Solid waste classification and identification method and system for recycling

Abstract

The invention relates to the technical field of image recognition, in particular to a solid waste classification and recognition method and system for recycling. The method comprises the following steps: acquiring a junk image, and analyzing the image to determine a junk area; determining a characteristic value according to the circular fitting feature and the gray value; determining a sudden change amplitude according to the characteristic value change; determining a to-be-selected sudden change time node of garbage change caused by the conveyor belt according to the value of the sudden change amplitude; target time nodes generating garbage state sudden change on the conveying belt are obtained through screening and cut, time periods are determined, and state change indexes are determined according to changes of the overall sudden change amplitude in each time period; and in combination with the characteristic value and the state change index, determining a possibility index that the garbage area is a zip-top can area, and screening the zip-top can area according to the possibility index of each garbage area. The ring-pull can classification and recognition stability can be improved, the ring-pull can classification and recognition effect is better, and the accuracy is higher.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition, and particularly relates to a method and system for classifying and recognizing solid waste for recycling. Background Art

[0002] Waste recycling is a key link in achieving sustainable development goals. The recycling process of aluminum saves more than 95% of the energy compared to the primary aluminum production process, so it has high recycling value, especially for aluminum cans. However, since aluminum products are not magnetic, aluminum cans cannot be screened out by a magnetic separator. In addition, the outer shell of aluminum cans is relatively soft and easily deformed, and it is very easy to be flattened on the conveyor belt, causing difficulties in identifying them during the screening process. Especially after the waste has been preliminarily screened by a magnetic separator, a wind separator, etc., the remaining waste not only contains aluminum cans but also other waste, which makes the secondary screening of flattened aluminum cans more complex and difficult.

[0003] In related technologies, manual screening is carried out, or specific image recognition and analysis are performed according to the characteristics of the overall image. However, due to the combination of flattened aluminum cans and a bumpy conveyor belt, the state of the aluminum cans themselves will change, resulting in unstable recognition of the characteristics of the aluminum cans, and thus the classification and recognition effect of the aluminum cans is poor and the accuracy is low. Summary of the Invention

[0004] In order to solve the technical problem in related technologies that due to the combination of flattened aluminum cans and a bumpy conveyor belt, the state of the aluminum cans themselves will change, resulting in unstable recognition of the characteristics of the aluminum cans, and thus the classification and recognition effect of the aluminum cans is poor and the accuracy is low, the present invention provides a method and system for classifying and recognizing solid waste for recycling, and the specific technical solutions adopted are as follows:

[0005] The present invention proposes a method for classifying and recognizing solid waste for recycling, and the method includes:

[0006] Obtain waste images at different sampling times on the conveyor belt, and perform image analysis to determine different waste areas; determine the characteristic values of the waste areas belonging to aluminum cans according to the matching characteristics and gray values of each waste area fitted with a circle;

[0007] Determine the mutation amplitude of the same waste area at each sampling time according to the change of the characteristic values of the same waste area at different positions on the conveyor belt at different sampling times; determine the candidate mutation time nodes for the conveyor belt to cause waste changes according to the numerical values of the mutation amplitudes;

[0008] Based on the candidate mutation time nodes corresponding to each garbage area, the target time nodes at which the garbage state mutates on the conveyor belt are screened. Taking the target time nodes as the cutting points, different time periods of each garbage area on the conveyor belt are determined. According to the change of the overall mutation amplitude in each time period, the state change index of the garbage area in each time period is determined;

[0009] Combining the characteristic value and the state change index, the possibility index that the garbage area is a beverage can area is determined, and the beverage can areas are screened according to the possibility index of each garbage area.

[0010] Further, the method for determining the characteristic value that the garbage area belongs to a beverage can according to the matching feature of fitting the garbage area with a circle and the gray value includes:

[0011] Perform elliptical fitting on the garbage area, and determine the distance between the two foci of the fitted ellipse as the circle matching distance;

[0012] Perform maximum-minimum normalization on the opposite number of the circle matching distance to obtain the circle matching index;

[0013] Calculate the average gray value of all pixel points after graying the garbage area to obtain the gray feature index;

[0014] Calculate the product of the circle matching index and the gray feature index, and perform maximum-minimum normalization to obtain the characteristic value.

[0015] Further, according to the change of the characteristic value of the same garbage area at different positions on the conveyor belt at different sampling times, the mutation amplitude of the same garbage area at each sampling time is determined, including:

[0016] Determine any sampling time as the target time, calculate the difference between the characteristic value of the target time and its previous time as the first difference; calculate the difference between the characteristic value of the target time and its next time as the second difference;

[0017] Take the normalized value of the sum of the first difference and the second difference as the mutation amplitude at the target time.

[0018] Further, according to the numerical value of the mutation amplitude, the candidate mutation time nodes at which the conveyor belt causes garbage changes are determined, including:

[0019] Take the time nodes at which the mutation amplitude is greater than the preset amplitude threshold as the candidate mutation time nodes.

[0020] Further, based on the candidate mutation time nodes corresponding to each garbage area, the target time nodes at which the garbage state mutates on the conveyor belt are screened, including:

[0021] Determine the frequencies corresponding to the sampling moments of all candidate mutation time nodes in the garbage areas, and use the preset number of candidate mutation time nodes with the largest frequencies as the target time nodes.

[0022] Furthermore, according to the change in the overall mutation amplitude in each time period, determine the state change index of the garbage area in each time period, including:

[0023] Determine the mutation stability within the time period according to the dispersion of all mutation amplitudes in the same time period;

[0024] Calculate the opposite of the mean of all mutation amplitudes in the same time period, and perform maximum-minimum normalization as the numerical feature index;

[0025] Use the product of the numerical feature index and the mutation stability as the state change index of the garbage area in the corresponding time period.

[0026] Furthermore, determine the mutation stability within the time period according to the dispersion of all mutation amplitudes in the same time period, including:

[0027] Calculate the opposite of the standard deviation of all mutation amplitudes in the same time period, and perform maximum-minimum normalization as the mutation stability.

[0028] Furthermore, combine the characteristic value and the state change index to determine the possibility that the garbage area is an aluminum can area, including:

[0029] Calculate the mean of the characteristic values of all sampling moments in the same time period as the segment characteristic coefficient;

[0030] Calculate the product of the segment characteristic coefficient and the state change index in the same time period as the influence index of the time period;

[0031] Use the mean of the influence indexes of all time periods as the possibility index that the corresponding garbage area is an aluminum can area.

[0032] Furthermore, screen the aluminum can areas according to the possibility index of each garbage area, including:

[0033] Use the garbage areas with the possibility index greater than the preset possibility threshold as the aluminum can areas.

[0034] On the other hand, the present invention also includes a solid waste classification and identification system for recycling. The system includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the steps of the method described in any one of the foregoing are implemented.

[0035] The present invention has the following beneficial effects:

[0036] The present invention obtains garbage images at different sampling moments on a conveyor belt, performs image analysis to determine different garbage areas; determines characteristic values indicating that the garbage areas belong to aluminum cans according to the matching characteristics and gray values of each garbage area fitted with a circle; the characteristic values characterize the aluminum can characteristics of the garbage areas themselves, so subsequent numerical analysis and change analysis can be carried out based on the characteristic values. Then, according to the change of the characteristic values of the same garbage area at different positions on the conveyor belt at different sampling moments, the mutation amplitude of the same garbage area at each sampling moment is determined; the candidate mutation time nodes of the garbage change caused by the conveyor belt are determined according to the numerical values of the mutation amplitude; according to the candidate mutation time nodes corresponding to each garbage area, the target time nodes of the garbage state mutation on the conveyor belt are screened, and taking the target time nodes as the cutting points, different time periods of each garbage area on the conveyor belt are determined. The division of the time periods can effectively avoid the "state reset" influence caused by the bumpy rollers, improve the accuracy and reliability of the overall numerical analysis. According to the change of the overall mutation amplitude in each time period, the state change index of the garbage area in each time period is determined; combining the characteristic values and the state change index, the possibility index of the garbage area being an aluminum can area is determined, and the aluminum can areas are screened according to the possibility index of each garbage area. In summary, the present invention can combine shape features and gray features, perform characteristic value analysis, and further determine a more accurate and reliable possibility index according to characteristic mutations, combined with the influence of the rollers on the conveyor belt, improve the stability of aluminum can classification recognition, make the classification recognition effect of aluminum cans better and the accuracy higher. Description of the Drawings

[0037] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0038] Figure 1 It is a flowchart of a method for classifying and recognizing solid waste for recycling provided by an embodiment of the present invention;

[0039] Figure 2 It is a schematic diagram of an aluminum can in a parallel state and a state biased towards the vertical state provided by an embodiment of the present invention;

[0040] Figure 3 It is a schematic diagram of the conveyor belt roller area provided by an embodiment of the present invention. Detailed Embodiments

[0041] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the solid waste classification and identification method and system for recycling proposed by the present invention, its specific implementation method, structure, characteristics and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0042] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0043] The following is a detailed description of a method for classifying and identifying solid waste for recycling provided by the present invention in conjunction with the accompanying drawings.

[0044] See also Figure 1 , which shows a flow chart of a solid waste classification and identification method for recycling provided by an embodiment of the present invention, the method comprising:

[0045] S101: Obtain garbage images on the conveyor belt at different sampling times, perform image analysis to determine different garbage areas; determine the characteristic value of the garbage area belonging to the can according to the matching features and grayscale values ​​of each garbage area and the circular fitting.

[0046] The scenario corresponding to the embodiment of the present invention is specifically a garbage sorting scenario. This solution is mainly for the sorting of aluminum cans. In garbage disposal sites, waste sorting lines, scrap recycling stations and other environments, the recycling of aluminum products such as cans is difficult. Traditional recycling methods such as magnetic separation and manual screening are inefficient, and cans are often misjudged or missed due to their morphological changes and different physical states. Therefore, image analysis can be combined to further classify and identify cans.

[0047] When garbage is recycled, the magnetic metal garbage in the originally recyclable garbage will be screened out when the garbage passes through the magnetic separator. However, since aluminum products are not magnetic, it is difficult to screen them out through the magnetic separator, so they will remain in the screened garbage for subsequent screening steps.

[0048] Since the main material of the can is aluminum, the material on the side is lighter and softer, and is easily deformed during garbage recycling. The bottom and top of the can are solidified by reinforcing ribs, so they are harder and less likely to deform than the can body. As a result, the garbage collected from the landfill is generally in two states:

[0049] One is flat after being crushed up and down, and the other may have partial deformation in the complete state. Aluminum is light in weight and large in volume, making it easy to be screened by a garbage air classifier and having a small volume. Therefore, the complete cans are present in the lighter part of the garbage after being screened by the garbage air classifier, while the crushed cans are relatively heavier and are likely to be in the same area as the rest of the heavier garbage after screening. This solution mainly analyzes the heavier group of garbage after the air classifier, that is, specifically analyzes the flat aluminum cans that have been crushed up and down.

[0050] In the embodiment of the present invention, image data at different time nodes are collected by a high-definition camera to obtain garbage images at different sampling moments. It should be noted that the garbage image is specifically the image information of the garbage above the conveyor belt. Therefore, during the collection process, means such as image preprocessing can be carried out to stably identify the garbage image and determine different garbage areas through image recognition means.

[0051] It should be noted that a garbage area represents the image area corresponding to a garbage individual. In the embodiment of the present invention, the division of garbage individuals can be performed through a pre-set artificial intelligence network model, and the specific method is a conventional technology in the art and will not be elaborated here. Of course, in some other embodiments, other division methods can also be used to determine the garbage area, and there is no limitation to this.

[0052] For the heavier part of the garbage after being screened by the garbage air classifier, the shape of the aluminum cans (all the aluminum cans mentioned subsequently are crushed ones) is mainly a flat cylindrical shape crushed up and down. Since it is smaller in volume and lower in center of gravity compared to the uncrushed aluminum cans, it is not easy to slide or roll on the conveyor belt. However, due to the reduced volume of the flat aluminum cans, they are easily blocked by other garbage, and there are two states on the conveyor belt, that is, the aluminum cans are parallel to the conveyor belt and perpendicular to the conveyor belt.

[0053] In different states, there are different image features in this area to judge the image-related features of this area belonging to the aluminum can. See Figure 2 , Figure 2 which is a schematic diagram of the aluminum can in a parallel state and a state inclined towards the vertical state provided by an embodiment of the present invention.

[0054] In the image, the shape of the aluminum can is variable due to being stepped on, but the changes in the upper and lower bottoms and lids of the aluminum can are relatively small, and it can be used as the main analysis target. On the conveyor belt, when it is parallel to the conveyor belt, the area is close to a circle, and when it changes towards the vertical direction, it is also similar to an ellipse. Therefore, circular fitting can be used for matching feature analysis.

[0055] Further, in some embodiments of the present invention, the characteristic value indicating that the garbage area belongs to an aluminum can is determined according to the matching characteristics of each garbage area with the circular fitting and the gray value, including: performing elliptical fitting on the garbage area to determine the distance between the two foci of the fitted ellipse as the circular matching distance; performing maximum-minimum normalization on the opposite number of the circular matching distance to obtain the circular matching index; calculating the average gray value of all pixel points after graying the garbage area to obtain the gray characteristic index; calculating the product of the circular matching index and the gray characteristic index, and performing maximum-minimum normalization to obtain the characteristic value.

[0056] Among them, the closer the distance between the two foci obtained by elliptical fitting is, the more circular the garbage area appears. Therefore, performing maximum-minimum normalization on the opposite number of the circular matching distance to obtain the circular matching index, the larger the value of the circular matching index, the more the corresponding garbage area conforms to the characteristics of the aluminum can parallel conveyor belt. And the smaller the value of the circular matching index, further discussion is needed.

[0057] In the embodiments of the present invention, since the magnetic separator is used to screen out the remaining magnetic metals, most of the remaining items on the conveyor belt have no metallic luster except for aluminum cans. Therefore, for each garbage area, the glossiness within the garbage area can be calculated. Due to the reflection characteristics of the metal material, the surface of the flattened aluminum can will generate relatively uniform reflected light, while other types of garbage (such as plastics or papers) usually do not have this uniformity. Therefore, in each garbage area of each image, the greater the reflective characteristic, the greater the possibility that it is an aluminum can area; superimposing the white characteristic corresponding to the aluminum can, the reflective characteristic can be specifically analyzed by the gray value, and has a better analysis effect. In the embodiments of the present invention, the average gray value of all pixel points after graying the garbage area is calculated to obtain the gray characteristic index.

[0058] Therefore, the characteristic value is obtained by combining the product of the circular matching index and the gray characteristic index. The larger the characteristic value, the more the corresponding garbage area conforms to the characteristics of the aluminum can. However, there are still aluminum cans that tend to be in a vertical state and do not present a circular feature in the image. In addition, the aluminum cans are prone to jolts on the conveyor belt, resulting in state changes. Therefore, further analysis is needed. For the specific analysis process, please refer to the subsequent embodiments.

[0059] S102: Determine the mutation amplitude of the same garbage area at each sampling moment according to the change of the characteristic value of the same garbage area at different positions on the conveyor belt at different sampling moments; determine the candidate mutation time nodes of the conveyor belt causing garbage changes according to the value of the mutation amplitude.

[0060] In the embodiments of the present invention, the optical flow method can be used to specifically analyze the same garbage area at different sampling times. Alternatively, since the conveyor belt itself moves at a constant speed, the same garbage area can also be matched and identified according to the uniform motion to determine the same garbage area at different sampling times.

[0061] In the time series of each area, since there are multiple states of the aluminum can, and due to the movement and jolting of the conveyor belt, its state is constantly changing. During the movement and jolting of the conveyor belt, the aluminum can will continuously change between two states parallel and perpendicular to the conveyor belt. Therefore, its characteristic values will show high and low fluctuations in time series. However, since the conveyor belt is jolted differently at several different positions during movement, the fluctuation amplitude of its characteristic values changes differently.

[0062] It can be understood that in the related art, specific analysis is carried out through the overall characteristic values. However, since there are roller supports on the conveyor belt, it will cause jolting at the roller positions, which in turn leads to easy changes in the state of the garbage. On a normal conveyor belt, its change is a small vibration, while when passing through the roller area on the conveyor belt, it is jolted more severely. Refer to Figure 3 , Figure 3 which is a schematic diagram of the conveyor belt roller area provided by an embodiment of the present invention, and it also has different effects on the state of the garbage on the conveyor belt. And the roller area will cause errors in the overall judgment. Therefore, further analysis needs to be combined with the roller area.

[0063] Furthermore, in some embodiments of the present invention, according to the change of the characteristic values of the same garbage area at different positions on the conveyor belt at different sampling times, the mutation amplitude of the same garbage area at each sampling time is determined, including: determining any sampling time as the target time, calculating the difference between the characteristic values of the target time and its previous time as the first difference; calculating the difference between the characteristic values of the target time and its next time as the second difference; taking the normalized value of the sum of the first difference and the second difference as the mutation amplitude at the target time.

[0064] That is to say, the difference between the characteristic values of any sampling time and its two adjacent sampling times before and after is subjected to mutation analysis. The larger the value of the mutation amplitude, the more obvious the state mutation at the corresponding target time.

[0065] Furthermore, in some embodiments of the present invention, according to the numerical value of the mutation amplitude, the candidate mutation time nodes for the conveyor belt to cause garbage changes are determined, including: taking the time nodes with the mutation amplitude greater than the preset amplitude threshold as the candidate mutation time nodes.

[0066] In an embodiment of the present invention, the preset amplitude threshold is the threshold value of the mutation amplitude. Since the larger the value of the mutation amplitude, the more obvious the state mutation at the corresponding target time, therefore, a threshold analysis is directly performed. Optionally, the preset amplitude threshold can be specifically, for example, 0.7. That is to say, the time nodes with a mutation amplitude greater than 0.7 are used as candidate mutation time nodes.

[0067] The candidate mutation time nodes represent the time nodes when the garbage state changes greatly. Whether the specific reason for the change is due to the roller needs to be further analyzed.

[0068] Since each area of the image vibrates at each moment, the vibration caused by the movement of the conveyor belt is a high-frequency vibration with a relatively low frequency. It is manifested as that each garbage area in the image will change (that is, due to the rollers on the conveyor belt, the garbage on the conveyor belt undergoes a large bump. This bump has a large amplitude, but its frequency changes with the distribution of the rollers. When the garbage is bumped at this position, the change range is large and the degree of change is large, making the garbage distribution on the conveyor belt appear similar to a reset effect). And due to the low-frequency high-amplitude fluctuation caused by the conveyor belt passing over the rollers (similar to the above, but this is a small-range vibration caused by the movement of the conveyor belt. Its amplitude is not large, but its frequency is relatively larger than the above bump), so the classification in each time period calculated according to the candidate mutation nodes above is not obvious. Among the candidate mutation time nodes, there may be mutations caused by non-axles when the conveyor belt moves, and further screening is needed.

[0069] S103: According to the candidate mutation time nodes corresponding to each garbage area, screen to obtain the target time nodes when the garbage state mutation occurs on the conveyor belt. Taking the target time nodes as the cutting points, determine different time periods of each garbage area on the conveyor belt. According to the change of the overall mutation amplitude in each time period, determine the state change index of the garbage area in each time period.

[0070] Since the reasons for the large changes in mutations occurring on the conveyor belt are different, it is necessary to distinguish the time nodes with a larger degree of mutation among them, by analyzing whether the changes between the nodes are caused by the rollers on the conveyor belt or are other relatively soft garbage itself.

[0071] Further, in some embodiments of the present invention, according to the candidate mutation time nodes corresponding to each garbage area, screening to obtain the target time nodes when the garbage state mutation occurs on the conveyor belt includes: determining the frequency of the sampling moments corresponding to the candidate mutation time nodes of all garbage areas, and taking the preset number of candidate mutation time nodes with the largest frequency as the target time nodes.

[0072] Among them, the preset quantity can be the same as the number of rollers. If there are 10 rollers on the conveyor belt, the preset quantity can be set to 10. In the embodiments of the present invention, since relatively large mutations may also occur at other positions due to the soft characteristics of the garbage itself, resulting in mutation characteristics, but the roller mutations are relatively stable. Therefore, the preset number of candidate mutation time nodes with the largest frequency is used as the target time node. The target time node caused by the roller mutation can be directly screened out.

[0073] After determining the target time node, taking the target time node as the cutting point, different time periods of each garbage area on the conveyor belt are determined. Then, within the same time period, other soft-textured garbage will change frequently, while the state change of hard materials such as aluminum cans is relatively small. The state changes in each time period can be specifically analyzed.

[0074] Furthermore, in some embodiments of the present invention, according to the change in the overall mutation amplitude in each time period, the state change index of each garbage area is determined, including: determining the mutation stability within the time period according to the dispersion of all mutation amplitudes within the same time period; calculating the opposite number of the mean value of all mutation amplitudes within the same time period, and performing maximum-minimum normalization processing as the numerical feature index; multiplying the numerical feature index by the mutation stability as the state change index of the garbage area within the corresponding time period.

[0075] Among them, the dispersion analysis can specifically use the standard deviation. That is to say, calculate the opposite number of the standard deviation of all mutation amplitudes within the same time period, and perform maximum-minimum normalization processing as the mutation stability. The mutation stability is a stability characteristic representing the overall mutation amplitude within the time period. The larger the value of the mutation stability, the more stable it is.

[0076] And the opposite number of the mean value of all mutation amplitudes within the same time period, after maximum-minimum normalization processing, is used as the numerical feature index. The larger the numerical feature index, the smaller the overall numerical value of the mutation amplitude, and the smaller the generated mutation.

[0077] Therefore, directly multiply the numerical feature index by the mutation stability as the state change index of the garbage area within the corresponding time period. Then, the larger the value of the state defense index, the more stable the state change of the garbage area within the corresponding time period, and the smaller the overall numerical value.

[0078] S104: Combine the characteristic values and the state change index to determine the possibility index that the garbage area is an aluminum can area, and screen the aluminum can area according to the possibility index of each garbage area.

[0079] Further, in some embodiments of the present invention, determining the possibility that the garbage area is a canned drink area by combining characteristic values and state change indicators includes: calculating the mean value of the characteristic values at all sampling moments within the same time period as the segment characteristic coefficient; calculating the product of the segment characteristic coefficient and the state change indicator within the same time period as the influence indicator for the time period; and taking the mean value of the influence indicators for all time periods as the possibility indicator for the corresponding garbage area being a canned drink area.

[0080] Among them, for the analysis of all time periods, that is, by combining the segment influence situations within all time periods, and since there are corresponding characteristic values and state change indicators at each sampling moment within the same time period, directly calculating the product of the segment characteristic coefficient and the state change indicator within the same time period as the influence indicator for the time period, the influence indicator for the time period represents the overall change characteristics and the similarity between the characteristics and canned drinks of the garbage area within the time period.

[0081] This similarity situation can only represent the characteristics of one time period, and it is also necessary to combine all time periods. Therefore, taking the mean value of the influence indicators for all time periods as the possibility indicator for the corresponding garbage area being a canned drink area. The possibility indicator represents the possibility that the garbage area is a canned drink area. Due to the combination of the analysis of the bumps of the overall conveyor belt, this possibility indicator has higher accuracy and objectivity.

[0082] Further, in some embodiments of the present invention, screening the canned drink area according to the possibility indicator of each garbage area includes: taking the garbage area with a possibility indicator greater than the preset possibility threshold as the canned drink area.

[0083] Among them, the preset possibility threshold is the threshold value of the possibility indicator. In the embodiments of the present invention, the preset possibility threshold can be set to 0.7, that is to say, taking the garbage area with a possibility indicator greater than 0.7 as the canned drink area. Thus, the accurate identification of the canned drink area is realized.

[0084] The present invention obtains garbage images at different sampling moments on a conveyor belt, performs image analysis to determine different garbage areas; determines characteristic values indicating that the garbage areas belong to aluminum cans based on the matching characteristics and gray values of each garbage area fitted with a circle; the characteristic values characterize the aluminum can characteristics of the garbage areas themselves, so subsequent numerical analysis and change analysis can be performed based on the characteristic values. Then, according to the changes in the characteristic values of the same garbage area at different positions on the conveyor belt at different sampling moments, the mutation amplitude of the same garbage area at each sampling moment is determined; the candidate mutation time nodes of the conveyor belt causing garbage changes are determined based on the numerical values of the mutation amplitudes; according to the candidate mutation time nodes corresponding to each garbage area, the target time nodes of the garbage state mutation on the conveyor belt are screened. Taking the target time nodes as cutting points, different time periods of each garbage area on the conveyor belt are determined. The division of the time periods can effectively avoid the "state reset" influence caused by bumpy rollers, improving the accuracy and reliability of the overall numerical analysis. According to the changes in the overall mutation amplitude in each time period, the state change index of the garbage area in each time period is determined; combining the characteristic values and the state change index, the possibility index of the garbage area being an aluminum can area is determined, and the aluminum can areas are screened according to the possibility index of each garbage area. In summary, the present invention can combine shape characteristics and gray characteristics to perform characteristic value analysis, and further determine a more accurate and reliable possibility index based on characteristic mutations and considering the influence of rollers on the conveyor belt, making the classification and recognition effect of aluminum cans better and more accurate.

[0085] The present invention also provides a classification and recognition system for solid waste for recycling. The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the foregoing classification and recognition method for solid waste for recycling are implemented.

[0086] It should be noted that: the above sequence of embodiments of the present invention is only for description and does not represent the superiority or inferiority of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0087] Each embodiment in this specification is described in a progressive manner. The same or similar parts among the embodiments can be referred to each other, and the key points of each embodiment are the differences from other embodiments.

Claims

1. A method for classifying and identifying solid waste for recycling, characterized in that: The method comprises: Obtain garbage images on the conveyor belt at different sampling times, perform image analysis to determine different garbage areas; determine the characteristic value of the garbage area belonging to the can based on the matching features and grayscale values ​​of each garbage area and the circular fitting; According to the changes in characteristic values ​​of the same garbage area at different positions on the conveyor belt at different sampling times, the mutation amplitude of the same garbage area at each sampling time is determined; according to the value of the mutation amplitude, the candidate mutation time node at which the conveyor belt causes garbage changes is determined; According to the candidate mutation time nodes corresponding to each garbage area, the target time nodes at which garbage status mutations occur on the conveyor belt are screened out, and the target time nodes are used as cutting points to determine the different time periods of each garbage area on the conveyor belt. According to the changes in the overall mutation amplitude in each time period, the status change index of the garbage area in each time period is determined; The characteristic value and the state change index are combined to determine the possibility index that the garbage area is a can area, and the can area is screened according to the possibility index of each garbage area.

2. A solid waste classification and identification method for recycling as claimed in claim 1, characterized in that: The method of determining the characteristic value of the garbage area belonging to the can according to the matching features and grayscale values ​​of each garbage area and the circular fitting includes: Performing ellipse fitting on the garbage area, and determining the distance between the two foci of the fitted ellipse as the circle matching distance; Normalizing the inverse of the circle matching distance to its maximum and minimum values ​​to obtain a circle matching index; Calculate the mean grayscale value of all pixels in the garbage area after grayscale conversion to obtain the grayscale feature index; The product of the circle matching index and the grayscale feature index is calculated, and the maximum and minimum values ​​are normalized to obtain the characteristic value.

3. A solid waste classification and identification method for recycling as claimed in claim 1, characterized in that: According to the changes in characteristic values ​​of the same garbage area at different positions on the conveyor belt at different sampling times, the mutation amplitude of the same garbage area at each sampling time is determined, including: Determine any sampling moment as the target moment, calculate the difference between the characteristic value of the target moment and the moment before it as the first difference; calculate the difference between the characteristic value of the target moment and the moment after it as the second difference; The normalized value of the sum of the first difference and the second difference is used as the mutation amplitude at the target moment.

4. A solid waste classification and identification method for recycling as claimed in claim 1, characterized in that: Determining a candidate mutation time node at which the conveyor belt causes garbage changes according to the value of the mutation amplitude includes: The time node at which the mutation amplitude is greater than the preset amplitude threshold is used as the candidate mutation time node.

5. A solid waste classification and identification method for recycling as claimed in claim 1, characterized in that: According to the candidate mutation time nodes corresponding to each garbage area, the target time nodes for generating garbage status mutation on the conveyor belt are screened, including: Determine the frequency of sampling moments corresponding to the candidate mutation time nodes of all garbage areas, and take a preset number of candidate mutation time nodes with the largest frequency as target time nodes.

6. A solid waste classification and identification method for recycling as claimed in claim 1, characterized in that: According to the change of the overall mutation amplitude in each time period, the state change indicators of the garbage area in each time period are determined, including: According to the dispersion of all mutation amplitudes in the same time period, the mutation stability within the time period is determined; Calculate the opposite of the mean of all mutation amplitudes in the same time period, and normalize the maximum and minimum values ​​as numerical feature indicators; The product of the numerical feature index and the mutation stability is used as the state change indicator of the garbage area in the corresponding time period.

7. A solid waste classification and identification method for recycling as claimed in claim 6, characterized in that: Based on the discreteness of all mutation amplitudes in the same time period, the mutation stability within the time period is determined, including: The inverse of the standard deviation of all mutation amplitudes in the same time period was calculated, and the maximum and minimum values ​​were normalized to obtain mutation stability.

8. A solid waste classification and identification method for recycling as claimed in claim 1, characterized in that: Combining the characteristic value and the state change indicator, determining the possibility that the garbage area is a can area includes: Calculate the mean of the characteristic values ​​at all sampling moments in the same time period as the segment characteristic coefficient; Calculate the product of the segment characteristic coefficient and the state change index in the same time period as the impact index of the time period; The mean of the impact indicators of all time periods is used as the possibility indicator that the corresponding garbage area is a can area.

9. A solid waste classification and identification method for recycling as claimed in claim 1, characterized in that: The can areas are screened based on the aforementioned likelihood indicators for each garbage area, including: The garbage area whose possibility index is greater than the preset possibility threshold is regarded as the can area.

10. A solid waste classification and identification system for recycling, the system comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 9 are implemented.

Images