-classification garbage can for intelligently identifying and automatically classifying garbage.

The technical scheme adopted by the invention for solving the technical problems is as follows: the invention discloses a four-classification garbage can for intelligently identifying and automatically classifying garbage, which comprises a can body and a controller, wherein a partition plate is arranged in the can body; the number of the partition plates is three, the partition plates divide the interior of the barrel body into four cavities in the length direction, and a driving structure is arranged at an upper port of the barrel body; the driving structure comprises a first motor, a second motor, a fixing plate, a screw rod, a first sliding block, a second sliding block, a sliding rod, a receiving groove and a touch switch; the number of the touch switches is two, and the two touch switches are symmetrically and fixedly connected to the side wall of the barrel body; the number of the fixed plates is two, the fixed plates are symmetrically and fixedly connected to the outer side wall of the barrel body, one end of one fixed plate is fixedly connected with a first motor, an output shaft of the first motor is fixedly connected with one end of a screw rod, the other end of the screw rod is hinged to one end of the other fixed plate, and a first sliding block is sleeved on the outer ring of the screw rod; the first sliding block is connected to the screw rod in a screw rod nut mode, and a second motor is fixedly connected to the first sliding block; the output shaft of the second motor is fixedly connected with a receiving groove; the receiving groove is semi-cylindrical, one side wall of the receiving groove is fixedly connected to an output shaft of a second motor, the other side wall of the receiving groove is provided with a first rod, and the first rod is hinged with a first sliding block; the second sliding block is connected to the sliding rod in a sliding mode, and two ends of the sliding rod are fixedly connected to the other ends of the two fixing plates.

Preferably, the two fixing plates are provided with elastic corrugated plates; one end of the corrugated plate is fixedly connected to the fixed plate, two sides of the other end of the corrugated plate are provided with a plate, and the middle position of the plate is hinged to the output shaft of the second motor and the outer ring of the first rod.

Preferably, the two vertical side walls in the receiving groove are provided with identification devices for identifying the garbage types; the identification device is electrically connected with the touch switch and the controller, and the identification device identifies the garbage by using a convolutional neural network single garbage identification method.

Preferably, the convolutional neural network single garbage identification method is applicable to the garbage can, and the identification method comprises the following steps:

s1: the method comprises the steps that a user firstly puts garbage into a receiving groove, an identification device calculates by using a background difference method to obtain a garbage image with a removed background, the garbage image is input into a multi-channel convolutional neural network model to be calculated layer by layer and then output representative garbage classification characteristics, and finally the garbage image is input into a classification layer of a convolutional neural network to identify the garbage, so that the classification of the garbage is identified, and the garbage classification is obtained;

preferably, the two garbage images in the background subtraction method are images shot by a camera after garbage is dumped for the previous time and images shot after garbage is received, so as to obtain garbage images with removed backgrounds.

Preferably, the convolutional neural network includes an initial layer, an intermediate layer and a classification layer.

Preferably, the initial layer is composed of a convolutional layer with a core size of 7 × 7 and a step size of 2 and a pooling layer with a core size of 3 × 3 and a step size of 2. The middle layer consists of N similar residual blocks (N is more than or equal to 16 and less than or equal to 33); the classification layer is composed of an average pooling layer and a full connection layer, the activation function of the full connection layer is softmax, the probability vector of the finally output garbage identification is Q, and the dimensionality of the vector Q is the number of garbage items in the garbage database.

Preferably, the classification layer queries a "garbage" database according to the identified garbage category to obtain the classification to which the garbage belongs.

The invention has the technical effects and advantages that:

through the cooperation between a motor and No. two motors, a motor removes the receiving tank and rotates in the corresponding cavity that belongs to rubbish and hold with the rubbish of receiving tank to accomplish waste classification, long service life, the consumption is little, and simple structure realizes easily, and categorised accuracy and classification efficiency are high when guaranteeing the cost.

The buckled plate through port on the staving removes along with the receiving tank, and the buckled plate covers port on to the staving, avoids the rainwater drippage in the staving to and prevent that rubbish from directly throwing into the staving after not classifying, cause the emergence of rubbish classification failure phenomenon.

Through the cooperation between the convolutional neural network initialization layer, the middle feature extraction layer and the classification layer, the garbage classification is accurate and the classification efficiency is high, the user operation is simple and convenient, and the trouble of the user to the garbage classification is solved.

Drawings

The invention is further described with reference to the following figures and embodiments.

FIG. 1 is a perspective view of the present invention;

FIG. 2 is an enlarged view of a portion of FIG. 1 at A;

FIG. 3 is a view showing the inner structure of the tub in the present invention;

fig. 4 is a diagram of the residual block-like structure of the present invention.

In the figure: the electric heating barrel comprises a barrel body 1, a partition plate 11, a driving structure 2, a first motor 21, a second motor 22, a fixing plate 23, a corrugated plate 231, a first plate 232, a screw rod 24, a first sliding block 25, a second sliding block 26, a sliding rod 27, a receiving groove 28, a first rod 281, a recognition device 282, a touch switch 29 and a controller 3.

Detailed Description

In order to make the technical means, the creation features, the achievement purposes and the effects of the invention easy to understand, the invention is further described with the following embodiments.

As shown in fig. 1 to 4, the four-classification trash can for intelligently identifying and automatically classifying trash according to the present invention includes a can body 1 and a controller 3; the barrel body 1 is internally provided with a partition plate 11; the number of the partition plates 11 is three, the partition plates 11 evenly divide the interior of the barrel body 1 into four cavities in the length direction, and the upper port of the barrel body 1 is provided with a driving structure 2; the driving structure 2 comprises a first motor 21, a second motor 22, a fixing plate 23, a screw rod 24, a first sliding block 25, a second sliding block 26, a sliding rod 27, a receiving groove 28 and a touch switch 29; the number of the touch switches 29 is two, and the two touch switches 29 are symmetrically and fixedly connected on the side wall of the barrel body 1; the number of the fixing plates 23 is two, the fixing plates 23 are symmetrically and fixedly connected to the outer side wall of the barrel body 1, one end of one fixing plate 23 is fixedly connected with a first motor 21, an output shaft of the first motor 21 is fixedly connected with one end of a screw rod 24, the other end of the screw rod 24 is hinged to one end of the other fixing plate 23, and a first sliding block 25 is sleeved on the outer ring of the screw rod 24; the first sliding block 25 is connected to the screw rod 24 in a screw rod nut mode, and the second motor 22 is fixedly connected to the first sliding block 25; the output shaft of the second motor 22 is fixedly connected with a receiving groove 28; the receiving groove 28 is semi-cylindrical, one side wall of the receiving groove 28 is fixedly connected with an output shaft of the second motor 22, the other side wall of the receiving groove 28 is provided with a first rod 281, and the first rod 281 is hinged with the second sliding block 26; the second sliding block 26 is slidably connected to the sliding rod 27, and two ends of the sliding rod 27 are fixedly connected to the other ends of the two fixing plates 23.

The two fixing plates 23 are provided with elastic corrugated plates 231; corrugated plate 231 one end rigid coupling is equipped with a board 232 on fixed plate 23, the other end both sides limit of corrugated plate 231, and a board 232 intermediate position articulates on No. two motor 22's output shaft and on the outer lane of a pole 281.

In an embodiment of the present invention, an identification device 282 for identifying a type of garbage is disposed in the receiving groove 28; in this embodiment, an infrared sensing device may be further combined, when the human body approaches to throw the garbage, the outer cover above the receiving groove 28 is opened, and when the garbage is thrown, the outer cover is closed; under the condition that the outer cover is closed, the LED light source is turned on, the recognition device 282 is used for photographing and recognizing, after recognition is finished, the LED light source is turned off, and the receiving groove 28 is translated and rotated to dump garbage according to the recognition result. The identification device 282 is electrically connected to the touch switch 29 and the controller 3, and the identification device 282 identifies the garbage by using a convolutional neural network single garbage identification method; when the garbage collection device works, garbage is put into the receiving groove 28, then the touch switch 29 is pressed, the touch switch 29 controls the recognition device 282 to recognize the garbage, then the recognition device 282 classifies the garbage by using a convolutional neural network single garbage recognition method, and then the recognition device 282 controls the first motor 21 and the second motor 22 to work through the controller 3; firstly, a first motor 21 rotates and drives a screw rod 24 to rotate, then a first sliding block 25 on the screw rod 24 makes linear motion along the screw rod 24, meanwhile, the first sliding block 25 drives a second motor 22 and a receiving groove 28 to move, when the receiving groove 28 moves to the garbage type containing cavity, the first motor 21 stops rotating, then the second motor 22 rotates and drives the receiving groove 28 to rotate around the central line of an output shaft of the second motor 22, then the receiving groove 28 gradually rotates, and after a whole circle of rotation, the second motor 22 stops rotating, and at the moment, garbage in the receiving groove 28 is poured into the cavity where the garbage is contained; the corrugated plate 231 at the upper end of the barrel body 1 moves along with the receiving groove 28, and the corrugated plate 231 covers the upper end of the barrel body 1, so that rainwater is prevented from dripping into the barrel body 1, and the phenomenon that garbage classification fails because the garbage is directly thrown into the barrel body 1 without being classified is prevented; the garbage in the receiving groove 28 is poured into the corresponding cavity through the rotation of the second motor 22, so that the garbage classification is completed, the service life is long, the power consumption is low, the structure is simple and easy to realize, the classification is accurate while the cost is ensured, and the classification efficiency is high.

The utility model provides a four categorised garbage bin are known for rubbish intelligent recognition automatic classification which characterized in that: the convolutional neural network single garbage identification method is suitable for the garbage can in claims 1-3, and comprises the following steps:

and in the background difference method, the two garbage images are respectively the image shot by the camera after the garbage is dumped for the previous time and the image shot after the garbage is received, so that the garbage image with the removed background is obtained.

The convolutional neural network includes an initialization layer, an intermediate layer, and a classification layer.

The initial layer consists of a convolutional layer with a core size of 7 × 7 and a step size of 2 and a pooling layer with a core size of 3 × 3 and a step size of 2. The middle layer consists of N similar residual blocks (N is more than or equal to 16 and less than or equal to 33); the classification layer is composed of an average pooling layer and a full connection layer, the activation function of the full connection layer is softmax, the probability vector of the finally output garbage identification is Q, and the dimensionality of the vector Q is the number of garbage items in the garbage database.

The structure of the similar residual block is shown in FIG. 4, when the input is a characteristic diagram F of H × W × C_iFirst, convolution operation was performed using the convolutional layer 1 × 1 and the convolutional layer 3 × 3 to obtain a characteristic diagram F of H × W × (C × R)₁The feature maps divided into R H × W × C are then input into the R branches, respectively, and aggregated into feature map F of H × W × C by the pixel addition operation₂Then, compressing the spatial information of the feature map by using a global pooling operation to obtain 1 × 1 × C feature map F₃And then passes through a convolution layer of 1 × 1 (containing convolution, ReLU activation function and batch)Quantity normalization processing) to obtain a characteristic diagram F of 1 × 1 × (C × R)₄Then inputting the weighted values into the softmax activation layer for attention weight value distribution, weighting the generated R groups of weight values and the corresponding H × W × C feature map on channel dimension to obtain R feature maps with attention weight, and adding and fusing the R weighted feature maps together to obtain a channel weighted feature map F_wThen, a feature diagram F of H × W × C is obtained through a convolution layer of 1 × 1₅Finally, obtaining a characteristic diagram F in the previous step by utilizing a residual error connection mode₅And input the feature map F_iAdding the obtained data to obtain the final output characteristic diagram F of the residual error-like block_o。

And the classification layer queries a 'garbage' database according to the identified garbage category to obtain the classification of the garbage.

In summary, the embodiments of the present invention have the following beneficial effects:

1. through the cooperation between a motor and No. two motors, a motor removes the receiving tank and rotates in the corresponding cavity that belongs to rubbish and hold with the rubbish of receiving tank to accomplish waste classification, long service life, the consumption is little, and simple structure realizes easily, and categorised accuracy and classification efficiency are high when guaranteeing the cost.

2. The buckled plate through port on the staving removes along with the receiving tank, and the buckled plate covers port on to the staving, avoids the rainwater drippage in the staving to and prevent that rubbish from directly throwing into the staving after not classifying, cause the emergence of rubbish classification failure phenomenon.

3. Through the cooperation between the convolutional neural network initialization layer, the middle feature extraction layer and the classification layer, the garbage classification is accurate and the classification efficiency is high, the user operation is simple and convenient, and the trouble of the user to the garbage classification is solved.

The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be considered to be within the technical scope of the present invention, and the technical solutions and the inventive concepts thereof according to the present invention should be equivalent or changed within the scope of the present invention.

Claims

1. The utility model provides a four categorised garbage bins of branch for rubbish intelligent recognition automatic classification, includes staving and controller, its characterized in that: a partition plate is arranged in the barrel body; the number of the partition plates is three, the partition plates divide the interior of the barrel body into four cavities in the length direction, and a driving structure is arranged at an upper port of the barrel body; the driving structure comprises a first motor, a second motor, a fixing plate, a screw rod, a first sliding block, a second sliding block, a sliding rod, a receiving groove and a touch switch; the number of the touch switches is two, and the two touch switches are symmetrically and fixedly connected to the side wall of the barrel body; the number of the fixed plates is two, the fixed plates are symmetrically and fixedly connected to the outer side wall of the barrel body, one end of one fixed plate is fixedly connected with a first motor, an output shaft of the first motor is fixedly connected with one end of a screw rod, the other end of the screw rod is hinged to one end of the other fixed plate, and a first sliding block is sleeved on the outer ring of the screw rod; the first sliding block is connected to the screw rod in a screw rod nut mode, and a second motor is fixedly connected to the first sliding block; the output shaft of the second motor is fixedly connected with a receiving groove; the receiving groove is semi-cylindrical, one side wall of the receiving groove is fixedly connected to an output shaft of a second motor, the other side wall of the receiving groove is provided with a first rod, and the first rod is hinged to a second sliding block; the second sliding block is connected to the sliding rod in a sliding mode, and two ends of the sliding rod are fixedly connected to the other ends of the two fixing plates.

2. The four-classification garbage can for intelligent garbage recognition and automatic classification as claimed in claim 1, wherein: elastic corrugated plates are arranged on the two fixing plates; one end of the corrugated plate is fixedly connected to the fixed plate, two sides of the other end of the corrugated plate are provided with a plate, and the middle position of the plate is hinged to the output shaft of the second motor and the outer ring of the first rod.

3. The four-classification garbage can for intelligent garbage recognition and automatic classification as claimed in claim 1, wherein: an identification device for identifying the garbage type is arranged in the receiving tank; the identification device is electrically connected with the touch switch and the controller, and the identification device identifies the garbage by using a convolutional neural network single garbage identification method.

4. The four-classification trash can identification for intelligent trash identification and automatic classification according to claim 3, wherein: the convolutional neural network single garbage identification method is suitable for the garbage can in claims 1-3, and comprises the following steps:

5. the four-classification garbage can for intelligent garbage recognition and automatic classification as claimed in claim 4, wherein: and in the background difference method, the two garbage images are respectively the image shot by the camera after the garbage is dumped for the previous time and the image shot after the garbage is received, so that the garbage image with the removed background is obtained.

6. The four-classification garbage can for intelligent garbage recognition and automatic classification as claimed in claim 4, wherein: the convolutional neural network includes an initialization layer, an intermediate layer, and a classification layer.

7. The four-classification garbage can for intelligent garbage recognition and automatic classification as claimed in claim 6, wherein: the initial layer consists of a convolutional layer with a core size of 7 × 7 and a step size of 2 and a pooling layer with a core size of 3 × 3 and a step size of 2. The middle layer consists of N similar residual blocks (N is more than or equal to 16 and less than or equal to 33); the classification layer is composed of an average pooling layer and a full connection layer, the activation function of the full connection layer is softmax, the probability vector of the finally output garbage identification is Q, and the dimensionality of the vector Q is the number of garbage items in the garbage database.

8. The four-classification garbage can for intelligent garbage recognition and automatic classification as claimed in claim 6, wherein: and the classification layer queries a 'garbage' database according to the identified garbage category to obtain the classification of the garbage.