A plastic bag quality detection system based on image recognition

Abstract

The application discloses a kind of based on image recognition's plastic bag quality detection system, it is related to quality detection technical field, including: image acquisition module, for obtaining original image sequence on production line;Strategy execution module, for according to dynamic strategy coefficient output final quality detection result;Result filing and feedback module, for the final quality detection result is aggregated generation and output for prompting operator to intervene early warning instruction;Core state perception module, for based on historical structured quality data, calculate system risk index;Dynamic decision and adjustment module, for system risk index is input to preset nonlinear adjustment function, to generate dynamic strategy coefficient.The application is adjusted detection model by dynamic strategy coefficient, and the quality trend of risk index perception is fused, and closed-loop control architecture is constructed, speed and precision collaborative optimization is realized, system oscillation is avoided, quality risk is early warned, and production efficiency and stability are improved.

Description

TECHNICAL FIELD

[0001] The application relates to the quality detection technical field, in particular to a plastic bag quality detection system based on image recognition. BACKGROUND

[0002] In the automatic production process of plastic bags, quality detection is an important link to guarantee the qualified rate of products. With the improvement of production line speed and quality standards, the detection system based on image recognition is widely used in real-time detection of quality indicators such as surface defects and size precision of plastic bags because of its non-contact and high efficiency. The technology realizes quality judgment through image acquisition by industrial cameras and other equipment combined with algorithm analysis, and can become an important means to replace manual sampling inspection.

[0003] The existing image recognition detection system has obvious defects. The traditional image recognition detection system mostly uses a fixed threshold detection model, which is difficult to achieve dynamic balance between detection speed and accuracy, and often causes system oscillation problems due to threshold switching. At the same time, it also lacks the ability to perceive the trend of quality changes and can only make judgments based on the current defect rate, which cannot early warn potential batch quality risks, leading to lag in production adjustment and affecting production efficiency and product quality stability. SUMMARY

[0004] The purpose of the application is to provide a plastic bag quality detection system and method based on image recognition, which solves the problems in the background art.

[0005] To solve the above technical problems, the application provides a plastic bag quality detection system based on image recognition, which comprises an image acquisition module for acquiring an original image sequence on a production line.

[0006] A strategy execution module is used to adaptively select a detection model from a preset speed priority model and an accuracy priority model according to a dynamic strategy coefficient. The strategy execution module is also used to adopt the detection model to process the original image sequence to output a final quality detection result.

[0007] A core state perception module is used to calculate a recent comprehensive defect rate and a change rate of the recent comprehensive defect rate based on the historical structured quality data. The core state perception module is also used to calculate a system risk index by combining the recent comprehensive defect rate and the change rate.

[0008] a result archiving and feedback module configured to aggregate the final quality inspection results to generate historical structured quality data recorded by inspection period, and configured to compare the system risk index with a preset early warning threshold, and generate and output an early warning instruction prompting an operator to intervene when the system risk index is higher than the early warning threshold for a preset number of consecutive inspection periods.

[0009] a dynamic decision and adjustment module configured to input the system risk index into a preset nonlinear adjustment function to generate the dynamic strategy coefficient;

[0010] Preferably, the image acquisition module is configured to capture the original image sequence by an industrial camera and a linear array sensor at an adjustable frame rate and resolution.

[0011] Preferably, the core state perception module is specifically configured to:

[0012] sum the total number of final defects in the historical structured quality data in a preset sliding time window, and sum the total number of final inspections;

[0013] The core state perception module is further configured to divide the sum of the total number of final defects by the sum of the total number of final inspections to calculate the recent comprehensive defect rate.

[0014] Preferably, the core state perception module is specifically configured to:

[0015] multiply the recent comprehensive defect rate by a preset defect rate weight to obtain a weighted defect rate;

[0016] multiply the change rate by a preset change rate weight to obtain a weighted change rate;

[0017] and add the weighted defect rate and the weighted change rate to calculate the system risk index.

[0018] Preferably, the nonlinear adjustment function is an S-shaped function.

[0019] The dynamic decision and adjustment module is configured to subtract a preset risk adjustment center point from the system risk index to obtain a risk deviation.

[0020] The dynamic decision and adjustment module is further configured to multiply the risk deviation by a preset adjustment curve steepness, and convert the result by the S-shaped function to generate the dynamic strategy coefficient.

[0021] Preferably, the strategy execution module is specifically configured to:

[0022] when the dynamic strategy coefficient is lower than a preset speed priority threshold, selecting the speed priority model as the detection model;

[0023] when the dynamic strategy coefficient is higher than a preset precision priority threshold, selecting the precision priority model as the detection model;

[0024] when the dynamic strategy coefficient is between the speed priority threshold and the precision priority threshold, adjusting the internal parameters of the speed priority model, the internal parameters of the precision priority model, or both, as the detection model, the internal parameters including a confidence threshold and an IOU threshold.

[0025] Preferably, the strategy execution module is further configured to generate a hardware adjustment instruction in response to the dynamic strategy coefficient.

[0026] The image acquisition module is further configured to receive the hardware adjustment instruction and dynamically adjust the frame rate and resolution of the original image sequence based on the hardware adjustment instruction.

[0027] Preferably, the historical structured quality data includes records in units of detection cycles.

[0028] Each of the records contains a final total number of detections and a final total number of defects corresponding to the detection cycle.

[0029] Compared with the prior art, the present application has the following beneficial effects:

[0030] 1. By dynamically adjusting the detection model through the dynamic strategy coefficient, a dynamic balance between detection speed and accuracy is achieved. When the risk is low, the speed priority model is automatically adopted to improve the detection speed; when the risk increases, the precision priority model is automatically switched to improve the detection accuracy, avoiding the performance fluctuations caused by the traditional fixed threshold method and solving the contradiction between speed and accuracy.

[0031] 2. By designing a risk index that fuses defect rates and their change rates, the quality change trend can be perceived in real time. By converting the strategy through an S-shaped nonlinear adjustment function, system oscillation is fundamentally avoided, and the response sensitivity of the detection system to production quality fluctuations is improved.

[0032] 3. By constructing a "perception-decision-execution" closed-loop control architecture, the detection strategy can be continuously optimized according to historical structured quality data. From the dynamic adjustment of image acquisition parameters to the adaptive selection of detection models, a data-driven intelligent adjustment mechanism is formed, realizing the transition from passive detection to active adaptation and providing technical support for early prevention of batch quality problems and stable production efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0033] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description only show some of the embodiments of the present application, and all other drawings obtained by those of ordinary skill in the art without creative effort based on these drawings also belong to the protection scope of the present application.

[0034] Figure 1 The logic block diagram of the system of the present application. DETAILED DESCRIPTION

[0035] The technical solutions in the embodiments of the present application will be clearly and completely described in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort also belong to the protection scope of the present application.

[0036] Embodiment 1:

[0037] Please refer to Figure 1 The present application provides a plastic bag quality detection system based on image recognition, comprising: an image acquisition module for acquiring an original image sequence on a production line;

[0038] A strategy execution module is used for adaptively selecting a detection model from a preset speed priority model and a precision priority model according to a dynamic strategy coefficient; the strategy execution module is also used for processing the original image sequence by using the detection model to output a final quality detection result;

[0039] A core state perception module is used for calculating a recent comprehensive defect rate and a change rate of the recent comprehensive defect rate based on the historical structured quality data; the core state perception module is also used for calculating a system risk index by combining the recent comprehensive defect rate and the change rate;

[0040] A result archiving and feedback module is used for aggregating the final quality detection result to generate historical structured quality data recorded by detection cycles; the result archiving and feedback module is also used for comparing the system risk index with a preset warning threshold, and generating and outputting a warning instruction for prompting an operator to intervene when the system risk index is higher than the warning threshold in a preset number of consecutive detection cycles;

[0041] A dynamic decision and adjustment module is used for inputting the system risk index into a preset nonlinear adjustment function to generate the dynamic strategy coefficient;

[0042] The embodiment provides a modular solution, through the close cooperation of the image acquisition module, the policy execution module, the result archiving and feedback module, the core state perception module and the dynamic decision and adjustment module, a complete closed-loop control system from data acquisition, state perception, dynamic decision to policy execution is constructed; the design of the system architecture aims to realize dynamic adaptive detection of the quality of the plastic bag production, actively perceive the quality change trend and adjust the detection strategy in real time, fundamentally solve the contradiction between speed and accuracy which is difficult to reconcile in the traditional detection method, and quickly respond to the quality problem trend, thereby improving the intelligentization and automation level of the whole production process.

[0043] Embodiment 2:

[0044] The image acquisition module is used to capture the original image sequence through an industrial camera and a linear array sensor at an adjustable frame rate and resolution;

[0045] In the embodiment, the image acquisition module can capture the high-speed flowing plastic bags on the production line at an adjustable frame rate and resolution to generate an original high-frame-rate image sequence by configuring the industrial camera and the linear array sensor; such configuration ensures that the original data containing the complete surface details of the plastic bags can be obtained, which provides a high-quality input basis for subsequent accurate detection, and the adjustability of the parameters is the hardware basis for realizing the adaptive detection strategy of the system, which responds to the instructions of the downstream modules for dynamic adjustment, and ensures that whether speed or accuracy is pursued, the data acquisition link can provide optimal support.

[0046] Embodiment 3:

[0047] The core state perception module is specifically configured to sum the total number of final defects in the historical structured quality data in a preset sliding time window, and sum the total number of final detections; the core state perception module is further configured to divide the sum of the total number of final defects by the sum of the total number of final detections to calculate the recent comprehensive defect rate;

[0048] The core state perception module is specifically configured to multiply the recent comprehensive defect rate by a preset defect rate weight to obtain a weighted defect rate, multiply the change rate by a preset change rate weight to obtain a weighted change rate, and add the weighted defect rate and the weighted change rate to calculate the system risk index;

[0049] The core state perception module obtains historical structured quality data from the result archiving and feedback module and performs operation in a preset sliding time window T, to quantify the current production quality state; the module calculates the recent comprehensive defect rate and the change rate of the recent comprehensive defect rate , and then the system risk index is calculated ;

[0050] The recent comprehensive defect rate is calculated based on statistics of historical data within a sliding time window T, and the calculation formula is:

[0051]

[0052] is the recent comprehensive defect rate of the current detection period;

[0053] is the current detection period;

[0054] is the preset sliding time window size;

[0055] is the final defect number of the i-th detection period;

[0056] is the final total number of detections of the i-th detection period;

[0057] This calculation method can reflect the real-time state of the current production quality;

[0058] The system risk index is calculated by weighted summation of the recent comprehensive defect rate and its change rate , thereby combining the real-time state and the change trend of quality, and the calculation formula is:

[0059]

[0060] is the system risk index of the current detection period;

[0061] is the defect rate weight (preset parameter);

[0062] is the recent comprehensive defect rate of the current detection period;

[0063] is the change rate weight (preset parameter);

[0064] is the defect rate change rate;

[0065] is the prediction time constant, a pre-set engineering parameter with time dimension used to adjust the response sensitivity of the system to the quality change trend;

[0066] Change rate The approximate calculation is , where is fixed as 1 detection period; this comprehensive risk assessment approach enables the system to not only focus on the current defect level, but also to predictively capture the signs of quality deterioration or improvement, thus providing a deeper and more forward-looking basis for decision-making.

[0067] Embodiment 4:

[0068] The nonlinear adjustment function is an S-shaped function; the dynamic decision-making and adjustment module is configured to subtract a preset risk adjustment center point from the system risk index to obtain a risk deviation; the dynamic decision-making and adjustment module is further configured to multiply the risk deviation by a preset adjustment curve steepness and convert the result through the S-shaped function to generate the dynamic strategy coefficient;

[0069] The dynamic decision-making and adjustment module receives the system risk index calculated by the core state perception module and converts it into a smooth dynamic strategy coefficient in the range of 0 to 1 using a preset S-shaped nonlinear adjustment function ; the preset S-shaped function simulates expert decision-making logic and enables smooth and nonlinear strategy transition according to the risk level, thus avoiding system oscillation that may be caused by traditional threshold switching methods; the formula for this conversion process is:

[0070]

[0071] is the dynamic strategy coefficient (value range [0, 1]);

[0072] is the adjustment curve steepness (preset parameter, controls the degree of strategy transition);

[0073] is the system risk index;

[0074] is the risk adjustment center point;

[0075] Through this function, the system risk This quantitative indicator is mapped to a guiding strategy parameter , where is set by experts according to the desired degree of strategy transition aggressiveness, and is the risk balance point defined by the business, at which the system speed and accuracy are most balanced (corresponding to ), thus providing a core driving force for the smooth switching between the "speed first" and "accuracy first" strategies for downstream modules.

[0076] Example 5:

[0077] The strategy execution module is specifically used to: select the speed-priority model as the detection model when the dynamic strategy coefficient is lower than the preset speed-priority threshold; select the accuracy-priority model as the detection model when the dynamic strategy coefficient is higher than the preset accuracy-priority threshold; and when the dynamic strategy coefficient is between the speed-priority threshold and the accuracy-priority threshold, adaptively adjust the internal parameters of the speed-priority model, the internal parameters of the accuracy-priority model, or simultaneously adjust the internal parameters of the two models to serve as the detection model, wherein the internal parameters include a confidence threshold and an IOU threshold.

[0078] The strategy execution module is further configured to generate hardware adjustment instructions in response to the dynamic strategy coefficients; the image acquisition module is further configured to receive the hardware adjustment instructions and dynamically adjust the acquisition frame rate and acquisition resolution of the original image sequence based on the hardware adjustment instructions.

[0079] The strategy execution module is the final executor of the system's decisions; it receives the dynamic strategy coefficients generated by the dynamic decision-making and adjustment module. And translate it into specific software algorithms and hardware parameter configurations; at the software level, when When the value approaches 0, the module selects and applies a speed-prioritized lightweight detection model to maximize processing efficiency; when... When the value approaches 1, switch to a precision-prioritized deep learning model to ensure the highest detection accuracy; for values ​​between 1 and 1, switch to a model prioritizing accuracy. The module achieves seamless integration of the two strategies by smoothly adjusting key parameters within the model; at the hardware level, the module will also... Value mapping is used to specify instructions for the image acquisition module, when When the value increases (biased towards precision), it instructs the image acquisition module to moderately increase the acquisition frame rate. and resolution This provides richer data support for high-precision algorithms; conversely, it reduces these parameters to alleviate system load. This hardware and software coordinated adjustment mechanism enables the system to function as an organic whole, dynamically and precisely allocating computing resources according to real-time risks, thereby achieving optimal system energy efficiency while meeting quality control requirements.

[0080] Example 6:

[0081] Historical structured quality data includes records in units of inspection cycles; each record contains the final total number of inspections and the final total number of defects corresponding to the inspection cycle;

[0082] The result archiving and feedback module is the only generating source of system history data, responsible for real-time aggregation and archiving of each detection result output by the strategy execution module; this module will count the "final total number of detections" and the "final total number of defects" at the end of each detection cycle and store this set of data as a structured record in the history database; through this continuous recording process, the system builds a traceable, time-series-based structured quality database; these records are not only the basis for evaluating the long-term performance of the system, but more importantly, they constitute the data cornerstone of the decision feedback loop, providing the core state perception module upstream with the authoritative input data necessary to calculate the recent defect rate and risk index, thus ensuring that the decisions of the entire system are always based on real, continuous production quality history, achieving truly data-driven intelligent control.

[0083] The above is only a preferred embodiment of the present application, and does not limit the present application in other forms. Any skilled person in the art can use the disclosed technical content to make changes or modifications to equivalent embodiments for other fields, but any simple modification, equivalent change and modification made to the above embodiments without departing from the technical solution of the present application, in accordance with the technical essence of the present application, still falls within the protection scope of the technical solution of the present application.

Claims

1. A plastic bag quality inspection system based on image recognition, characterized in that, include: The image acquisition module is used to acquire raw image sequences from the production line; The strategy execution module is used to adaptively select a detection model from a preset speed-priority model and a precision-priority model based on dynamic strategy coefficients; the strategy execution module is also used to process the original image sequence using the detection model to output the final quality detection result; The core state perception module is used to calculate the recent comprehensive defect rate and the rate of change of the recent comprehensive defect rate based on historical structured quality data; the core state perception module is also used to combine the recent comprehensive defect rate and the rate of change to calculate the system risk index. The result archiving and feedback module is used to aggregate the final quality inspection results to generate historical structured quality data recorded according to the inspection cycle. The result archiving and feedback module is also used to compare the system risk index with a preset warning threshold, and when the system risk index is higher than the warning threshold for a preset number of consecutive inspection cycles, generate and output a warning instruction to prompt the operator to intervene. The dynamic decision-making and adjustment module is used to input the system risk index into a preset nonlinear adjustment function to generate the dynamic strategy coefficients; The nonlinear adjustment function is an S-shaped function; The dynamic decision-making and adjustment module is used to subtract a preset risk adjustment center point from the system risk index to obtain the risk deviation. The dynamic decision-making and adjustment module is further used to multiply the risk deviation by a preset adjustment curve steepness and transform the result through the S-shaped function to generate the dynamic strategy coefficient.

2. The image recognition-based plastic bag quality inspection system according to claim 1, characterized in that, The image acquisition module is used to capture the original image sequence at an adjustable frame rate and resolution using an industrial camera and a linear array sensor.

3. The image recognition-based plastic bag quality inspection system according to claim 1, characterized in that, The core state perception module is specifically used for: Within a preset sliding time window, the final total number of defects in the historical structured quality data is summed, and the final total number of detections is summed. The core state perception module is also used to divide the sum of the final total number of defects by the sum of the final total number of detections to calculate the recent comprehensive defect rate.

4. The image recognition-based plastic bag quality inspection system according to claim 1, characterized in that, The core state perception module is specifically used for: Multiply the recent comprehensive defect rate by a preset defect rate weight to obtain the weighted defect rate; Multiply the rate of change by a preset rate of change weight to obtain a weighted rate of change; The weighted defect rate is then added to the weighted rate of change to calculate the system risk index.

5. The image recognition-based plastic bag quality inspection system according to claim 1, characterized in that, The strategy execution module is specifically used for: When the dynamic strategy coefficient is lower than the preset speed priority threshold, the speed priority model is selected as the detection model; When the dynamic strategy coefficient is higher than the preset accuracy priority threshold, the accuracy priority model is selected as the detection model. When the dynamic strategy coefficient is between the speed priority threshold and the accuracy priority threshold, the detection model is obtained by adaptively adjusting the internal parameters of the speed priority model, the internal parameters of the accuracy priority model, or simultaneously adjusting the internal parameters of the two models. The internal parameters include the confidence threshold and the IOU threshold.

6. The image recognition-based plastic bag quality inspection system according to claim 1, characterized in that, The strategy execution module is also used to generate hardware adjustment instructions in response to the dynamic strategy coefficients; The image acquisition module is also used to receive the hardware adjustment command and, based on the hardware adjustment command, dynamically adjust the acquisition frame rate and acquisition resolution of the original image sequence.

7. The image recognition-based plastic bag quality inspection system according to claim 1, characterized in that, The historical structured quality data includes records in units of detection cycles; Each record contains the final total number of inspections and the final total number of defects corresponding to the inspection cycle.

Images