Garbage bag identification device and method

Abstract

The garbage bag identification device is characterized in that two cameras are arranged on the left side and the right side of the surface of the device side by side, the first camera is an RGB camera, and the surface of the second camera is covered with a color spectrum optical filter; the calculation module is used for identifying and judging the garbage bags according to the video data obtained by the first camera and the second camera, and when the pixel positions of the garbage bags identified by the first camera and the second camera are the same, the garbage bags are judged to be compliant garbage bags; and when the pixel positions of the garbage bags identified by the first camera and the second camera respectively are different, or only one camera detects the garbage bags, the garbage bags are judged to be non-compliant garbage bags, and an alarm signal is output. The device and method can improve the recognition accuracy and consistency of the garbage bag in different lighting environments, and can be applied to various environments.

Description

Technical Field

[0001] The invention belongs to the technical field of garbage disposal, and in particular relates to a garbage bag identification device and method. Background Art

[0002] The total amount of garbage in cities is now increasing year by year. In order to facilitate garbage recycling and treatment, garbage classification is now advocated, requiring people to throw different garbage into different trash cans. Currently, many places adopt garbage classification by placing trash cans marked with garbage classification signs in the community, and the degree of intelligence is relatively low.

[0003] Trash can now be sorted by the color of the bags. However, there is no good automatic monitoring solution for dumping trash bags. When handling the collected trash bags, garbage stations are often unable to effectively distinguish between compliant and non-compliant trash, resulting in increased pollution and processing costs. Some garbage stations have introduced recognition systems to identify compliant trash bags. However, changes in environmental and lighting conditions hinder accurate recognition, making it difficult for existing technologies to consistently identify non-compliant trash bags. Summary of the invention

[0004] In response to the current problems, the present invention proposes a garbage bag identification device and method, which improves the recognition accuracy and consistency in different environments by setting up dual cameras, and applies narrow-band filtering and visual algorithms to solve the problem of reduced recognition rate caused by color deviation under different lighting conditions (daylight, white light, fluorescent light, LED white light, LED warm light), so as to effectively distinguish between compliant and non-compliant garbage bags.

[0005] The specific plan is as follows: A garbage bag identification device includes a housing, an antenna, a power interface, an Internet interface, and a built-in computing module and a wireless module. Two cameras are arranged side by side on the surface of the device, wherein the first camera is an RGB camera for acquiring video data under ordinary light, and the surface of the second camera is covered with a color spectrum filter for acquiring video data under light of a specific wavelength band; The computing module is used to identify and judge the garbage bag based on the video data obtained by the first camera and the second camera: pixel alignment is performed on the video data obtained by the first camera and the second camera, and then object detection, analysis and identification are performed on the video data according to the visual image algorithm. When the pixel positions of the garbage bag respectively identified by the first camera and the second camera are the same, the garbage bag is judged to be a compliant garbage bag. When the pixel positions of the garbage bag respectively identified by the first camera and the second camera are different, or only one camera detects the garbage bag, it is judged to be a non-compliant garbage bag, and the device outputs an alarm signal.

[0006] Furthermore, an image is sent through light, the Internet or wireless communication and / or an alarm signal is externally output 232 , and the sent image displays a garbage bag detection frame and / or a blurred upper body image of a person.

[0007] Furthermore, the person image is identified and blurred based on the video data acquired by the first camera.

[0008] Furthermore, pixel alignment of the video data acquired by the first camera and the second camera refers to selecting a group of corresponding pixels in the images captured by the two cameras, punctuating the pixels, and inputting them into the center of the picture for processing to automatically align the two images to ensure that the pixels of the two images always correspond.

[0009] Further, the Internet method includes email, and the Internet interface can be an RJ45 interface.

[0010] Furthermore, the wireless communication method includes 4G, 5G or WIFI.

[0011] Furthermore, the power supply is a 12V DC power supply.

[0012] Furthermore, the spectral filter can only pass light with a wavelength of 400nm-450nm.

[0013] A method for identifying garbage bags based on the above-mentioned garbage bag identification device comprises the following steps: S1: placing the garbage bag identification device near the garbage bin, and making the first camera and the second camera face the garbage bag identification area; S2 turns on the first camera and the second camera at the same time to capture video data; S3 synchronizes the pixels of the video data obtained by the first camera and the second camera, detects whether there are non-background objects in the video data through a visual image algorithm, performs feature analysis on the non-background objects, identifies whether they are garbage bags, and marks their edges in the form of detection frames. When the first camera and the second camera respectively identify a garbage bag at the same pixel position, the garbage bag is determined to be compliant and no alarm signal is output. When the pixel positions of the garbage bags identified by the first camera and the second camera are different, or only one camera detects a garbage bag, the identified garbage bag is determined to be non-compliant and an alarm signal is output.

[0014] The garbage bag identification device and method provided by the present invention improves the accuracy and consistency of identification in different environments by setting up dual cameras, applies narrowband filtering and visual algorithms to process data in real time, and promptly alarms and performs processing according to the judgment results. The device and method can be used in a variety of environments, including urban areas, parks, and indoor facilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0016] Figure 1 It is a schematic diagram of a garbage bag identification device of the present invention; Figure 2 It is a schematic diagram of the dual camera imaging principle; Figure 3 This is the schematic diagram of the dual-camera garbage bag recognition algorithm; Figure 4 This is the garbage bag identification method of the present invention. DETAILED DESCRIPTION

[0017] In the following description, a large number of specific details are provided to provide a more thorough understanding of the present invention. However, it is apparent to those skilled in the art that the present invention can be implemented without one or more of these details. In other examples, in order to avoid confusion with the present invention, some technical features well known in the art are not described.

[0018] In order to further thoroughly understand the contents of the present invention, detailed steps will be proposed in the following description to illustrate the method proposed by the present invention. Obviously, the implementation of the present invention is not limited to the special details familiar to those skilled in the art. The preferred embodiments of the present invention are described in detail as follows, but it should be understood that the present invention can be implemented in different forms and should not be interpreted as being limited to the embodiments proposed here. When the terms "comprise" and / or "include" are used in this specification, it indicates that the features, integral bodies, steps, operations, elements and / or components are present, but does not exclude the presence or addition of one or more other features, integral bodies, steps, operations, elements, components and / or combinations thereof.

[0019] like Figure 1 As shown, a garbage bag identification device includes a housing, an antenna, a power interface, an Internet interface, and a built-in computing module and a wireless module. Two side-by-side cameras are arranged on the surface of the device, wherein the first camera is an RGB camera for acquiring video data under ordinary light, and the surface of the second camera is covered with a color spectrum filter for acquiring video data under light of a specific wavelength band; The computing module is used to identify and judge the garbage bag based on the video data obtained by the first camera and the second camera: pixel alignment is performed on the video data obtained by the first camera and the second camera, and then object detection, analysis and identification are performed on the video data according to the visual image algorithm. When the pixel positions of the garbage bag respectively identified by the first camera and the second camera are the same, the garbage bag is judged to be a compliant garbage bag. When the pixel positions of the garbage bag respectively identified by the first camera and the second camera are different, or only one camera detects the garbage bag, it is judged to be a non-compliant garbage bag, and the device outputs an alarm signal.

[0020] In one embodiment, the spectral filter can only pass light with a wavelength of 400nm-450nm.

[0021] In one embodiment, the device adopts a dual-lens single-core processor design, where the first lens is for RGB imaging; the second lens is used to filter the special band light source for imaging, and the parameters of the two lenses are as follows: Photosensitive element: 1 / 2.8" CMOS sensor; Imaging resolution: 2 million pixels (1920*1080); Optical lens: 6mm lens / 8mm lens.

[0022] In one embodiment, the device uses a dual-core integrated high-performance NNIE engine with 1Tops computing power, a HiSilicon 3516av300 processor, dual-core, ARM Cortex-A7@900MHz, 32KB ICache, 32KB D-Cache, 256KB L2Cache, supports NEON acceleration, and integrates an FPU processing unit.

[0023] like Figure 2 As shown, the ordinary RGB camera and the camera with special bands work at the same time, one way to obtain traditional video images, and the other way to achieve special band light imaging, forming two different videos for post-processing and analysis.

[0024] Since the two cameras are arranged left and right, the device may not be absolutely level when it is installed. Therefore, it is necessary to synchronize the images of the left and right cameras of the device. After installation, select a group of corresponding pixels in the left and right cameras respectively, use a group of corresponding pixels for punctuation, and input them into the center XY of the picture for processing to automatically align the two images. The two images will keep pixel correspondence in the later stage.

[0025] After collecting video data, two sets of videos corresponding to two pixels can be obtained through dual-lens calibration. The computing module of the device processes these two sets of videos. The RGB camera is responsible for detecting whether there is a garbage bag in the picture, and the special band camera detects whether there is an object that meets the special light band for preliminary detection. When the same pixel position in the two cameras meets the conditions, the detected object is a garbage bag that meets the requirements. If it is inconsistent, it is a garbage bag that does not meet the regulations.

[0026] In one embodiment, when RGB detects a garbage bag, it compares it with the garbage bag pixel position of the special band camera. If the comparison is consistent, they are all waste bags that meet the requirements. At this time, the device does not output any results to the outside world. When RGB detects a garbage bag, it compares it with the garbage bag pixel position of the special band camera. If the comparison is unsuccessful, the garbage bag that does not meet the regulations is marked. The RGB camera will perform upper body bokeh processing on the portrait and output the result map of the garbage bag position.

[0027] like Figure 3 As shown, the overall workflow of the algorithm is briefly summarized as follows: Video acquisition and preprocessing: Get YUV data and process the data to meet algorithm requirements.

[0028] Object detection: Detect whether non-background objects appear in the video.

[0029] Object tracking: Determine whether a detected object is a new object based on its position.

[0030] Target analysis: Analyze the features of a new target, confirm whether it is a garbage bag, and mark its edges.

[0031] Dual algorithm result matching: compares new targets in the two cameras, if they are in the same position, it means compliance, if they are not in the same position, it means non-compliance.

[0032] The intelligent analysis capabilities of the garbage bag identification device include: RGB garbage bag intelligent detection function: can identify garbage bags appearing in pictures; New garbage bag detection: Based on the background learning principle, determine whether the identified garbage bag is a new garbage bag and the pixel area of ​​the new garbage bag on the screen; RGB person detection function: detect and analyze people appearing in the video and frame the area of ​​the people; RGB People Blur: Blurs the upper body of detected people in the video.

[0033] Special band camera garbage bag detection function: including identifying qualified garbage bags through special band light imaging, and determining whether the identified garbage bag is a new garbage bag based on the background learning principle and the pixel area of ​​the new garbage bag in the image; Dual-screen garbage bag judgment: If there is a new garbage bag in the RGB image and there is also a garbage bag in the special band camera, if the pixels of the new garbage bag in the two images are the same, it can be judged that the new garbage bag in the two images is a compliant garbage bag. On the other hand, the new garbage bag is a non-compliant garbage bag; Result output: When a new garbage bag is judged to be an illegal garbage bag, if there is a person in the picture at this time, the upper body of the person will be blurred, and an image with a garbage bag detection frame and a blurred upper body of the person will be output.

[0034] Camera software features include: Configuration function: configure the camera's IP, user name, port number, primary and secondary data streams, volume and other parameters; Picture adjustment function: you can configure the output parameters of the camera imaging parameters; Communication function: The camera supports wired network, wifi, 4G and other communication modes. Each mode can be used alone. When combined, only one of Wi-Fi and 4G can be selected. Video viewing: You can preview pictures in real time through the built-in IE of the camera; Email configuration: You can set the address of the push service, the content of the email sent, etc.

[0035] Linkage strategy: Alarm linkage can be set to trigger email, light and 232 external output alarms at close range.

[0036] In one embodiment, the alarm signal may be output externally via lighting, sending an image, and / or 232 , wherein the sent image displays a garbage bag detection frame and / or a blurred upper body image of a person.

[0037] In one embodiment, the image of a person may be recognized and blurred based on the video data acquired by the first camera.

[0038] In one embodiment, the Internet method includes email, and the Internet interface can be an RJ45 interface.

[0039] In one embodiment, the power source is a 12V DC power source.

[0040] In one embodiment, the spectral filter can only pass light with a wavelength of 400nm-450nm.

[0041] like Figure 4 As shown, a recognition method based on the above-mentioned garbage bag recognition device comprises the following steps: S1: placing the garbage bag identification device near a garbage bin, and making the first camera and the second camera face the garbage bag identification area, such as the garbage bin; S2 turns on the first camera and the second camera at the same time to capture video data; S3 synchronizes the pixels of the video data obtained by the first camera and the second camera, detects whether there are non-background objects in the video data through a visual image algorithm, performs feature analysis on the non-background objects, identifies whether they are garbage bags, and marks their edges in the form of detection frames. When the first camera and the second camera respectively identify a garbage bag at the same pixel position, the garbage bag is determined to be compliant and no alarm signal is output. When the pixel positions of the garbage bags identified by the first camera and the second camera are different, or only one camera detects a garbage bag, the identified garbage bag is determined to be non-compliant and an alarm signal is output.

[0042] The recognition accuracy of the recognition algorithm of the present invention is: when the light is bright and the fluctuation value does not change much, the recognition accuracy is over 85%.

[0043] Algorithm processing capability: 1920*1080 detection, 10 frames per second*2 channels; Detection speed: 50ms.

[0044] The garbage bag identification device and method provided by the present invention improves the accuracy and consistency of identification in different environments by setting up dual cameras, applies narrowband filtering and visual algorithms to process data in real time, and promptly alarms and performs processing according to the judgment results. The device and method can be used in a variety of environments, including urban areas, parks, and indoor facilities.

[0045] The above-described embodiments are only preferred embodiments for fully illustrating the present invention, and the protection scope of the present invention is not limited thereto. Equivalent substitutions or changes made by those skilled in the art based on the present invention are within the protection scope of the present invention. The protection scope of the present invention shall be subject to the claims.

Claims

1. A garbage bag identification device, comprising a housing, an antenna, a power interface, an Internet interface, and a built-in computing module and a wireless module. Two cameras are arranged side by side on the surface of the device, wherein: The first camera is an RGB camera, which is used to obtain video data under ordinary light, and the surface of the second camera is covered with a color spectrum filter, which is used to obtain video data under light of a specific wavelength band; The computing module is used to identify and judge the garbage bag based on the video data obtained by the first camera and the second camera: pixel alignment is performed on the video data obtained by the first camera and the second camera, and then object detection, analysis and identification are performed on the video data according to the visual image algorithm. When the pixel positions of the garbage bag respectively identified by the first camera and the second camera are the same, the garbage bag is judged to be a compliant garbage bag. When the pixel positions of the garbage bag respectively identified by the first camera and the second camera are different, or only one camera detects the garbage bag, it is judged to be a non-compliant garbage bag, and the device outputs an alarm signal.

2. A garbage bag identification device according to claim 1, characterized in that: An image is sent through light, the Internet or wireless communication and / or an external alarm signal is output 232 , wherein the sent image displays a garbage bag detection frame and / or a blurred upper body image of a person.

3. The garbage bag identification device according to claim 2, further characterized in that: The person image is recognized and blurred based on the video data acquired by the first camera.

4. The garbage bag identification device according to claim 1, further characterized in that: The pixel alignment of the video data obtained by the first camera and the second camera refers to selecting a group of corresponding pixels in the images taken by the two cameras, punctuating the pixels, and inputting them into the center of the picture for processing to automatically align the two images to ensure that the pixels of the two images always correspond.

5. The garbage bag identification device according to claim 2, further characterized in that: The Internet method includes email, and the Internet interface can be an RJ45 interface.

6. The garbage bag identification device according to claim 2, further characterized in that: The wireless communication method includes 4G, 5G or WIFI.

7. The garbage bag identification device according to claim 1, further characterized in that: The power supply is a 12V DC power supply.

8. The garbage bag identification device according to claim 1, further characterized in that: The spectral filter can only pass light with a wavelength of 400nm-450nm.

9. A method for identifying a garbage bag based on the garbage bag identification device according to any one of claims 1 to 8, characterized in that: The steps include: S1: placing the garbage bag identification device near the garbage bin, and making the first camera and the second camera face the garbage bag identification area; S2 turns on the first camera and the second camera at the same time to capture video data; S3 synchronizes the pixels of the video data acquired by the first camera and the second camera, detects whether there are non-background objects in the video data through a visual image algorithm, performs feature analysis on the non-background objects, identifies whether they are garbage bags, and marks their edges in the form of detection frames. When the first camera and the second camera respectively identify a garbage bag at the same pixel position, the garbage bag is determined to be compliant and no alarm signal is output. When the pixel positions of the garbage bags identified by the first camera and the second camera are different, or only one camera detects a garbage bag, the identified garbage bag is determined to be non-compliant, an alarm signal is output and / or an image is sent, and the sent image displays a garbage bag detection frame and / or a blurred image of the upper body of a person.

Images