Visual algorithm detection method for recycling garbage bags based on knife breaking and bag breaking and biaxial winding

Abstract

The present invention relates to the field of garbage recycling technology, and discloses a visual algorithm detection method for recycling garbage bags based on breaking knives and bags and double-axis winding, comprising the following steps: when the machine starts working, the breaking knife located above the domestic garbage bag performs the bag breaking operation driven by the pneumatic push rod; after the breaking knife is completed, the visual camera located above the bag breaking area transmits the image of the garbage bag in the working area to the host computer, and the host computer uses the image processing OpenCV library to detect whether the transmitted image is blurred; then it is judged whether the plastic bag is broken successfully. Compared with the prior art, the present invention not only makes the garbage bag breaking and recycling process fully automated and intelligent, but also the safety feedback mechanism ensures the safety of the operator. The use of the pneumatic push rod breaking knife and the double reels realizes the recycling of garbage bag winding while achieving accurate bag breaking; the real-time detection of the visual camera ensures the success of bag breaking and bag collection, and avoids the situation where the garbage bag has not been completely recycled or the garbage has not been completely released.

Description

Technical Field

[0001] The present invention relates to the technical field of garbage recycling, and in particular to a visual algorithm detection method for recycling garbage bags based on knife breaking and bag breaking and double-axis winding. Background Art

[0002] Garbage classification is conducive to the recycling and treatment of garbage, and can turn garbage into treasure. In order to improve the recycling rate of garbage, the garbage classification policy is gradually being implemented across the country. In this context, when people do garbage classification, they usually use garbage bags to put the classified garbage into the garbage recycling bin. At this time, the discarded garbage and plastic garbage bags are not actually classified. Today's garbage classification is basically done by manually unpacking bags and putting in garbage. This increases labor costs and classification difficulties. Therefore, a bag-breaking recycling equipment is needed in the garbage classification and recycling process, which can separate garbage and plastic garbage bags well for classification and processing. Most of the detection schemes for this process rely on manual experience and have a low degree of intelligence. They often cannot correctly judge whether the garbage bags are broken and whether the garbage bags have been recycled;

[0003] Most of the existing bag-breaking recycling devices rely solely on sharp props to simply break the garbage bags. This method not only requires manual cooperation, but also has certain risks. In smart trash cans, the mixing of garbage bags will interfere with the identification and classification of domestic waste, but the detection scheme of existing equipment is not smart and cannot handle garbage bags mixed with garbage. As for whether the garbage bag is successfully broken, the existing technology generally relies on simple image processing operations, with low recognition accuracy and is greatly affected by light;

[0004] When the tool rotates and approaches the bagged garbage, the cutting edge contacts the surface of the bagged garbage. Due to the appropriate rotation speed, the milling blade cannot damage the surface of the garbage bag, but only wraps the garbage bag upward along the chip groove, thereby completing the task of garbage bag recycling. Regarding the phenomenon of rotational winding, its characteristics are: it has the effect of converging surrounding objects. The cyclone will compress the surrounding air toward the center of the cyclone. For example, an industrial drill bit uses the extrusion principle to transport the surrounding blanks to the chip groove at the center of the drill bit. However, the above example is only the implementation of this principle for rigid or discontinuous bodies. Applying this principle to bag breaking technology can not only break the bagged garbage, but also collect the broken garbage bags.

[0005] The present invention proposes a set of intelligent garbage bag breaking and recycling solutions. Only by putting the garbage bags into the working area, the bag breaking and recycling work can be completed. Based on the multi-level feedback control method, the solution is safe and reliable. The high-definition camera located at the top of the working area is equipped with a neural network algorithm and an image processing algorithm to realize bag breaking detection and garbage bag recycling detection. Summary of the invention

[0006] The purpose of the present invention is to provide a visual algorithm detection method for recycling garbage bags based on knife breaking and bag breaking and double-axis winding to solve the problems in the background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A visual algorithm detection method for recycling garbage bags based on knife breaking and bag breaking and biaxial winding includes the following steps:

[0009] S1: When the machine starts working, the breaking knife located above the garbage bag performs the bag breaking operation driven by the pneumatic push rod;

[0010] S2: After the knife breaking is completed, the visual camera located above the bag breaking area transmits the image of the garbage bag in the working area to the host computer, and the host computer uses the image processing OpenCV library to detect whether the transmitted image is blurred;

[0011] S3: If the image is clear, perform filtering and noise reduction; if it is blurry, determine the blur type;

[0012] S4: If it is out-of-focus blur, restore the image by detecting the de-focus blur radius and then perform a filtering and denoising operation; if it is motion blur, restore the image by detecting the motion blur direction and scale and then perform a filtering and denoising operation;

[0013] S5: then judging whether the plastic bag is broken successfully;

[0014] S6: If the bag breaking is unsuccessful, the knife breaking will repeat S1-S5;

[0015] S7: If the bag is broken successfully, the tilting motor drives the bottom plate of the working area to tilt, and the garbage bag is brought close to the double reels, and the double reels perform the winding operation;

[0016] S8: While the dual reels are performing the winding operation, the visual camera located above the work area continuously transmits the work area image to the host computer for a series of identifications to determine whether the plastic bag is successfully wound and recycled. The specific process is as follows:

[0017] S8.1: The host computer uses the image processing OpenCV library to detect whether the transmitted image is blurred;

[0018] S8.2: If the image is clear, perform filtering and noise reduction; if it is blurry, determine the blur type;

[0019] S8.3: If the image is out-of-focus blur, the image is restored by detecting the out-of-focus blur radius and then a filtering and denoising operation is performed; if the image is motion blur, the image is restored by detecting the direction and scale of the motion blur and then a filtering and denoising operation is performed;

[0020] S8.4: Finally, determine whether the plastic bag is successfully rolled and recycled.

[0021] Preferably, in S5: a large number of images of successful bag breaking are trained by a convolutional neural network, and when the crack heat map of the image output by the neural network is greater than a set threshold, the bag is broken successfully.

[0022] Preferably, in S5: a large number of images of successful bag breaking are trained through a convolutional neural network. If the crack heat map of the image output by the neural network is less than a set threshold, the breaking knife repeats the bag breaking operation. The crack heat map is a garbage bag image trained by a convolutional neural network. Different thermal values ​​are assigned according to the size and depth of the opening of the garbage bag. The classifier can determine whether the plastic bag is successfully broken by the proportion of deep thermal values.

[0023] Preferably, in S8.3: the host computer calculates the proportion of garbage bag pixels in the working area by binarization, edge detection and other methods, and outputs a proportion value k.

[0024] Preferably, if k is less than 0.95, the double reels continue to perform the winding operation, and if k is greater than 0.95, the broken bag recovery work is completed.

[0025] Preferably, a camera located at a fixed position above the working area transmits images in real time, and a series of image preprocessing operations are performed to calculate the proportion of garbage bags in the working area, that is, the proportion value k.

[0026] The visual algorithm detection method based on knife breaking and bag breaking and biaxial winding for recycling garbage bags provided by the present invention has the following beneficial effects:

[0027] Compared with the prior art, the present invention not only makes the garbage bag breaking and recycling process fully automated and intelligent, but also ensures the safety of the operator through the safety feedback mechanism. The use of pneumatic push rod breaking knife and double reels can achieve accurate bag breaking and garbage bag winding and recycling at the same time. The real-time detection of the visual camera ensures the success of bag breaking and bag collection, avoiding the situation where the garbage bag has not been completely recycled or the garbage has not been completely released. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a schematic diagram of the overall process of an embodiment of the present invention. DETAILED DESCRIPTION

[0029] like Figure 1 As shown, the visual algorithm detection method for recycling garbage bags based on knife breaking and bag breaking and biaxial winding provided by the present invention includes the following steps:

[0030] S1: When the machine starts working, the breaking knife located above the garbage bag performs the bag breaking operation driven by the pneumatic push rod;

[0031] S2: After the knife breaking is completed, the visual camera located above the bag breaking area transmits the image of the garbage bag in the working area to the host computer, and the host computer uses the image processing OpenCV library to detect whether the transmitted image is blurred;

[0032] S3: If the image is clear, perform filtering and noise reduction; if it is blurry, determine the blur type;

[0033] S4: If it is out-of-focus blur, restore the image by detecting the de-focus blur radius and then perform a filtering and denoising operation; if it is motion blur, restore the image by detecting the motion blur direction and scale and then perform a filtering and denoising operation;

[0034] S5: Then determine whether the plastic bag is broken successfully: a large number of images of successful bag breaking are trained through the convolutional neural network. When the crack heat map of the image output by the neural network is greater than the set threshold, the bag is broken successfully; a large number of images of successful bag breaking are trained through the convolutional neural network. If the crack heat map of the image output by the neural network is less than the set threshold, the breaking knife repeats the bag breaking operation. The crack heat map is a garbage bag image trained by the convolutional neural network. Different heat values ​​are assigned according to the size and depth of the garbage bag opening. The classifier can determine whether the plastic bag is broken successfully by the proportion of deep heat values;

[0035] S6: If the bag breaking is unsuccessful, the knife breaking will repeat S1-S5;

[0036] S7: If the bag is broken successfully, the tilting motor drives the bottom plate of the working area to tilt, and the garbage bag is brought close to the double reels, and the double reels perform the winding operation;

[0037] S8: While the dual reels are performing the winding operation, the visual camera located above the work area continuously transmits the work area image to the host computer for a series of identifications to determine whether the plastic bag is successfully wound and recycled. The specific process is as follows:

[0038] S8.1: The host computer uses the image processing OpenCV library to detect whether the transmitted image is blurred;

[0039] S8.2: If the image is clear, perform filtering and noise reduction; if it is blurry, determine the blur type;

[0040] S8.3: If the image is out-of-focus blur, the image is restored by detecting the out-of-focus blur radius and then a filtering and denoising operation is performed; if the image is motion blur, the image is restored by detecting the direction and scale of the motion blur and then a filtering and denoising operation is performed;

[0041] S8.4: Finally, determine whether the plastic bag is successfully rolled and recycled: the upper computer calculates the proportion of garbage bag pixels in the working area through binarization, edge detection and other methods, and outputs the proportion value k; if k is less than 0.95, the dual reels continue to perform the winding operation; if k is greater than 0.95, the broken bag recycling work is completed; the camera located in a fixed position above the working area transmits the image in real time, and calculates the proportion of the garbage bag in the working area through a series of image preprocessing operations, that is, the proportion value k.

[0042] Safety feedback: To ensure the safe and stable operation of the machine, the high-definition camera located above the working area will continuously detect abnormal objects entering the working area, such as human hands, non-garbage organisms, etc., and use a lightweight target detection network for real-time detection. When there are abnormal objects, the lower computer controls the alarm to sound an alarm and emit a flashing red light. When there are no abnormal objects, operations such as breaking bags and recycling garbage bags are performed.

[0043] When used, the visual algorithm detection method based on knife breaking and bag breaking and double-axis winding recycling garbage bag includes the following steps: when the machine starts working, the knife breaking located above the domestic garbage bag performs the bag breaking operation driven by the pneumatic push rod; after the knife breaking is completed, the visual camera located above the bag breaking area transmits the image of the garbage bag in the working area to the host computer, and the host computer uses the image processing OpenCV library to detect whether the transmitted image is blurred; if the image is clear, perform filtering and noise reduction operations; if it is blurred, determine the type of blur; if it is out-of-focus blur, perform filtering and noise reduction operations after restoring the image by detecting the defocus blur radius; if it is motion blur, perform filtering and noise reduction operations after restoring the image by detecting the motion blur direction and scale; then determine whether the plastic bag is broken successfully; if the bag is broken unsuccessfully, repeat the knife breaking operation Line S1-S5; if the bag is broken successfully, then the tilting motor drives the bottom plate of the working area to tilt, and the garbage bag is brought close to the double reels, and the double reels perform the winding operation; while the double reels perform the winding operation, the visual camera located above the working area continuously transmits the working area image to the host computer for a series of identifications to determine whether the plastic bag is successfully wound and recycled. The specific process is as follows: the host computer uses the image processing OpenCV library to detect whether the transmitted image is blurred; if the image is clear, perform filtering and denoising operations; if blurred, determine the type of blur; if it is out-of-focus blur, perform filtering and denoising operations after restoring the image by detecting the defocus blur radius; if it is motion blur, perform filtering and denoising operations after restoring the image by detecting the motion blur direction and scale; finally, determine whether the plastic bag is successfully wound and recycled.

[0044] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0045] The embodiments described in the above embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

Claims

1. A visual algorithm detection method for recycling garbage bags based on knife breaking and bag breaking and biaxial winding, characterized by: The following steps are involved: S1: When the machine starts working, the breaking knife located above the garbage bag performs the bag breaking operation driven by the pneumatic push rod; S2: After the knife breaking is completed, the visual camera located above the bag breaking area transmits the image of the garbage bag in the working area to the host computer, and the host computer uses the image processing OpenCV library to detect whether the transmitted image is blurred; S3: If the image is clear, perform filtering and noise reduction; if it is blurry, determine the blur type; S4: If the image is out-of-focus blur, restore the image by detecting the out-of-focus blur radius and then perform a filtering and noise reduction operation; If it is motion blur, the motion blur direction and scale are detected, and the image is restored and then a filtering and noise reduction operation is performed; S5: Then determine whether the garbage bag is broken successfully, specifically including: A large number of successful bag breaking images are trained through convolutional neural networks. When the crack heat map of the neural network output image is greater than the set threshold, the bag is broken successfully; if the crack heat map of the neural network output image is less than the set threshold, the bag breaking operation is repeated by the breaking knife; S6: If bag breaking is unsuccessful, the breaking knife repeats steps S1-S5; S7: If the bag is broken successfully, the tilting motor drives the bottom plate of the working area to tilt, and the garbage bag is brought close to the double reels, and the double reels perform the winding operation; S8: While the dual reels are performing the winding operation, the visual camera located above the work area continuously transmits the work area image to the host computer for a series of identifications to determine whether the garbage bag is successfully wound and recycled. The specific process is as follows: S8.1: The host computer uses the image processing OpenCV library to detect whether the transmitted image is blurred; S8.2: If the image is clear, perform filtering and noise reduction operations; if it is blurry, determine the blur type; S8.3: If the image is out-of-focus blur, restore the image by detecting the out-of-focus blur radius and then perform a filtering and noise reduction operation; If it is motion blur, the motion blur direction and scale are detected, and the image is restored and then a filtering and noise reduction operation is performed; S8.4: Finally, determine whether the garbage bag is successfully rolled and recycled.

2. The visual algorithm detection method for recycling garbage bags based on knife breaking and bag breaking and biaxial winding according to claim 1 is characterized by: In step S8.3: the host computer calculates the proportion of garbage bag pixels in the working area through binarization and edge detection, and outputs the proportion value k.

3. The visual algorithm detection method for recycling garbage bags based on knife breaking and bag breaking and biaxial winding according to claim 2 is characterized by: If k is less than 0.95, the double reels continue to perform the winding operation, and if k is greater than 0.95, the broken bag recovery work is completed.

4. The visual algorithm detection method for recycling garbage bags based on knife breaking and bag breaking and biaxial winding according to claim 1 is characterized by: A camera located at a fixed position above the working area transmits images in real time, and a series of image preprocessing operations are performed to calculate the proportion of garbage bags in the working area, that is, the proportion value k.

Images