Intelligent monitoring method and system for quality of feed packaging bag

Abstract

This invention relates to the field of packaging bag quality monitoring technology, specifically to an intelligent monitoring method and system for feed packaging bags. The method includes: collecting side structure data of feed packaging bags; constructing a side structure map of the feed packaging bags using a structural boundary topology modeling algorithm based on the spatial distribution relationship of the bag arrangement; identifying the exposed contact areas of bundled plastic woven bags; analyzing the relative motion direction of the exposed contact areas using the attitude trajectory data of the feed packaging bags to generate potential friction risk zones; collecting structural surface feature data; constructing a multi-dimensional feature residual anomaly detection function to identify abnormal areas within the potential friction risk zones; and outputting the mapping relationship between abnormal areas and feed packaging bags based on boundary overlap indices and spatial projection intersection rules to generate a structured quality monitoring record of the feed packaging bags. This invention achieves intelligent identification and structured monitoring of potential friction anomaly areas in bundled feed packaging bags after transportation.

Description

Technical Field

[0001] This invention relates to the field of packaging bag quality monitoring technology, specifically to an intelligent monitoring method and system for feed packaging bag quality. Background Technology

[0002] With the rapid growth of packaging and transportation scale in the feed industry, bundled plastic woven bags frequently experience quality problems such as surface wear, bag damage, and fading labels during logistics, seriously affecting product delivery integrity and warehousing management efficiency. Packaging bags typically use flexible composite material structures, and the bag body is constructed through processes such as sewing and heat pressing. They are then handled and stacked in bundles during stacking and transportation. In actual warehousing and logistics, bundled feed packaging bags are easily subjected to external handling, forklift collisions, vehicle vibrations, and shelf compression, which can easily lead to misalignment of the side bag structure, curling of corners, or partial exposure of the bag surface. In particular, bags on the upper layer or edge of the stack are more likely to be exposed to frequent physical contact.

[0003] During actual transportation and stacking, feed packaging bags often have some edges or corners exposed for a long time due to irregular stacking methods, accumulated transportation vibrations, and complex posture changes, forming potentially high-risk friction areas. These exposed areas may experience abnormal problems such as boundary erosion, surface depressions, or local deformation due to repeated contact with adjacent bags, the ground, or transportation equipment during transportation. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent monitoring method and system for the quality of feed packaging bags, so as to solve the abnormal problems mentioned in the background art, such as boundary erosion, surface depression or local deformation caused by repeated contact with adjacent bags, the ground or transportation devices.

[0005] To achieve the above objectives, the technical solution of the present invention is: a method for intelligent monitoring of feed packaging bag quality, comprising:

[0006] S1. After the feed packaging bags are transported, collect the side structure data of the feed packaging bags. Based on the spatial distribution relationship of the bag arrangement, use the structural boundary topology modeling algorithm to construct the side structure map of the feed packaging bags.

[0007] S2. Identify the exposed contact area of ​​bundled plastic woven bags based on the side structure map of feed packaging bags, and analyze the relative motion direction of the exposed contact area by combining the attitude trajectory data of feed packaging bags to generate potential friction risk areas.

[0008] S3. Collect structural surface feature data in the potential friction risk zone and construct a multi-dimensional feature residual anomaly detection function to identify abnormal areas in the potential friction risk zone;

[0009] S4. Map the abnormal area to the side structure map of the feed packaging bag. Based on the boundary overlap index and spatial projection intersection rule, output the mapping relationship between the abnormal area and the feed packaging bag, and generate a structured quality monitoring record of the feed packaging bag.

[0010] Preferably, in step S1, the structural boundary topology modeling algorithm is a three-dimensional topology modeling algorithm based on the boundary coordinate data of the feed packaging bag and the stacking level index relationship, used to construct the side structure map of the feed packaging bag. The specific method is as follows:

[0011] Based on the side structure data of feed packaging bags, the boundary coordinates of the feed packaging bags are extracted to form boundary coordinate data; a stacking hierarchy index relationship is established according to the arrangement order of the feed packaging bags in the bundled state; a topological connection operation is performed based on the boundary coordinate data and the stacking hierarchy index relationship to generate a three-dimensional spatial topological structure; in the three-dimensional spatial topological structure, nodes are defined as boundary feature points of feed packaging bags, and edges are the contact boundaries between adjacent bags; duplicate and non-contact nodes are deleted to generate a side structure map of feed packaging bags.

[0012] Preferably, in step S2, the process of identifying the exposed contact area of ​​bundled plastic woven bags based on the side structure map of feed packaging bags includes:

[0013] The spatial exposure degree between nodes in the side structure diagram of feed packaging bags is calculated, and the exposed node set is determined based on the minimum distance from the node to the outer boundary surface and the visible projection ratio. The exposed node set is aggregated according to the spatial adjacency relationship to form the exposed contact area. Within the exposed contact area, the attitude vector difference and relative displacement component between adjacent bags are calculated by combining the attitude trajectory data of feed packaging bags. The relative motion direction of the contact area in the bundled structure is analyzed based on the pose change rate. Based on the relative motion direction and contact duration within the exposed contact area, if the frequency of change of the relative motion direction and the contact duration both exceed the preset threshold, the exposed contact area is marked as a potential friction risk area.

[0014] Preferably, in S3, the structural surface feature data refers to the set of multimodal feature data collected on the surface of the feed packaging bag corresponding to the potential friction risk zone;

[0015] The multidimensional feature residual anomaly detection function is a joint residual evaluation function constructed based on structural surface feature data. Specifically, it performs standard feature model fitting on the multidimensional surface feature vectors in the potential friction risk zone, calculates the residual value of each feature dimension within the sliding detection window, compares the residual value with the residual value of the historical stable region, and identifies abnormal regions where the residual value abruptly surges.

[0016] Preferably, in step S3, the method for identifying abnormal areas within the potential friction risk zone is as follows:

[0017] Based on the potential friction risk area, the surface feature data of the structure is obtained and input into the multidimensional feature residual anomaly detection function for residual analysis; the texture gradient residual, reflectivity change residual, concavity and convexity height difference residual and color level residual are calculated and mapped to the corresponding nodes of the side structure map of the feed packaging bag to form a residual response heat map; an anomaly identification heat threshold is set, and continuous spatial sub-regions with residual heat exceeding the anomaly identification heat threshold are extracted and marked as abnormal regions.

[0018] Preferably, in S4, the boundary overlap index refers to a calculation parameter used to quantify the degree of spatial overlap between the abnormal region and a single feed packaging bag in the map structure;

[0019] The spatial projection intersection rule is a rule for determining the spatial correspondence between abnormal areas and packaging bag nodes;

[0020] The mapping relationship between the abnormal area and the feed packaging bag is based on the abnormal area. It is located in the side structure diagram of the feed packaging bag and bound to the actual physical part of the feed packaging bag, forming a two-way correspondence between the abnormality of the packaging bag and the specific structural position of the packaging bag.

[0021] On the other hand, the present invention provides an intelligent monitoring system for the quality of feed packaging bags, including a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement the aforementioned intelligent monitoring method for the quality of feed packaging bags.

[0022] Compared with the prior art, the above-mentioned technical solution of the present invention has the following beneficial technical effects:

[0023] 1. In this invention, a side structure map is constructed based on the actual stacking arrangement of feed packaging bags, which can more clearly identify the contact relationship and exposed areas between packaging bags, thereby helping to discover potential friction risks in the early stage and avoid structural wear of the entire batch of bags due to neglecting small area contact.

[0024] 2. In this invention, based on the fusion analysis mechanism of the side structure map and attitude trajectory data of feed packaging bags, potential damaged areas caused by exposure or friction during transportation can be quickly identified without disassembling the bundled packaging bags, thereby realizing intelligent monitoring of the overall quality of the bundled packaging bags. Attached Figure Description

[0025] Figure 1 This is a flowchart of one embodiment of the present invention. Detailed Implementation

[0026] Example 1, as Figure 1As shown, the specific implementation steps of the intelligent monitoring method for feed packaging bag quality proposed in this invention are as follows:

[0027] S1. After the feed packaging bags are transported, collect the side structure data of the feed packaging bags. Based on the spatial distribution relationship of the bag arrangement, use the structural boundary topology modeling algorithm to construct the side structure map of the feed packaging bags.

[0028] S2. Identify the exposed contact area of ​​bundled plastic woven bags based on the side structure map of feed packaging bags, and analyze the relative motion direction of the exposed contact area by combining the attitude trajectory data of feed packaging bags to generate potential friction risk areas.

[0029] S3. Collect structural surface feature data in the potential friction risk zone and construct a multi-dimensional feature residual anomaly detection function to identify abnormal areas in the potential friction risk zone;

[0030] S4. Map the abnormal area to the side structure map of the feed packaging bag. Based on the boundary overlap index and spatial projection intersection rule, output the mapping relationship between the abnormal area and the feed packaging bag, and generate a structured quality monitoring record of the feed packaging bag.

[0031] In this embodiment, in step S1, the structural boundary topology modeling algorithm is a three-dimensional topology modeling algorithm based on the boundary coordinate data of the feed packaging bag and the stacking level index relationship. It is used to construct the side structure map of the feed packaging bag. The specific method is as follows:

[0032] Based on the side structure data of feed packaging bags, the boundary coordinates of the feed packaging bags are extracted to form boundary coordinate data; a stacking hierarchy index relationship is established according to the arrangement order of the feed packaging bags in the bundled state; a topological connection operation is performed based on the boundary coordinate data and the stacking hierarchy index relationship to generate a three-dimensional spatial topological structure; in the three-dimensional spatial topological structure, nodes are defined as boundary feature points of feed packaging bags, and edges are the contact boundaries between adjacent bags; duplicate and non-contact nodes are deleted to generate a side structure map of feed packaging bags.

[0033] In this embodiment, the spatial distribution relationship of bag arrangement refers to the set of information on the arrangement position and relative posture of each individual bag in three-dimensional space relative to other bags when the feed packaging bags are in a bundled state. This relationship is used to characterize the stacking form and local spatial structure features of the bundled packaging bags during actual transportation or storage. The process of obtaining the spatial distribution relationship of bag arrangement includes the following steps: acquiring the boundary point cloud data of each packaging bag in the bundled state based on sensors or vision acquisition modules, generating the boundary coordinate set of each packaging bag through envelope contour recognition and edge point fitting algorithms; constructing a stacking level index matrix between bags based on parameters such as the spatial contact relationship, overlap height level, and contact strength of the contact surface between the boundary coordinate sets, which is used to represent the vertical and horizontal contact arrangement between bags; fusing and mapping the boundary coordinate set and the stacking index relationship to a unified three-dimensional coordinate system; extracting local stacking density parameters, contact posture angle parameters, and boundary distortion parameters, which are the spatial distribution relationship of bag arrangement.

[0034] In this embodiment, after the feed packaging bags have completed the transportation process, their side structure data is acquired using a two-dimensional or three-dimensional acquisition device for their static state after being bundled and stacked. The side structure data includes the boundary outline, corner point position, and relative arrangement order information of each packaging bag. The data can be acquired through a visual imaging system, laser scanning equipment, or structured light measurement module. After acquisition, the boundary data is preprocessed, including noise reduction, edge enhancement, and morphological correction, to improve the accuracy of boundary coordinate extraction.

[0035] Based on the preprocessed boundary image data, a boundary feature point extraction operation is performed to obtain the boundary coordinate data of each packaging bag, which represents the set of coordinate points of the packaging bag's perceptible boundary in space. At the same time, according to the actual stacking order of the packaging bags and the relative height hierarchy formed during the transmission process, a stacking hierarchy index relationship is constructed. This index relationship is used to clarify the vertical or horizontal hierarchy number of each packaging bag in the bundled structure and its possibility of contact with adjacent bags.

[0036] Subsequently, the structural boundary topology modeling algorithm is invoked, taking the boundary coordinate data and stacking hierarchy index relationship as input, and performing a three-dimensional topology connection operation: constructing connection edges between nodes through spatial projection analysis, defining topology nodes as boundary feature points, and connection edges representing boundary segments that have contact or proximity relationships in physical space; and using a topology filtering mechanism to remove duplicate nodes (such as the same boundary being identified multiple times) and non-contact nodes (boundary points whose spatial distance exceeds the contact threshold), ensuring that the topology structure retains only physically valid connections.

[0037] The final output is a side structure diagram of the feed packaging bag. This diagram is a graph structure representation that describes the spatial arrangement and boundary topology of the packaging bags. Nodes represent boundary points, and edges represent the actual or potential contact boundaries between adjacent bags.

[0038] Side structure data of feed packaging bags refers to the set of data obtained from the outer side of feed packaging bags after bundled transportation, which describes the arrangement and boundary configuration of the packaging bags. Specifically, it includes the following data types: boundary contour image data, feed packaging bag boundary coordinates, contour structure connectivity information, bag arrangement index information, and initial criteria for contact status between adjacent bags.

[0039] In this embodiment, step S2, which involves identifying the exposed contact area of ​​bundled plastic woven bags based on the side structure map of feed packaging bags, includes:

[0040] The spatial exposure degree between nodes in the side structure diagram of feed packaging bags is calculated, and the exposed node set is determined based on the minimum distance from the node to the outer boundary surface and the visible projection ratio. The exposed node set is aggregated according to the spatial adjacency relationship to form the exposed contact area. Within the exposed contact area, the attitude vector difference and relative displacement component between adjacent bags are calculated by combining the attitude trajectory data of feed packaging bags. The relative motion direction of the contact area in the bundled structure is analyzed based on the pose change rate. Based on the relative motion direction and contact duration within the exposed contact area, if the frequency of change of the relative motion direction and the contact duration both exceed the preset threshold, the exposed contact area is marked as a potential friction risk area.

[0041] In this embodiment, the spatial exposure degree refers to the degree of spatial openness of a node in the side structure diagram of the feed packaging bag relative to the outer surface of the bundle. It is defined as the weighted function value of the minimum Euclidean distance from the node to the outer contour surface of the diagram and the visible projected area. In actual calculation, by constructing a three-dimensional boundary voxel model, the exposure ratio of each node under orthogonal projection is obtained, and it is normalized and combined with the corresponding boundary distance to form a node spatial exposure degree index. The exposed node set refers to the set of nodes with spatial exposure degrees greater than a preset exposure threshold, reflecting the boundary area in the bundled structure that is most likely to have direct contact with the external environment. Furthermore, by dividing... The spatial adjacency relationships between exposed nodes are analyzed, and continuous boundary region blocks are formed using graph clustering methods, which are defined as exposed contact areas. The attitude trajectory data is the spatial position and orientation change information of feed packaging bags during transportation. The acquisition methods include accelerometer and gyroscope data fusion modeling or high-frequency video 3D reconstruction, which are uniformly converted into attitude vector sequences in this embodiment. Within the exposed contact area, the attitude vector difference and spatial displacement change rate between adjacent packaging bags are extracted to construct a relative motion direction vector set. The attitude change rate and contact duration are calculated by combining the time series window to form a multi-dimensional contact dynamic index group.

[0042] In this embodiment, in S3, the structural surface feature data refers to the set of multimodal feature data collected on the surface of the feed packaging bag corresponding to the potential friction risk zone; the structural surface feature data includes, but is not limited to: texture gradient distribution data, surface reflectivity change data, boundary unevenness height difference data and surface color gradation change information, which are used to characterize the changes in the microscopic physical state of the feed packaging bag surface under different pressure and friction conditions;

[0043] The multidimensional feature residual anomaly detection function is a joint residual evaluation function constructed based on structural surface feature data. Specifically, it performs standard feature model fitting on the multidimensional surface feature vectors in the potential friction risk zone, calculates the residual value of each feature dimension within the sliding detection window, compares the residual value with the residual value of the historical stable region, and identifies abnormal regions where the residual value abruptly surges.

[0044] In this embodiment, the structural surface feature data is a dataset obtained from multimodal sensing of the feed packaging bag surface in the potential friction risk area. It reflects the microscopic physical changes of the bag surface under pressure, friction, or stress. Specifically, it includes the following data types: texture gradient distribution data, obtained by analyzing the grayscale gradient direction histogram of the surface texture pattern to extract the directional changes in surface texture after local friction, used to determine whether local disturbances have occurred in the surface fiber arrangement; surface reflectivity change data, measured using a linear light source or structured light reflection model to measure local reflectivity changes, reflecting the changing pattern of surface optical characteristics with friction damage; boundary unevenness height difference data, calculated using a laser displacement sensor or structured light scanning to calculate the surface height matrix and determine the amount of local unevenness changes to characterize surface indentations and protrusions; and surface color gradation change information, captured by a multispectral imaging device to capture color gradation changes in different bands, used to determine the micro-color changes produced by friction on the plastic surface.

[0045] In this embodiment, the method for identifying abnormal regions in the potential friction risk zone in step S3 is as follows:

[0046] Based on the potential friction risk area, the surface feature data of the structure is obtained and input into the multidimensional feature residual anomaly detection function for residual analysis; the texture gradient residual, reflectivity change residual, concavity and convexity height difference residual and color level residual are calculated and mapped to the corresponding nodes of the side structure map of the feed packaging bag to form a residual response heat map; an anomaly identification heat threshold is set, and continuous spatial sub-regions with residual heat exceeding the anomaly identification heat threshold are extracted and marked as abnormal regions.

[0047] In this embodiment, the texture gradient residual is used to characterize the deviation between the local texture direction change and the standard texture model; the reflectivity change residual is used to represent the difference between the actual reflectivity and the light reflection intensity under normal conditions; the unevenness height difference residual is used to reflect the abnormal degree of the change in surface height after deformation under force; the color level residual is used to evaluate the extent to which the image color level value deviates from the historical normal state; the above residual values ​​are all quantitative difference indicators obtained by comparing with the standard baseline, reflecting the response intensity of the surface microstructure under the influence of friction.

[0048] In this embodiment, in S4, the boundary overlap index refers to a calculation parameter used to quantify the degree of spatial overlap between the abnormal area and a single feed bag in the map structure; the boundary overlap index is specifically: in the three-dimensional structure map, the boundary contour of the abnormal area is projected onto the standard projection plane of the side structure map of the feed bag, and the overlap area between the projected contour and the spatial node area of ​​each single packaging bag is calculated.

[0049] The spatial projection intersection rule is a rule for determining the spatial correspondence between abnormal areas and packaging bag nodes;

[0050] The spatial projection intersection rules include the following two types of judgment logic: if the boundary overlap exceeds the preset overlap threshold, the abnormal area is considered to belong to the packaging bag node; if the abnormal area overlaps with multiple packaging bag nodes, the maximum overlap assignment strategy is adopted.

[0051] The mapping relationship between the abnormal area and the feed packaging bag is based on the abnormal area. It is located in the side structure diagram of the feed packaging bag and bound to the actual physical part of the feed packaging bag, forming a two-way correspondence between the abnormality of the packaging bag and the specific structural position of the packaging bag.

[0052] In this embodiment, the mapping relationship between the abnormal area and the feed packaging bag is output. The specific method includes: numbering each abnormal area and recording its maximum residual heat value, spatial center position, and boundary contour; identifying the map structure node that intersects with the abnormal area according to the spatial projection intersection rule; and binding each structure node to the specific physical area of ​​the feed packaging bag according to its definition in the side structure map of the feed packaging bag.

[0053] The structured quality monitoring record for feed packaging bags is a formatted representation of the mapping relationship between abnormal areas, spatial locations, and abnormal index values ​​identified during the monitoring process and the corresponding bag body, and stored as a data record unit with a structured hierarchy. The record structure includes the following fields: bag body number information field, abnormal area spatial identifier field, abnormal type field, abnormal index field, and monitoring timestamp field.

[0054] Example 2: The present invention proposes an intelligent monitoring system for the quality of feed packaging bags, which is applied to the intelligent monitoring method for the quality of feed packaging bags proposed in Example 1. It includes a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement the intelligent monitoring method for the quality of feed packaging bags in Example 1.

[0055] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited thereto. Various changes can be made within the scope of knowledge possessed by those skilled in the art without departing from the spirit of the present invention.

Claims

1. A method for intelligent monitoring of feed packaging bag quality, characterized in that, Includes the following steps: S1. After the feed packaging bags are transported, collect the side structure data of the feed packaging bags. Based on the spatial distribution relationship of the bag arrangement, use the structural boundary topology modeling algorithm to construct the side structure map of the feed packaging bags. S2. Identify the exposed contact area of ​​bundled plastic woven bags based on the side structure map of feed packaging bags, and analyze the relative motion direction of the exposed contact area by combining the attitude trajectory data of feed packaging bags to generate potential friction risk areas. In step S2, the process of identifying the exposed contact area of ​​bundled plastic woven bags based on the side structure map of feed packaging bags includes: calculating the spatial exposure degree between each node in the side structure map of feed packaging bags, determining the exposed node set based on the minimum distance from the node to the outer boundary surface and the visible projection ratio; aggregating the exposed node set according to spatial adjacency to form the exposed contact area; within the exposed contact area, combining the attitude trajectory data of feed packaging bags, calculating the attitude vector difference and relative displacement components between adjacent bags, analyzing the relative motion direction of the contact area in the bundled structure based on the pose change rate; determining the relative motion direction and contact duration within the exposed contact area, and if the frequency of change of the relative motion direction and the contact duration both exceed a preset threshold, then the exposed contact area is marked as a potential friction risk area; S3. Collect structural surface feature data in the potential friction risk zone and construct a multi-dimensional feature residual anomaly detection function to identify abnormal areas in the potential friction risk zone; In S3, the structural surface feature data refers to the set of multimodal feature data collected on the surface of the feed packaging bag corresponding to the potential friction risk zone; the multidimensional feature residual anomaly detection function is a joint residual evaluation function constructed based on the structural surface feature data, specifically: performing standard feature model fitting on the multidimensional surface feature vector in the potential friction risk zone, calculating the residual value of each feature dimension within the sliding detection window, comparing the residual value with the residual value of the historical stable area, and identifying abnormal areas where the residual value abruptly increases; In step S3, the method for identifying abnormal regions in the potential friction risk zone is as follows: Based on the potential friction risk zone, extract the surface feature data of the structure and input it into the multidimensional feature residual abnormality detection function for residual analysis; calculate the texture gradient residual, reflectivity change residual, concavity-convexity height difference residual, and color gradation residual, and map them to the corresponding nodes of the side structure map of the feed packaging bag to form a residual response heat map; set an abnormality identification heat threshold, extract continuous spatial sub-regions whose residual heat exceeds the abnormality identification heat threshold, and mark them as abnormal regions; S4. Map the abnormal area to the side structure map of the feed packaging bag. Based on the boundary overlap index and spatial projection intersection rule, output the mapping relationship between the abnormal area and the feed packaging bag, and generate a structured quality monitoring record of the feed packaging bag.

2. The intelligent monitoring method for feed packaging bag quality according to claim 1, characterized in that: In S1, the structural boundary topology modeling algorithm is a three-dimensional topology modeling algorithm based on the boundary coordinate data of the feed packaging bag and the stacking level index relationship. It is used to construct the side structure map of the feed packaging bag. The specific method is as follows: Based on the side structure data of feed packaging bags, the boundary coordinates of the feed packaging bags are extracted to form boundary coordinate data; a stacking hierarchy index relationship is established according to the arrangement order of the feed packaging bags in the bundled state; a topological connection operation is performed based on the boundary coordinate data and the stacking hierarchy index relationship to generate a three-dimensional spatial topological structure; in the three-dimensional spatial topological structure, nodes are defined as boundary feature points of feed packaging bags, and edges are the contact boundaries between adjacent bags; duplicate and non-contact nodes are deleted to generate a side structure map of feed packaging bags.

3. The intelligent monitoring method for feed packaging bag quality according to claim 2, characterized in that: In S4, the boundary overlap index refers to a calculation parameter used to quantify the degree of spatial overlap between the abnormal region and a single feed packaging bag in the map structure. The spatial projection intersection rule is a rule for determining the spatial correspondence between abnormal areas and packaging bag nodes; The mapping relationship between the abnormal area and the feed packaging bag is based on the abnormal area. It is located in the side structure diagram of the feed packaging bag and bound to the actual physical part of the feed packaging bag, forming a two-way correspondence between the abnormality of the packaging bag and the specific structural position of the packaging bag.

4. A smart monitoring system for the quality of feed packaging bags, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: The processor executes a computer program to implement the intelligent monitoring method for feed packaging bag quality as described in any one of claims 1-3.

Images