CN115601594A - Visual data processing method and system based on intelligent recovery - Google Patents
Visual data processing method and system based on intelligent recovery Download PDFInfo
- Publication number
- CN115601594A CN115601594A CN202211275342.5A CN202211275342A CN115601594A CN 115601594 A CN115601594 A CN 115601594A CN 202211275342 A CN202211275342 A CN 202211275342A CN 115601594 A CN115601594 A CN 115601594A
- Authority
- CN
- China
- Prior art keywords
- visual data
- data processing
- algorithm
- original image
- recycled
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02W—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO WASTEWATER TREATMENT OR WASTE MANAGEMENT
- Y02W30/00—Technologies for solid waste management
- Y02W30/10—Waste collection, transportation, transfer or storage, e.g. segregated refuse collecting, electric or hybrid propulsion
Landscapes
- Processing Of Solid Wastes (AREA)
Abstract
The invention discloses a visual data processing method based on intelligent recovery, which is embedded into a visual data processing device based on intelligent recovery, and the device comprises: the garbage recycling bin comprises a control module arranged on a bin cover body of the garbage recycling bin, a camera communicated with the control module and used for acquiring newly thrown-in recycled materials in the garbage recycling bin, and a recycled material delivery opening which is arranged on the bin cover body of the garbage recycling bin and is controlled to be opened, wherein the recycled material delivery opening is connected with the control module; the method comprises the following steps: acquiring an image, detecting a recovered object and judging a newly-entered recovered object; the invention provides a visual data processing method and system based on intelligent recovery, which are low in cost and high in accuracy.
Description
Technical Field
The invention relates to the field of intelligent garbage recycling, in particular to a visual data processing method and system based on intelligent recycling.
Background
In the current garbage recycling application, garbage recycling boxes are arranged in areas such as districts and office buildings at fixed points, residents can put garbage into the garbage recycling boxes by themselves, and then the garbage can is cleaned by garbage cleaning personnel at regular intervals. To facilitate use, the waste bin is typically placed in a number of different positions. But the classification that each garbage collection case was retrieved is different, often appears at present, and the resident often makes mistakes corresponding garbage collection bucket when rubbish is put in.
Most classification work of the existing treatment mode depends on manpower, so that automatic garbage classification and recovery are not really realized. At present, high-tech intelligent garbage cans such as solar automatic compression garbage cans and electric compression garbage cans for plastic bottles, glass bottles and pop cans appear in developed countries, and the garbage cans are intelligently managed in a scheduling mode.
How to effectively realize the bottle recycling with low cost and high accuracy rate and improve the bottle classification and identification results becomes a problem to be solved.
Disclosure of Invention
The invention overcomes the defects of the prior art, and provides the intelligent recovery-based visual data processing method and system with a low-cost recovery device and high accuracy.
The technical scheme of the invention is as follows:
the visual data processing method based on intelligent recovery is embedded into a visual data processing device based on intelligent recovery, and the device comprises: the garbage recycling bin comprises a control module arranged on a bin cover body of the garbage recycling bin, a camera communicated with the control module and used for acquiring newly thrown-in recycled materials in the garbage recycling bin, and a recycled material delivery opening which is arranged on the bin cover body of the garbage recycling bin and is controlled to be opened, wherein the recycled material delivery opening is connected with the control module; the method comprises the following steps:
an image acquisition step: tong (Chinese character of 'tong')The camera is used for acquiring an image of the bearing condition of the current recycled object in the garbage recycling bin as a background image P A (ii) a And acquiring an image of the current load condition of the recycling object in the garbage recycling bin as an original image P B ；
Detecting a recovered substance: preliminarily detecting a background map P by a first target detection algorithm A Original image P B All detected recoveries in (a);
judging newly-entered recovery materials: processing the data acquired in the step of detecting the recovered substances by a second recognition detection algorithm, and positioning a background image P by the second recognition detection algorithm A With respect to the original image P B At the position of (1), screening out the original image P B And background picture P A The position of the detection object in which there is a difference;
synchronously carrying out third frame difference algorithm processing on the data acquired in the step of detecting the recycled objects to acquire the positions of the detected objects with differences; combining the second re-recognition detection algorithm and the third frame difference algorithm to obtain the new recycled material in the original image P B The position of (1);
and (3) classification step: and processing the newly input recycled objects confirmed in the newly input recycled object judging step through a fourth classification detection algorithm to obtain a classification result of the target detection object.
Further, after the sensor detects the recovered object, the camera is started after a control instruction is set for a certain time interval, and the original image P is obtained B 。
Further, screening out the original image P in the judgment step of newly-entered recycled material B And background picture P A There are the first three detector positions with the greatest difference.
Further, the second Re-recognition detection algorithm adopts a Re-Identification technology or a softmax superposed triple hard combined training mechanism algorithm.
Further, the fourth classification detection algorithm adopts a MobileNet algorithm, which includes an SE attention module and an Sa module.
Further, the third frame difference algorithm process is to process the background image P A Original image P B And performing frame difference processing to obtain a difference detection object larger than a threshold value.
Further, the first target detection algorithm uses the YOLOV5 algorithm and adds an attention mechanism in the backbone network structure thereof to optimize the capability of resisting the light interference.
Further, after the classification step, the current background picture P is updated A For the original image P B。
A visual data processing system based on intelligent recovery, comprising a visual data processing method of any one of the preceding claims 1 to 8 for implementing an embedded intelligent recovery based method; the system comprises a camera, a cloud server, a communication module and a sensor, wherein the cloud server is used for processing data acquired by the camera, the communication module is connected with the control module and communicated with the cloud server, and the sensor is used for triggering the control module to send a control instruction for detecting the input of a recycled material.
Further, the cloud server further comprises terminal equipment, and the terminal equipment acquires data and instructions on the cloud server through the communication module.
The invention has the advantages that:
according to the invention, the barrel cover of the existing garbage recycling box is improved, and the camera is arranged on one side of the barrel cover close to the garbage recycling box, so that the garbage loading condition in the garbage recycling box can be obtained, and low-cost detection can be directly realized. The image identification method with high accuracy provides key data for the subsequent scheduling of the garbage recycling bin.
The invention starts from the problem of detecting the recycled objects in practical application, and optimizes and applies a corresponding algorithm to more accurately position and classify the newly recycled objects, namely, after a user carries out garbage delivery operation, the types and the positions of the delivered objects of the user are quickly identified. Meanwhile, a delivery port for the recycled materials of bottles is designed on the barrel cover of the intelligent garbage recycling box, the delivery identification is particularly carried out on plastic bottles, glass bottles, pop cans and paper bottles, the types of the fed materials are standardized, a paid recycling mechanism is set, a control module can select a low-end chip, the barrel cover transformation cost is reduced, an identification task is completed by cloud service, and the garbage barrel cleaning management cost and the equipment hardware cost are reduced as far as possible.
Drawings
FIG. 1a is a flow chart of a method for processing visual data based on smart recycling according to the present invention;
FIG. 1b is a process for intelligent recovery-based visual data processing according to the present invention;
FIG. 2 is a logic diagram of an image processing algorithm of the vision data processing method based on intelligent recovery of the invention;
FIG. 3 is a schematic diagram of a system for processing visual data based on intelligent recovery according to the present invention;
FIG. 4 is a schematic view of a recycle delivery port of the present invention.
Detailed Description
The invention is further described with reference to the following figures and detailed description. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
As shown in fig. 1 to 4, a visual data processing method based on smart recycling embeds a visual data processing apparatus based on smart recycling, the apparatus including: the garbage recycling bin comprises a control module arranged on a bin cover body of the garbage recycling bin, a camera communicated with the control module and used for acquiring newly thrown-in recycled materials in the garbage recycling bin, and a recycled material delivery opening which is arranged on the bin cover body of the garbage recycling bin and is controlled to be opened, wherein the recycled material delivery opening is connected with the control module; the method specifically comprises the following steps:
s11: an image acquisition step: acquiring an image of the current bearing condition of the recycling object in the garbage recycling bin through a camera as a background image P A (ii) a And acquiring an image of the current load condition of the recycling object in the garbage recycling bin as an original image P B . That is, after the sensor detects the recovered object, the camera is started after the control command interval is set to obtain the original image P B . To monitor the garbage recoveryThe effect of the current situation in the bucket can effectively monitor whether new recovery objects enter or not so as to start subsequent recovery object detection.
S12: detecting a recovered substance: preliminarily detecting a background map P by a first target detection algorithm A Original image P B All recoveries detected.
Preferably, the first target detection algorithm is implemented by using a YOLOV5 algorithm and optimized, and referring to fig. 2, an attention mechanism is added to the backbone network structure, so that the target detection frame is positioned more accurately, the light interference resistance of the target detection frame is enhanced, and the target detection method is better suitable for the complex environment in the existing garbage recycling bin. It may also preferably be done using SSD, FASTRCNN algorithm.
S13: judging newly-entered recovery materials: processing the data acquired in the step of detecting the recycled materials by a second re-identification detection algorithm, and positioning a background picture P by the second re-identification detection algorithm A With respect to the original image P B At the position of (1), screening out the original image P B And background picture P A Where there is a difference in the position of the detector.
Wherein, the position of the detected object is screened out to newly enter the original image P in the step of judging the recovered object B And background picture P A The first three positions of the detection object with the largest difference are enough to confirm the position of the newly-entered recovery object.
Preferably, the second Re-recognition detection algorithm can use a Re-Identification technology, called ReID technology for short, or further use a softmax plus triplet hard combined training mechanism, so that the Re-recognition detection algorithm is suitable for the requirement of accurate Re-recognition in the current application scene.
Referring to fig. 1b, synchronously performing a third frame difference algorithm on the data obtained in the step of detecting the recycling object to obtain the positions of the detecting objects with differences; combining the second recognition detection algorithm and the third frame difference algorithm to obtain the original image P of the newly-entered and recycled object B Of (c) is used. Wherein the third frame difference algorithm process is to the background image P A Original image P B Frame difference processing is carried out to obtain a difference detection object larger than a threshold value, so that the position of a newly-entered recovery object can be effectively confirmed.
S14: and (3) classification step: and processing the newly input recycled objects confirmed in the newly input recycled object judging step through a fourth classification detection algorithm to obtain a classification result of the target detection object. The fourth classification detection algorithm adopts a MobileNet algorithm, and a design SE attention module and a Sa module are added to the fourth classification detection algorithm to improve the classification accuracy. It can also be preferably implemented using the ShuffleNet, efficientNet technology.
Preferably, the classification step may be followed by an update to update the current background map P A For the original image P B . So as to process the next new incoming object and repeat the steps S11-S14.
Referring to fig. 3, the present invention further discloses a vision data processing system based on intelligent recovery, including the above-mentioned vision data processing method based on intelligent recovery method; the system comprises a camera, a cloud server, a communication module and a sensor, wherein the cloud server is used for processing data acquired by the camera, the communication module is connected with the control module and communicated with the cloud server, and the sensor is used for triggering the control module to send a control instruction for detecting the input of a recycled material.
Preferably, the cloud server further comprises a terminal device, and the terminal device acquires data and instructions on the cloud server through the communication module.
The utility model discloses a retrieve thing delivery port, including delivery port body 1 and delivery port apron 2, delivery port apron 2 passes through hinge 3 with delivery port body 1 and is connected, and is concrete, and hinge 3 cooperation has the delivery port of lockhole, makes delivery port department in normally closed state, and the user is fixed in the only identifier on retrieving the delivery port through electronic equipment scanning, sends signal to server, assigns the order control by the server and retrieves delivery port and unblanks, and delivery port is automatic to be popped open. The mode is low in cost, convenient to maintain and replace and more beneficial to popularization and use.
Can also set up detection device at the recovery thing entrance of delivery port body 1, detection device detects the delivery article number, and detects not influenced by weather and temperature. This allows for more efficient detection of the number of delivered items. The detection device adopts an infrared detection device, which can not be influenced by weather and temperature, can realize accurate detection and has lower cost.
Can also set up the camera based on AIOT around the recovery thing entrance of delivery port body 1, acquire the information of recovery thing through the camera. The camera can be used for assisting the detection device to count the number of the recovered delivery items. The camera can also be used for acquiring the information of the delivered objects, namely when the detection device detects the delivered objects, the camera starts to shoot the delivered objects, so that the content information of the delivered objects is identified according to the images. The mechanism that the detection device can trigger the camera to operate at the moment can be used for better high-efficiency operation and saving electric energy.
The communication module generally adopts an AIOT technology to better guarantee the operation of the system. In order to standardize the separate putting of the objects by the residents, a paid recovery mechanism can be set to promote the use of the residents. This scheme can sweep the sign indicating number back at the user, uncaps automatically, and the user post the back, triggers the response in real time, real-time identification, and the camera requirement is not harsh, and need not the industrial camera of high definition, compresses the picture, and the picture size can be reduced to general high definition picture size 1/50 ~ 1/30, effectively improves user's use and experiences.
To sum up, this design is transplanted conveniently, and the movable design to the conventional garbage bin of equidimension not, all can the key formula upgrade, occupies smallly, and very easy the popularization, public area convenient to use and cost are lower. The delivery opening of the recycled material is reasonable in design, when the user uses the device, the recycled material is naturally thrown down from top to bottom, and points can be obtained in real time after the user throws the cover off, so that the device is convenient to use. A cascade model fusion technology is adopted in data processing, the robustness of recognition is improved, and the problems of inaccurate recognition and the like caused by large change of the posture after an object is smashed can be effectively solved.
The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, many modifications and amendments can be made without departing from the spirit of the present invention, and these modifications and amendments should also be considered as within the scope of the present invention.
Claims (10)
1. Visual data processing method based on intelligent recovery, characterized by that, imbed the visual data processing apparatus based on intelligent recovery, the apparatus includes: the garbage recycling bin comprises a control module arranged on a bin cover body of the garbage recycling bin, a camera communicated with the control module and used for acquiring newly thrown-in recycled materials in the garbage recycling bin, and a recycled material delivery opening which is arranged on the bin cover body of the garbage recycling bin and is controlled to be opened, wherein the recycled material delivery opening is connected with the control module; the method comprises the following steps:
an image acquisition step: obtaining an image of the current load condition of the recycling object in the garbage recycling bin through a camera as a background image P A (ii) a And acquiring an image of the current load condition of the recycling object in the garbage recycling bin as an original image P B ；
Detecting a recovered substance: preliminarily detecting a background map P by a first target detection algorithm A Original image P B All detected recoveries in (a);
judging newly-entered recovery materials: processing the data acquired in the step of detecting the recovered substances by a second recognition detection algorithm, and positioning a background image P by the second recognition detection algorithm A With respect to the original image P B At the position of (1), screening out the original image P B And background picture P A The position of the detection object in which there is a difference;
synchronously carrying out third frame difference algorithm processing on the data acquired in the step of detecting the recycled objects to acquire the positions of the detected objects with differences; combining the second recognition detection algorithm and the third frame difference algorithm to obtain the original image P of the newly-entered and recycled object B The position of (1);
and (3) classification step: and processing the newly input recycled objects confirmed in the newly input recycled object judging step through a fourth classification detection algorithm to obtain a classification result of the target detection object.
2. The visual data processing method based on smart recycling according to claim 1, wherein: after the sensor detects the recovered object, the camera is started after a control instruction interval of a certain time to acquire the original image P B 。
3. The visual data processing method based on smart recycling according to claim 1, wherein: screening out original image P in judgment step of newly-entered recycled object B And background picture P A There are the first three detector positions with the greatest difference.
4. The visual data processing method based on smart recycling according to claim 1, wherein: the second recognition detection algorithm adopts a Re-Identification technology or a softmax superposed triple hard combined training mechanism algorithm.
5. The visual data processing method based on smart recycling according to claim 1, wherein: the fourth classification detection algorithm adopts a MobileNet algorithm, which comprises an SE attention module and a Sa module.
6. The visual data processing method based on smart recycling according to claim 1, wherein: the third frame difference algorithm processing is to the background image P A Original image P B And performing frame difference processing to obtain a difference detection object larger than a threshold value.
7. The visual data processing method based on smart recycling according to claim 1, wherein: the first target detection algorithm uses the YOLOV5 algorithm and adds an attention mechanism in its backbone network structure to optimize its light interference resistance.
8. The visual data processing method based on smart recycling according to claim 1, wherein: after further including the classification step, updating the current background map P A Is an original image P B。
9. Visual data processing system based on intelligence is retrieved, its characterized in that: the method for processing visual data based on intelligent recovery method is realized by any one of the claims 1 to 8; the system comprises a camera, a cloud server, a communication module and a sensor, wherein the cloud server is used for processing data acquired by the camera, the communication module is connected with the control module and communicated with the cloud server, and the sensor is used for triggering the control module to send a control instruction for detecting the input of a recycled material.
10. The intelligent recovery-based visual data processing system of claim 9, wherein: the cloud server further comprises terminal equipment, and the terminal equipment acquires data and instructions on the cloud server through the communication module.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211275342.5A CN115601594A (en) | 2022-10-18 | 2022-10-18 | Visual data processing method and system based on intelligent recovery |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211275342.5A CN115601594A (en) | 2022-10-18 | 2022-10-18 | Visual data processing method and system based on intelligent recovery |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115601594A true CN115601594A (en) | 2023-01-13 |
Family
ID=84849901
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211275342.5A Pending CN115601594A (en) | 2022-10-18 | 2022-10-18 | Visual data processing method and system based on intelligent recovery |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115601594A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116246215A (en) * | 2023-05-11 | 2023-06-09 | 小手创新(杭州)科技有限公司 | Method for identifying new articles based on visual algorithm, barrel cover and intelligent recycling bin |
-
2022
- 2022-10-18 CN CN202211275342.5A patent/CN115601594A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116246215A (en) * | 2023-05-11 | 2023-06-09 | 小手创新(杭州)科技有限公司 | Method for identifying new articles based on visual algorithm, barrel cover and intelligent recycling bin |
| CN116246215B (en) * | 2023-05-11 | 2024-01-09 | 小手创新(杭州)科技有限公司 | Method for identifying new articles based on visual algorithm, barrel cover and intelligent recycling bin |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109684979B (en) | Image recognition technology-based garbage classification method and device and electronic equipment | |
| CN109201514B (en) | Waste sorting recycle method, garbage classification device and classified-refuse recovery system | |
| CN110697273B (en) | Household garbage intelligent recognition and automatic classification system and method based on iterative learning control | |
| CN208165794U (en) | A kind of intelligent classification dustbin | |
| CN113135368B (en) | Intelligent garbage front-end classification system and method | |
| CN206489651U (en) | " internet+" Packaging Bottle intelligence recovery system | |
| CN111559588A (en) | Intelligent garbage can for classified garbage throwing and classified garbage throwing method | |
| CN111178182A (en) | Real-time detection method for garbage loss behavior | |
| CN111611970A (en) | Urban management monitoring video-based disposable garbage behavior detection method | |
| CN206546593U (en) | House refuse intelligent classification reclaims cloud identifying system | |
| CN115601594A (en) | Visual data processing method and system based on intelligent recovery | |
| CN111056191A (en) | Intelligent garbage collection system capable of classifying and delivering multiple garbage | |
| CN211392481U (en) | Image recognition big data-based garbage classification system | |
| CN111738118A (en) | Garbage classification management method and system based on visual recognition | |
| CN107123076A (en) | A kind of waste management system and waste management method | |
| CN111217047A (en) | Artificial intelligent garbage classification device and method based on IC card management | |
| CN113657143B (en) | Garbage classification method based on classification and detection combined judgment | |
| CN110406861A (en) | A kind of dustbin automatically opening dustbin cover | |
| CN111767804A (en) | Recyclable garbage image classification method and system based on artificial intelligence | |
| CN212048939U (en) | Intelligent garbage recycling device for community | |
| CN117088015A (en) | Intelligent garbage disposal method, system and storage medium | |
| CN110775483A (en) | Community intelligent garbage recycling device and method, electronic equipment and storage medium | |
| CN114715562B (en) | Recognition method for kitchen waste illegal throwing behavior | |
| CN212515840U (en) | Garbage classification management system based on visual identification | |
| CN112508103A (en) | Perishable garbage image identification and assessment management method based on garbage collection and transportation vehicle |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |