Tobacco leaf surface mildew identification and tracking method and system

Abstract

The invention discloses a tobacco leaf surface mildew identification tracking method and system, and relates to the technical field of mildew identification, and the method comprises the steps: obtaining a tobacco leaf image on the surface of a conveying belt, and carrying out the hidden danger mildew feature identification, and obtaining a hidden danger image; performing hidden danger light band identification based on the hidden danger image to obtain a screened hidden danger area; judging whether tobacco leaf shielding exists or not based on the screening hidden danger area; if not, abnormal light band identification is carried out based on the screening hidden danger area, and a first abnormal area is obtained; if yes, determining a jitter range based on the tobacco leaf shielding position and the corresponding hidden danger light band; adjusting and screening tobacco leaves shielded in the hidden danger area based on the shaking range, and then collecting an adjustment image; performing abnormal light band identification based on the adjusted image to obtain a second abnormal region; obtaining a corresponding ray tracing instruction and abnormal early warning information based on the first abnormal area or the second abnormal area; and tracking the corresponding abnormal region based on the ray tracking instruction. And the tobacco leaf surface mildew identification and tracking accuracy is improved.

Description

Technical Field

[0001] This invention relates to the field of mold identification technology, and more specifically to a method and system for identifying and tracking mold growth on the surface of tobacco leaves. Background Technology

[0002] Tobacco leaves typically undergo initial curing and re-curing processes before being stored in warehouses to await further processing. During storage, the cured tobacco leaves undergo natural fermentation, which improves their combustibility, aroma, and color. Because the natural fermentation process is lengthy and requires controlled temperature and humidity levels within the storage space, these conditions often create ideal environments for mold growth, such as Aspergillus and Penicillium. Rapid mold proliferation can lead to surface mold growth on the tobacco leaves. Failure to promptly remove moldy leaves can negatively impact the overall quality of the tobacco.

[0003] In existing technologies, before processing naturally fermented tobacco leaves for flavoring, shredding, etc., moldy tobacco leaves are typically selected from the broken-up fermented tobacco leaves through manual screening to ensure the quality of the tobacco leaves. However, since the number of tobacco leaves to be screened is large, and the accuracy of manual screening is often affected by various factors, such as the experience and concentration of the screening personnel, the accuracy of mold identification during the screening process is low, and it is easy to miss some, which will affect the quality of the tobacco leaves.

[0004] Therefore, how to improve the accuracy of identifying and tracking mold growth on tobacco leaves, and thus ensure the quality of tobacco leaves, is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] In view of this, the present invention provides a method and system for identifying and tracking mold growth on the surface of tobacco leaves, which improves the accuracy of identifying and tracking mold growth on the surface of tobacco leaves, thereby ensuring the quality of tobacco leaves.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A method for identifying and tracking mold growth on tobacco leaf surfaces includes:

[0008] Images of tobacco leaves on the conveyor belt surface were acquired and their potential mold and mildew characteristics were identified to obtain potential hazard images;

[0009] Based on the hazard image, hazard light band identification is performed to obtain the screened hazard areas;

[0010] Based on the screening of potential hazard areas, it is determined whether there is tobacco leaf obstruction;

[0011] If not, then based on the screened potential hazard area, abnormal light band identification is performed to obtain the first abnormal area;

[0012] If so, the shaking range is determined based on the location of the tobacco leaf obstruction and the corresponding potential hazard light band;

[0013] After adjusting the obscured tobacco leaves in the screened potential hazard area based on the shaking range, an adjusted image was acquired.

[0014] Based on the adjusted image, abnormal light bands are identified to obtain a second abnormal region;

[0015] Based on the first or second abnormal region, corresponding ray tracing instructions and abnormal warning information are obtained;

[0016] The ray tracing command is used to track the corresponding abnormal area.

[0017] Preferred areas are selected to identify potential hazards, specifically including:

[0018] Based on the image of the potential hazard, identify the affected moldy area containing the moldy characteristics of the potential hazard;

[0019] Based on the aforementioned areas with potential mold growth, obtain the corresponding regional spectral data;

[0020] Based on the regional spectral data, the hidden danger light band is identified to obtain the spectral data of the screening region containing the hidden danger light band;

[0021] The potential moldy areas corresponding to the spectral data of the selected areas are used as the selected potential hazard areas.

[0022] Preferably, the jitter range is determined, specifically including:

[0023] Based on the identified potential hazard areas, the occlusion contour information corresponding to each tobacco leaf occlusion location is identified;

[0024] The amount of tobacco leaves accumulated at the corresponding tobacco leaf occlusion position is determined based on the occlusion contour information.

[0025] The corresponding flat area is determined based on the amount of tobacco leaves piled up.

[0026] The initial shaking range is obtained based on the position of the tobacco leaf obstruction and its corresponding flat area;

[0027] The corresponding expanded area is obtained based on the hidden danger light band corresponding to the position of the tobacco leaf obstruction;

[0028] The initial jitter range is optimized based on the expanded area to obtain the final jitter range.

[0029] Preferably, the method for obtaining the initial jitter range is as follows:

[0030] The target distance is the straight-line distance from the highest edge point to the conveyor belt surface in the three-dimensional edge information corresponding to the tobacco leaf blocking position.

[0031] The area difference is obtained based on the mapping relationship between the target distance and the preset area difference.

[0032] The flat area corresponding to the tobacco leaf shading position is obtained by adding the area difference and the flat area;

[0033] Identify the corresponding tiling edge information based on the tiling range;

[0034] The initial jitter range is obtained by integrating all the tiling ranges based on the information of each tiling edge.

[0035] Preferably, the method for generating ray tracing instructions is as follows:

[0036] The abnormal region is defined as either the first abnormal region or the second abnormal region.

[0037] Based on the abnormal region, the corresponding abnormal optical band is obtained;

[0038] The visibility of the tracking ray is determined based on the anomalous light band.

[0039] The real-time movement position of the abnormal area is identified based on the tobacco leaf image on the conveyor belt surface;

[0040] The ray tracing command is generated based on the ray visibility and the real-time moving position.

[0041] Preferred options also include:

[0042] When multiple abnormal areas exist within a first preset time period, the number of the potential moldy features in all the abnormal areas is identified as the number of potential hazards.

[0043] Based on the number of potential hazards, determine whether it exceeds a set value;

[0044] If so, then generate a shutdown command and the first prompt message;

[0045] Otherwise, the irradiation sequence and irradiation duration of each abnormal region are determined based on the abnormal light band corresponding to each abnormal region;

[0046] Based on the irradiation sequence, the irradiation duration, and the corresponding ray tracing command, a corresponding target ray tracing command is obtained;

[0047] Based on the target ray tracing command, the target tracking device is controlled to perform real-time tracking of all the abnormal areas.

[0048] Preferably, the method for obtaining the irradiation sequence of the abnormal region is as follows:

[0049] The light difference is based on the difference between the abnormal light band and the preset standard light band corresponding to each abnormal region.

[0050] Sort all the light wave differences in descending order to obtain the light wave difference sequence;

[0051] Based on the light wave difference sequence, the abnormal regions corresponding to each light wave difference are sorted to obtain the irradiation order corresponding to each abnormal region.

[0052] Preferred options also include:

[0053] Integrate multiple mold characteristics identified within a second preset time period;

[0054] Based on the comparison of each of the aforementioned mold characteristics, the mold similarity is obtained;

[0055] Based on the mold similarity, determine whether it is greater than a preset similarity;

[0056] If so, then the corresponding abnormal conveyor belt area and conveyor belt conveying rate are determined based on the abnormal area corresponding to each of the moldy characteristics;

[0057] The timing of tobacco leaf feeding is determined based on the abnormal conveyor belt area and the conveyor belt speed.

[0058] The target tobacco raw material can be traced based on the time of tobacco leaf feeding.

[0059] Based on the target tobacco raw material, an inspection instruction and a corresponding second prompt message are generated;

[0060] The target tobacco raw material is subjected to a safety inspection based on the inspection command and the second prompt information.

[0061] Preferably, the method for determining the timing of tobacco leaf feeding is as follows:

[0062] The conveying distance of the moldy tobacco leaves is determined based on the abnormal conveyor belt area corresponding to each of the moldy characteristics.

[0063] The time when the moldy tobacco leaves arrive at the conveyor belt is determined by the ratio of the conveying distance to the conveyor belt speed, and this time is taken as the unloading time of the tobacco leaves.

[0064] A system for identifying and tracking mold growth on the surface of tobacco leaves includes: a module for identifying potential mold growth, a module for judging tobacco leaf occlusion, a module for adjusting occluded tobacco leaves, a module for recognizing adjusted images, and a module for tracking abnormal areas.

[0065] The hidden danger mold identification module is used to acquire images of tobacco leaves on the surface of the conveyor belt and identify hidden danger mold features to obtain a hidden danger image; based on the hidden danger image, the hidden danger light band is identified to obtain the screened hidden danger area;

[0066] The tobacco leaf obstruction detection module is used to determine whether there is tobacco leaf obstruction based on the screened potential hazard area; if not, it performs abnormal light band identification based on the screened potential hazard area to obtain the first abnormal area; if so, it determines the shaking range based on the tobacco leaf obstruction position and its corresponding potential hazard light band.

[0067] The obscured tobacco leaf adjustment module is used to adjust the obscured tobacco leaves in the screened hazard area based on the shaking range and then collect the adjusted image;

[0068] The adjusted image recognition module is used to identify abnormal light bands based on the adjusted image to obtain a second abnormal region;

[0069] The abnormal region tracking module is used to obtain corresponding ray tracing instructions and abnormal warning information based on the first abnormal region or the second abnormal region; and to track the corresponding abnormal region based on the ray tracing instructions.

[0070] As can be seen from the above technical solution, compared with the prior art, the present invention discloses a method and system for identifying and tracking mold growth on the surface of tobacco leaves, which has the following beneficial effects:

[0071] 1. This invention identifies areas in the conveyor belt where tobacco leaves may be moldy by recognizing images of tobacco leaves on the conveyor belt surface. When potential moldy features are found in the tobacco leaf images on the conveyor belt surface, spectral data analysis is performed on the corresponding areas instead of directly performing spectral data analysis on the entire conveyor belt surface image. This improves the recognition speed while reducing the computational burden in the mold identification process. By performing spectral data analysis on abnormal areas, it is possible to determine whether the abnormal areas contain moldy features, making it easier to capture subtle changes that are difficult to detect with the naked eye, thereby improving the accuracy of moldy feature identification.

[0072] 2. This invention obtains the shielding contour information corresponding to each tobacco leaf shielding position, thereby obtaining the tobacco leaf accumulation amount and the flat area required to occupy on the conveyor belt for each shielding position. Based on each flat area, an initial shaking range is determined, which ensures that the accumulated tobacco leaves at each shielding position can be effectively moved during the shaking of the conveyor belt. The expansion area is determined by the hidden danger light band corresponding to each tobacco leaf shielding position, and the initial shaking range is optimized based on the expansion area, which facilitates ensuring that the tobacco leaves at the shielding positions with a high degree of hidden danger are fully moved, thereby improving the accuracy of determining abnormal light bands.

[0073] 3. This invention generates timely abnormal warning prompts, which helps to remind relevant personnel to remove tobacco leaves with moldy characteristics from the conveyor belt in a timely manner, thus preventing the spread of mold problems and ensuring the stable operation of the tobacco production line and the quality of tobacco leaves. Attached Figure Description

[0074] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0075] Figure 1 The flowchart of a method for identifying and tracking mold growth on the surface of tobacco leaves provided by the present invention.

[0076] Figure 2 The flowchart illustrates the method for determining the jitter range provided by this invention.

[0077] Figure 3 Provided by the present invention Detailed Implementation

[0078] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0079] Example 1

[0080] like Figure 1 As shown in the figure, an embodiment of the present invention discloses a method for identifying and tracking mold growth on the surface of tobacco leaves, including:

[0081] Images of tobacco leaves on the conveyor belt surface were acquired and their potential mold and mildew characteristics were identified to obtain potential hazard images;

[0082] Based on the images of potential hazards, the light bands of potential hazards are identified to obtain the screened hazard areas;

[0083] Based on the screening of potential hazard areas, determine whether there is tobacco leaf obstruction;

[0084] If not, then based on the screening of potential hazard areas, abnormal light bands are identified to obtain the first abnormal area;

[0085] If so, the shaking range is determined based on the location of the tobacco leaf obstruction and the corresponding potential hazard light band;

[0086] After adjusting the shaking range to filter out obscured tobacco leaves in the potential hazard area, the image was collected and adjusted.

[0087] Based on the adjusted image, abnormal light bands are identified to obtain the second abnormal region;

[0088] Based on the first or second abnormal region, obtain the corresponding ray tracing instructions and abnormal warning information;

[0089] The ray tracing command is used to track the corresponding abnormal areas.

[0090] Example 2

[0091] This invention discloses a method for identifying and tracking mold growth on the surface of tobacco leaves, comprising:

[0092] Images of tobacco leaves on the conveyor belt surface are acquired and their potential mold and mildew characteristics are identified to obtain potential hazard images.

[0093] Preferably, after natural fermentation, the tobacco leaves may become more compact, which is not conducive to subsequent processing. Therefore, before processing the fermented tobacco leaves, they need to be broken up. Loosening the tobacco leaves facilitates subsequent processes such as seasoning, shredding, and rolling. The broken tobacco leaves are generally conveyed to the next workstation via a conveyor belt. During the conveying process, the presence of moldy characteristics in the tobacco leaves is identified. When moldy tobacco leaves are found, they can be removed from the conveyor belt in a timely manner to prevent moldy tobacco leaves from mixing with non-moldy tobacco leaves and affecting the quality of the tobacco leaves.

[0094] Preferably, in this embodiment, an image acquisition device positioned above the conveyor belt acquires images of tobacco leaves on the conveyor belt surface. When the image of tobacco leaves on the conveyor belt surface contains features of potential mold growth, it indicates that the image may or may not contain moldy tobacco leaves. As a natural plant product, tobacco leaves may have various natural textures, spots, etc., on their surface. Especially after natural fermentation, the natural textures and spots on the surface of tobacco leaves may be similar to mold features in some cases. There is a high degree of similarity between potential mold growth features and mold features. Potential mold growth features are determined by relevant personnel based on historical mold features and then uploaded to the identification system. Specific potential mold growth features are not specifically limited in this embodiment.

[0095] Preferably, when the image of tobacco leaves on the conveyor belt surface contains moldy features, it indicates that the image of tobacco leaves on the conveyor belt surface contains moldy tobacco leaves. At this time, a prompt message or a stop command can be generated directly so that relevant personnel can remove the moldy tobacco leaves from the conveyor belt in a timely manner. Since the growth and reproduction of mold on tobacco leaves is a complex process, its morphology and color may vary depending on the type of mold, the growth environment, and the growth stage. Therefore, the specific moldy features are not specifically limited in this application embodiment. They can be determined by relevant personnel based on historical experimental data and then uploaded to the identification system.

[0096] Preferably, when the tobacco leaf image on the conveyor belt surface contains potential moldy features, it indicates that the tobacco leaf image on the conveyor belt surface may contain moldy tobacco leaves, which needs to be further confirmed. The tobacco leaf image on the conveyor belt surface is divided by the identified potential moldy features, and the conveyor belt area containing the potential moldy features is located as the potential area. The area corresponding to the potential area is not specifically limited in this embodiment of the application, as long as it can be guaranteed that the identified potential area contains the potential moldy features.

[0097] Based on the images of potential hazards, the light bands of potential hazards are identified to obtain the selected potential hazard areas.

[0098] Preferred areas are selected to identify potential hazards, specifically including:

[0099] Identify the affected moldy areas containing moldy characteristics based on the images of potential hazards;

[0100] Obtain the corresponding regional spectral data based on the areas with potential mold growth;

[0101] Based on regional spectral data, hazard light band identification is performed to obtain spectral data of the screened region containing hazard light bands;

[0102] The potential hazard areas are selected based on the spectral data of the selected areas.

[0103] Preferably, the regional spectral data corresponding to the potentially moldy area can be obtained by scanning the tobacco leaves corresponding to the potentially moldy area using a spectral imaging device.

[0104] Preferably, when the regional spectral data contains the hazard light band, the characterization screening area may contain moldy tobacco leaves. However, the moldy characteristics of the moldy tobacco leaves are relatively vague. In this case, it is necessary to analyze whether there is tobacco leaf obstruction in the screening hazard area.

[0105] Based on the screening of potential hazard areas, it is determined whether there is tobacco leaf obstruction.

[0106] Preferably, the hazard areas containing the hazard light band in the regional spectral data can be identified as the conveyor belt areas of concern, and the corresponding conveyor belt images of concern can be determined from the tobacco leaf images on the conveyor belt surface. The conveyor belt images of concern are then imported into a trained occlusion recognition model to determine whether the conveyor belt images of concern contain tobacco leaf occlusion locations. The occlusion recognition model is obtained after training on a large amount of sample data, which includes sample images of the tobacco leaf surface of the conveyor belt with different degrees of tobacco leaf occlusion, and artificial labels corresponding to each sample image. The specific generation method of the occlusion recognition model is not specifically limited in the embodiments of this application, as long as it can determine whether the conveyor belt images of concern contain tobacco leaf occlusion locations.

[0107] If not, then based on the screening of potential hazard areas, abnormal light bands are identified to obtain the first abnormal area.

[0108] Preferably, the first abnormal region is obtained, specifically including:

[0109] Obtain the corresponding screening spectral data based on the screening of potential hazard areas;

[0110] Based on the screening of spectral data, abnormal light bands are identified to obtain the first abnormal spectral data containing the abnormal light bands.

[0111] The potential hazard area corresponding to the first abnormal spectral data is selected as the first abnormal area.

[0112] Preferably, in this embodiment, the screening spectral data corresponding to the screening potential hazard area is obtained by scanning the tobacco leaves corresponding to the screening potential hazard area with a spectral imaging device. When the tobacco leaves become moldy, the spectral characteristics of the mold spots will change. For example, the reflectivity of the mold spots to green light decreases, while the reflectivity to red light and near-infrared light increases. The specific abnormal light bands are not specifically limited in this embodiment of the application. They can be determined by relevant personnel based on historical experimental data and uploaded to the identification system. The degree of mold spots on the moldy tobacco leaves in the screening potential hazard area can be analyzed through the abnormal light bands.

[0113] Preferably, by identifying the tobacco leaf images on the conveyor belt surface, it is easier to discover areas in the conveyor belt that may have mold. When there are potential mold characteristics in the conveyor belt surface image, spectral data analysis is then performed on the corresponding area, rather than directly performing spectral data analysis on the entire conveyor belt surface tobacco leaf image. This improves the recognition speed while reducing the computational pressure in the mold recognition process.

[0114] If so, the shaking range is determined based on the location of the tobacco leaf obstruction and the corresponding potential hazard light band.

[0115] Preferred, such as Figure 2 As shown, the range of jitter is determined, specifically including:

[0116] Based on the screening of potential hazard areas, the occlusion contour information corresponding to each tobacco leaf occlusion location is identified;

[0117] The amount of tobacco leaves accumulated at the corresponding occlusion location is determined based on the occlusion contour information.

[0118] The corresponding flat area is determined based on the amount of tobacco leaf accumulation.

[0119] The initial shaking range is obtained based on the position of the tobacco leaves being covered and their corresponding flat area;

[0120] The corresponding expanded area is obtained based on the potential light band corresponding to the location of tobacco leaf obstruction.

[0121] The initial jitter range is optimized based on the expanded area to obtain the final jitter range.

[0122] Preferably, when the area obstructed by tobacco leaves exceeds a preset obstruction threshold, a screening hazard area is determined to have tobacco leaf obstruction. The corresponding shaking range is determined based on each tobacco leaf obstruction location, rather than determining the shaking range after only one or two obstructed tobacco leaves appear. Setting a preset obstruction threshold improves the effectiveness of the determined shaking range. The specific preset obstruction threshold can be 8 or 10; the exact number is not limited in this embodiment and can be set by relevant technical personnel according to actual needs. To improve the accuracy of the determined shaking range, the shaking range is determined based on all tobacco leaf obstruction locations and the corresponding hazard light band for each tobacco leaf obstruction location.

[0123] The preferred method for determining the amount of tobacco leaves accumulated at the location where the tobacco leaves are covered is as follows:

[0124] Construct a three-dimensional image of the tobacco leaf occlusion location based on the tobacco leaf image on the conveyor belt surface;

[0125] The occlusion contour information corresponding to the occlusion position of the tobacco leaf is determined from the 3D stereo image corresponding to the occlusion position of the tobacco leaf based on the edge detection algorithm;

[0126] The occlusion contour information serves as the three-dimensional edge information corresponding to the occlusion position of the tobacco leaf.

[0127] The amount of tobacco leaves accumulated at the location where tobacco leaves are obstructed is obtained based on three-dimensional edge information.

[0128] Preferably, in another embodiment, the display effect of the tobacco leaf obstruction position in different directions is obtained from the tobacco leaf image on the conveyor belt surface, and then the display effects in multiple directions are combined to obtain a three-dimensional image corresponding to the tobacco leaf obstruction position.

[0129] Preferably, in this embodiment, the shading contour information corresponding to the tobacco leaf shading position is determined by three-dimensional modeling. The method for determining the shading contour information is not specifically limited in this embodiment, as long as the tobacco leaf accumulation amount corresponding to the shading position can be calculated based on the determined shading contour information. Based on the above method, the tobacco leaf accumulation amount corresponding to each tobacco leaf shading position can be determined.

[0130] Preferably, based on the preset flat-laying mapping relationship and the amount of tobacco leaf pile, the flat-laying area corresponding to each tobacco leaf shading position is determined, and the preset flat-laying mapping relationship is the correspondence between the amount of tobacco leaf pile and the flat-laying area.

[0131] Preferably, the flat area corresponding to the tobacco leaf covering position is the area required after the piled tobacco leaves are laid flat. The larger the amount of tobacco leaves piled up, the larger the corresponding flat area. The flat area corresponding to each amount of tobacco leaves piled up can be determined by a preset flat mapping relationship. The preset flat mapping relationship can be regarded as the conversion relationship between the amount of tobacco leaves piled up and the flat area. The specific content of the preset flat mapping relationship is not specifically limited in this application embodiment. It can be determined by relevant personnel based on historical experimental data and then uploaded to the recognition system.

[0132] Preferably, the method for obtaining the initial jitter range is as follows:

[0133] The target distance is determined by the straight-line distance from the highest edge point to the conveyor belt surface in the three-dimensional edge information corresponding to the position of tobacco leaf obstruction.

[0134] The area difference is obtained based on the mapping relationship between the target distance and the preset area difference.

[0135] The flat area corresponding to the position where the tobacco leaves are blocked is obtained by adding the area difference and the flat area;

[0136] Identify the corresponding tile edge information based on the tile range;

[0137] The initial jitter range is obtained by integrating all the tile ranges based on the edge information of each tile.

[0138] Preferably, the center of any tobacco leaf shading position is determined as the center point of the flat-laying range. An initial shaking range is formed by spreading outward from the center point. Since the tobacco leaf may shift during the shaking process, the area of ​​the flat-laying range is larger than the flat-laying area, that is, there is an area difference between the range area and the flat-laying area.

[0139] Preferably, the preset area difference mapping relationship includes area differences corresponding to different straight-line distances. The specific content is not specifically limited in this application embodiment, and can be determined by relevant personnel based on historical experimental data and then uploaded to the recognition system.

[0140] Preferably, based on the preset edge expansion mapping relationship and the hidden danger light band corresponding to each tobacco leaf shading position, the edge expansion area corresponding to each tobacco leaf shading position is determined, and the edge expansion mapping relationship is the correspondence between the hidden danger light band and the edge expansion area.

[0141] Preferably, the closer the potential hazard light band is to the abnormal light band, the greater the probability that there is moldy tobacco at the corresponding tobacco leaf shading location. Sufficient shaking range needs to be provided for the tobacco leaf shading location. Based on the potential hazard light band corresponding to each tobacco leaf shading location, the expansion area corresponding to each tobacco leaf shading location is determined. After determining the expansion area corresponding to each tobacco leaf shading location, the corresponding flat edge information can be optimized based on the expansion area, thereby optimizing the initial shaking range. The optimized initial shaking range can be determined as the final shaking range.

[0142] Preferably, the method for determining the expanded area corresponding to the tobacco leaf shading position is as follows:

[0143] Calculate the abnormal difference wave between the potential hazard light band and the preset abnormal light band;

[0144] The expansion area corresponding to the abnormal difference wave is determined based on the preset expansion mapping relationship;

[0145] The smaller the abnormal difference value, the larger the corresponding expansion area.

[0146] The specific content of the preset edge expansion mapping relationship is not specifically limited in the embodiments of this application.

[0147] Preferably, by obtaining the shielding contour information corresponding to each tobacco leaf shielding position, the amount of tobacco leaf accumulation corresponding to each tobacco leaf shielding position is obtained, and then the flat area that each tobacco leaf shielding position needs to occupy on the conveyor belt is obtained. Based on each flat area, the initial shaking range is determined, which makes it easier to ensure that the accumulated tobacco leaves at each tobacco leaf shielding position can be effectively moved during the shaking of the conveyor belt.

[0148] Furthermore, by determining the expansion area based on the potential hazard light band corresponding to each tobacco leaf shading position, and optimizing the initial shaking range based on the expansion area, it is possible to ensure that the tobacco leaves at the tobacco leaf shading positions with a high degree of hazard are moved sufficiently, thereby improving the accuracy of determining abnormal light bands.

[0149] After adjusting the shaking range to filter out obscured tobacco leaves in the potential hazard area, the images were collected and adjusted.

[0150] Preferably, at least one local conveyor belt to be controlled is determined based on the shaking range, and the vibration frequency corresponding to the at least one local conveyor belt to be controlled is increased to promote the movement of tobacco leaves within the shaking range, thereby reducing the shading of tobacco leaves and improving the accuracy of determining abnormal light bands.

[0151] Preferably, the area of ​​the conveyor belt to be adjusted is a portion of the complete conveyor belt. The conveyor belt is in a vibrating state during normal operation. When it is necessary to adjust the vibration frequency of the area of ​​the conveyor belt to be adjusted, the specific adjustment value is not specifically limited in this embodiment, as long as the adjusted vibration frequency is higher than the original vibration frequency. By adjusting the vibration frequency of the conveyor belt corresponding to the shaking range, the movement of tobacco leaves within the shaking range is promoted, reducing tobacco leaf shading. Acquiring the adjusted image for abnormal light band identification can improve the accuracy of determining abnormal light bands.

[0152] Preferably, after determining the final vibration range, it is necessary to adjust the vibration frequency of the local conveyor belt to be adjusted within the vibration range. The vibration range may correspond to one local conveyor belt to be adjusted, or it may correspond to multiple local conveyor belts to be adjusted. The specific number is not specifically limited in this embodiment.

[0153] Based on the adjusted image, abnormal light bands are identified to obtain the second abnormal region.

[0154] Preferably, the second abnormal region is obtained, specifically including:

[0155] Obtain the corresponding adjusted spectral data based on the adjusted image;

[0156] Based on the adjusted spectral data, abnormal light bands are identified to obtain second abnormal spectral data containing abnormal light bands;

[0157] The adjusted image corresponding to the second anomalous spectral data is used as the second anomalous region.

[0158] Preferably, in this embodiment, the adjustment spectral data corresponding to the adjusted image is obtained by scanning the tobacco leaf corresponding to the adjusted image using a spectral imaging device. When the tobacco leaf becomes moldy, the spectral characteristics at the mold spot will change. For example, the reflectivity of the mold spot to green light decreases, while the reflectivity to red light and near-infrared light increases. The specific abnormal light band is not specifically limited in this embodiment of the application. It can be determined by relevant personnel based on historical experimental data and uploaded to the recognition system. The degree of mold spots on the moldy tobacco leaf in the adjusted image can be analyzed through the abnormal light band.

[0159] Based on the first or second abnormal region, corresponding ray tracing instructions and abnormal warning information are obtained.

[0160] Preferably, the method for generating ray tracing instructions is as follows:

[0161] Based on the first or second abnormal region as the abnormal region;

[0162] Obtain the corresponding abnormal optical bands based on the abnormal regions;

[0163] Determine the visibility of the tracking ray based on the anomalous light band;

[0164] Based on the real-time movement position of abnormal areas identified in tobacco leaf images on the conveyor belt surface;

[0165] Ray tracing instructions are generated based on ray visibility and real-time movement position.

[0166] Preferably, different abnormal light bands can reflect the degree of mold on moldy tobacco leaves. Therefore, the corresponding ray visibility can be set based on the abnormal light band corresponding to the moldy tobacco leaves. The greater the light wave difference between the abnormal light band and the preset standard light band, the more severe the mold is, and the higher the corresponding visible ray visibility. There is a corresponding relationship between the light wave difference and the ray visibility. Based on this relationship, the ray visibility corresponding to any light wave difference can be determined. The higher the ray visibility, the stronger the prompting to relevant personnel when using ray tracing to track moldy tobacco leaves based on ray visibility. The specific content of this relationship is not specifically limited in this application embodiment. It can be determined by relevant personnel based on historical experimental data and then uploaded to the identification system.

[0167] Preferably, since the conveyor belt is constantly in motion, when the moldy tobacco leaves in the abnormal area are not removed in time, the target tracking device needs to track the movement along with the movement of the moldy tobacco leaves. The tobacco leaf image on the conveyor belt surface is a real-time conveying image. Therefore, the real-time moving image corresponding to the abnormal area can be determined through the tobacco leaf image on the conveyor belt surface. Through feature recognition, the real-time moving position of the abnormal area, that is, the real-time moving position of the moldy tobacco leaves in the abnormal area, can be determined from the real-time moving image.

[0168] Preferably, by generating timely abnormal warning information, relevant personnel can be reminded to remove tobacco leaves with moldy characteristics from the conveyor belt in a timely manner, which helps to prevent the spread of mold problems and thus helps to ensure the stable operation of the tobacco production line and the quality of tobacco leaves.

[0169] The ray tracing command is used to track the corresponding abnormal areas.

[0170] Preferably, after generating a ray tracing command based on the real-time movement position and ray visibility, the ray tracing command is sent to the target tracking device, which controls the target tracking device to track the moldy tobacco leaves in the abnormal area in real time according to the real-time movement position and ray visibility, effectively guiding relevant personnel to remove the tobacco leaves with moldy characteristics from the abnormal area.

[0171] Preferably, the tracking device is a device with ray tracing capability. The tracking range of the tracking device is limited, that is, different abnormal areas may correspond to different tracking devices. When determining the target tracking device corresponding to the abnormal area, the tracking range of the target tracking device can include the abnormal area. The target tracking device is determined based on the abnormal area, and ray tracing is performed based on the target tracking device, so as to ensure that the tracking rays emitted by the target tracking device can accurately illuminate the abnormal area.

[0172] Preferred options also include:

[0173] When multiple abnormal areas exist within the first preset time period, the number of hidden moldy features in all abnormal areas is identified as the number of hidden dangers.

[0174] Determine whether the number of potential hazards exceeds the set value;

[0175] If so, then generate a shutdown command and the first prompt message;

[0176] Otherwise, the irradiation sequence and irradiation duration for each abnormal region are determined based on the abnormal light band corresponding to each abnormal region.

[0177] The corresponding target ray tracing command is obtained based on the irradiation sequence, irradiation duration, and corresponding ray tracing command.

[0178] The target tracking device is controlled by the target ray tracing command to track all abnormal areas in real time.

[0179] Preferably, in this embodiment, the duration of the first preset time period is 30 seconds or 50 seconds. The specific duration is not specifically limited in this embodiment. If multiple abnormal areas are detected within the first preset time period, that is, multiple abnormal areas need to be ray-tracked by the same target tracking device within the first preset time period, when the number of potential hazards is not higher than the preset number, the degree of moldiness of the moldy tobacco leaves in each abnormal area can be further analyzed, and the target tracking device can be controlled to perform ray tracking on each abnormal area in turn based on the analysis results. When the number of potential hazards is higher than the preset number, the conveyor belt can be stopped by generating a stop command, so as to allow more time for relevant personnel to remove the moldy tobacco leaves. The specific preset number is not specifically limited in this embodiment and can be set by relevant personnel according to actual needs.

[0180] The preferred method for obtaining the irradiation sequence of abnormal areas is as follows:

[0181] The light wave difference is based on the difference between the abnormal light band and the preset standard light band corresponding to each abnormal region.

[0182] Sort all the light wave differences in descending order to obtain the light wave difference sequence;

[0183] The abnormal regions corresponding to each light wave difference are sorted based on the light wave difference sequence to obtain the irradiation order corresponding to each abnormal region.

[0184] Preferably, the abnormal light bands corresponding to the abnormal areas can reflect the degree of mold on the moldy tobacco leaves. Therefore, the irradiation sequence and irradiation duration corresponding to each abnormal area can be determined based on the light wave difference between the abnormal light band and the preset standard light band. The light wave differences corresponding to each abnormal area are sorted in descending order to obtain a light wave difference sequence. Based on the light wave difference sequence, the abnormal areas corresponding to each light wave difference are sorted to determine the irradiation sequence corresponding to each abnormal area.

[0185] Preferably, based on a preset dwell time mapping relationship, the illumination dwell time corresponding to each abnormal area is determined. The preset dwell time mapping relationship is the correspondence between the light wave difference and the illumination dwell time. The specific content of the preset dwell time mapping relationship is not limited in this application embodiment, and can be determined by relevant personnel based on historical experimental data and then uploaded to the recognition system.

[0186] Preferably, multiple abnormal areas are irradiated in turn based on the irradiation sequence and irradiation dwell time. During the irradiation process, the irradiation needs to be carried out according to the ray visibility corresponding to each abnormal area. This optimization strategy helps to ensure that the tracking resources of the target tracking device are reasonably allocated, thereby improving the tracking efficiency.

[0187] Preferred options also include:

[0188] Integrate multiple mold characteristics identified within a second preset time period;

[0189] Based on the comparison of each mold growth characteristic, the mold growth similarity is obtained;

[0190] Determine whether the similarity is greater than the preset similarity based on the similarity of mold growth.

[0191] If so, then the corresponding abnormal conveyor belt area and conveyor belt conveying speed are determined based on the abnormal area corresponding to each moldy feature;

[0192] Determine the timing of tobacco leaf feeding based on abnormal conveyor belt areas and conveyor belt speed;

[0193] The target tobacco raw materials are traceable based on the moment of tobacco leaf feeding;

[0194] Based on the target tobacco leaf raw material, an inspection instruction and corresponding secondary prompt information are generated;

[0195] Safety inspections of target tobacco raw materials are conducted based on inspection instructions and second prompts.

[0196] Preferably, in this embodiment, the duration of the second preset time period is set to 3 minutes or 5 minutes. The specific duration is not specifically limited in this embodiment. By comparing each mold growth feature identified within the preset time period, the similarity between multiple mold growth features is determined, thereby facilitating the tracing of the cause of mold growth in tobacco leaves. Mold growth features can include mold texture, mold color, etc. By integrating multiple mold growth features to form a feature set, a preset similarity measurement method is used to compare the similarity of all mold growth features. The preset similarity measurement method can be cosine similarity, Euclidean distance, Manhattan distance, etc. The specific measurement method is not specifically limited in this embodiment.

[0197] The preferred method for determining the timing of tobacco leaf feeding is as follows:

[0198] The conveying distance of moldy tobacco leaves is determined based on the abnormal conveyor belt area corresponding to each moldy feature;

[0199] The time when moldy tobacco leaves arrive at the conveyor belt is determined by the ratio of the conveying distance to the conveyor belt speed.

[0200] Preferably, when the mold similarity is higher than a preset similarity, it indicates that the mold features identified at different times within the second preset time period have a high degree of similarity. The feeding time of moldy tobacco leaves with different mold characteristics is determined by the abnormal conveyor belt area and conveyor belt speed for each mold feature, and the source of the tobacco raw materials is traced based on the feeding time of each moldy tobacco leaf. For any mold feature, the conveying distance corresponding to the moldy tobacco leaf can be determined based on the abnormal conveyor belt area. Then, by dividing the conveyor belt speed and the conveying distance, the feeding time of the moldy tobacco leaf, i.e., the time when the moldy tobacco leaf arrives at the conveyor belt, can be obtained. Based on the above method, the feeding time corresponding to each mold feature can be determined.

[0201] Preferably, since the production and processing of tobacco leaves is a continuous operation, including breaking up tobacco leaves, conveying tobacco leaves, seasoning, shredding, rolling, and packaging, and each operation has an operation log, multiple tobacco leaf raw materials with mold characteristics can be traced based on the operation log and each feeding time. Through targeted inspections, potential mold risks can be discovered and dealt with in a timely manner, reducing the probability of mold growth in tobacco leaves from the source.

[0202] Based on the feedback from the inspection instructions generated for the tobacco raw materials to the terminal devices of the relevant inspection personnel, the system reminds the relevant inspection personnel to monitor the storage environment temperature and humidity of the tobacco raw materials in order to promptly detect and deal with potential mold risks.

[0203] Example 3

[0204] like Figure 3 As shown, a tobacco leaf surface mold identification and tracking system includes: a potential mold identification module, a tobacco leaf occlusion discrimination module, an occlusion tobacco leaf adjustment module, an adjustment image recognition module, and an abnormal area tracking module;

[0205] The hidden danger mold identification module is used to acquire images of tobacco leaves on the surface of the conveyor belt and identify hidden danger mold features to obtain a hidden danger image; based on the hidden danger image, the hidden danger light band is identified to obtain the screened hidden danger area;

[0206] The tobacco leaf obstruction detection module is used to determine whether there is tobacco leaf obstruction based on the screened potential hazard area; if not, it performs abnormal light band identification based on the screened potential hazard area to obtain the first abnormal area; if so, it determines the shaking range based on the tobacco leaf obstruction position and its corresponding potential hazard light band.

[0207] The obscured tobacco leaf adjustment module is used to adjust the obscured tobacco leaves in the screened hazard area based on the shaking range and then collect the adjusted image;

[0208] The adjusted image recognition module is used to identify abnormal light bands based on the adjusted image to obtain a second abnormal region;

[0209] The abnormal region tracking module is used to obtain corresponding ray tracing instructions and abnormal warning information based on the first abnormal region or the second abnormal region; and to track the corresponding abnormal region based on the ray tracing instructions.

[0210] Preferably, in this embodiment, the functional implementation process of each functional module corresponds one-to-one with the above method content, and will not be described in detail here.

[0211] Example 4

[0212] Based on the same inventive concept, the present invention also provides a computer device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0213] Memory, used to store computer programs;

[0214] When the processor executes a program stored in the memory, it is able to implement a method for identifying and tracking mold growth on the surface of tobacco leaves, as described in Example 1 or 2.

[0215] The electronic device may include a processor, a communications interface, a memory, and a communication bus, wherein the processor, communications interface, and memory communicate with each other via the communication bus. The processor can invoke logical instructions in the memory to execute a method for identifying and tracking mold growth on the surface of tobacco leaves, as described in Embodiment 1 or 2.

[0216] Furthermore, when the logical instructions in the aforementioned memory can be implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0217] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0218] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for identifying and tracking mold growth on the surface of tobacco leaves, characterized in that, include: Images of tobacco leaves on the conveyor belt surface were acquired and their potential mold and mildew characteristics were identified to obtain potential hazard images; Based on the hazard image, hazard light band identification is performed to obtain the screened hazard areas; Based on the screening of potential hazard areas, it is determined whether there is tobacco leaf obstruction; If not, then based on the screened potential hazard area, abnormal light band identification is performed to obtain the first abnormal area; If so, the shaking range is determined based on the location of the tobacco leaf obstruction and the corresponding potential hazard light band; After adjusting the obscured tobacco leaves in the screened potential hazard area based on the shaking range, an adjusted image was acquired. Based on the adjusted image, abnormal light bands are identified to obtain a second abnormal region; Based on the first or second abnormal region, corresponding ray tracing instructions and abnormal warning information are obtained; The ray tracing command is used to track the corresponding abnormal area.

2. The method for identifying and tracking mold growth on the surface of tobacco leaves according to claim 1, characterized in that, The identified potential hazard areas include: Based on the image of the potential hazard, identify the affected moldy area containing the moldy characteristics of the potential hazard; Based on the aforementioned areas with potential mold growth, obtain the corresponding regional spectral data; Based on the regional spectral data, the hidden danger light band is identified to obtain the spectral data of the screening region containing the hidden danger light band; The potential moldy areas corresponding to the spectral data of the selected areas are used as the selected potential hazard areas.

3. The method for identifying and tracking mold growth on the surface of tobacco leaves according to claim 1, characterized in that, Determine the range of jitter, specifically including: Based on the identified potential hazard areas, the occlusion contour information corresponding to each tobacco leaf occlusion location is identified; The amount of tobacco leaves accumulated at the corresponding tobacco leaf occlusion position is determined based on the occlusion contour information. The corresponding flat area is determined based on the amount of tobacco leaves piled up. The initial shaking range is obtained based on the position of the tobacco leaf obstruction and its corresponding flat area; The corresponding expanded area is obtained based on the hidden danger light band corresponding to the position of the tobacco leaf obstruction; The initial jitter range is optimized based on the expanded area to obtain the final jitter range.

4. The method for identifying and tracking mold growth on the surface of tobacco leaves according to claim 3, characterized in that, The method for obtaining the initial jitter range is as follows: The target distance is the straight-line distance from the highest edge point to the conveyor belt surface in the three-dimensional edge information corresponding to the tobacco leaf blocking position. The area difference is obtained based on the mapping relationship between the target distance and the preset area difference. The flat area corresponding to the tobacco leaf shading position is obtained by adding the area difference and the flat area; Identify the corresponding tiling edge information based on the tiling range; The initial jitter range is obtained by integrating all the tiling ranges based on the information of each tiling edge.

5. The method for identifying and tracking mold growth on the surface of tobacco leaves according to claim 1, characterized in that, The method for generating ray tracing commands is as follows: The abnormal region is defined as either the first abnormal region or the second abnormal region. Based on the abnormal region, the corresponding abnormal optical band is obtained; The visibility of the tracking ray is determined based on the anomalous light band. The real-time movement position of the abnormal area is identified based on the tobacco leaf image on the conveyor belt surface; The ray tracing command is generated based on the ray visibility and the real-time moving position.

6. The method for identifying and tracking mold growth on the surface of tobacco leaves according to claim 1, characterized in that, Also includes: When multiple abnormal areas exist within a first preset time period, the number of the potential moldy features in all the abnormal areas is identified as the number of potential hazards. Based on the number of potential hazards, determine whether it exceeds a set value; If so, then generate a shutdown command and the first prompt message; Otherwise, the irradiation sequence and irradiation duration of each abnormal region are determined based on the abnormal light band corresponding to each abnormal region; Based on the irradiation sequence, the irradiation duration, and the corresponding ray tracing command, a corresponding target ray tracing command is obtained; Based on the target ray tracing command, the target tracking device is controlled to perform real-time tracking of all the abnormal areas.

7. The method for identifying and tracking mold growth on the surface of tobacco leaves according to claim 6, characterized in that, The method for obtaining the irradiation sequence of the abnormal region is as follows: The light difference is based on the difference between the abnormal light band and the preset standard light band corresponding to each abnormal region. Sort all the light wave differences in descending order to obtain the light wave difference sequence; Based on the light wave difference sequence, the abnormal regions corresponding to each light wave difference are sorted to obtain the irradiation order corresponding to each abnormal region.

8. The method for identifying and tracking mold growth on the surface of tobacco leaves according to claim 1, characterized in that, Also includes: Integrate multiple mold characteristics identified within a second preset time period; Based on the comparison of each of the aforementioned mold characteristics, the mold similarity is obtained; Based on the mold similarity, determine whether it is greater than a preset similarity; If so, then the corresponding abnormal conveyor belt area and conveyor belt conveying rate are determined based on the abnormal area corresponding to each of the moldy characteristics; The timing of tobacco leaf feeding is determined based on the abnormal conveyor belt area and the conveyor belt speed. The target tobacco raw material can be traced based on the time of tobacco leaf feeding. Based on the target tobacco raw material, an inspection instruction and a corresponding second prompt message are generated; The target tobacco raw material is subjected to a safety inspection based on the inspection command and the second prompt information.

9. The method for identifying and tracking mold growth on the surface of tobacco leaves according to claim 8, characterized in that, The method for determining the timing of tobacco leaf feeding is as follows: The conveying distance of the moldy tobacco leaves is determined based on the abnormal conveyor belt area corresponding to each of the moldy characteristics. The time when the moldy tobacco leaves arrive at the conveyor belt is determined by the ratio of the conveying distance to the conveyor belt speed, and this time is taken as the unloading time of the tobacco leaves.

10. A tobacco leaf surface mold identification and tracking system, used to execute the tobacco leaf surface mold identification and tracking method as described in any one of claims 1-9, characterized in that, include: The module includes a hidden danger and mold identification module, a tobacco leaf occlusion detection module, an occlusion tobacco leaf adjustment module, an adjustment image recognition module, and an abnormal area tracking module. The hidden mold identification module is used to acquire images of tobacco leaves on the surface of the conveyor belt and identify hidden mold features to obtain a hidden mold image; Based on the hazard image, hazard light band identification is performed to obtain the screened hazard areas; The tobacco leaf obstruction detection module is used to determine whether there is tobacco leaf obstruction based on the screened potential hazard area; If not, then based on the screened potential hazard area, abnormal light band identification is performed to obtain the first abnormal area; If so, the shaking range is determined based on the location of the tobacco leaf obstruction and the corresponding potential hazard light band; The obscured tobacco leaf adjustment module is used to adjust the obscured tobacco leaves in the screened hazard area based on the shaking range and then collect the adjusted image; The adjusted image recognition module is used to identify abnormal light bands based on the adjusted image to obtain a second abnormal region; The abnormal region tracking module is used to obtain corresponding ray tracing instructions and abnormal warning information based on the first abnormal region or the second abnormal region. The ray tracing command is used to track the corresponding abnormal area.

Images