Water level sensor intelligent quality detection and sorting system

Abstract

The present invention discloses an intelligent quality detection and sorting system for water level sensors, which relates to the field of data processing technology. The system includes: calling a water level sensor detection program; placing the water level sensor to be detected in a fixture, performing linkage station quality detection, and obtaining a multi-dimensional quality detection data stream of the sensor; constructing a multi-channel for data processing to obtain a sensor quality detection fusion feature set; building a sensor quality adaptive evaluator to obtain sensor quality evaluation parameters; setting a qualified quality threshold, and performing qualified product sorting control on the sensor quality evaluation parameters. The present invention solves the technical problems in the prior art that the quality evaluation is not accurate and comprehensive enough, and the sorting control efficiency is low, and it is difficult to meet the requirements of efficient and accurate production and detection of water level sensors, and achieves the technical effect of improving the accuracy and efficiency of water level sensor detection and sorting, ensuring product quality, and improving the automation and intelligence level of the production process.

Description

Technical Field

[0001] The invention relates to the technical field of data processing, and in particular to an intelligent quality detection and sorting system for a water level sensor. Background Art

[0002] In the field of industrial production, the quality inspection and sorting of water level sensors are crucial. However, the current water level sensor detection technology faces many challenges. Existing detection systems often lack systematic and intelligent integration, and the detection process mostly relies on manual operation. From sensor placement to detection and subsequent sorting at each station, there are many manual links, which not only leads to low efficiency, but also easily introduces errors due to human factors. Moreover, in terms of data processing, there is a lack of effective multi-channel data processing architecture, which makes it difficult to efficiently integrate and process multi-type data (numerical, character, image, etc.) from different detection stations (such as painting, laser marking, CCD detection, bolt height detection and finished product detection, etc.), and it is impossible to fully explore the key features in the data. At the same time, the quality assessment model is not accurate and adaptive enough, and it is difficult to accurately evaluate according to the diverse quality assessment factor indicators of the sensor, which in turn affects the accuracy and efficiency of qualified product sorting, and cannot meet the needs of modern industrial production for high quality and high yield of water level sensors.

[0003] The existing technology has technical problems such as inaccurate and incomplete quality assessment and low efficiency of sorting control, which makes it difficult to meet the needs of efficient and accurate production and testing of water level sensors. Summary of the invention

[0004] The present application provides an intelligent quality inspection and sorting system for water level sensors, which is used to solve the technical problems in the prior art that the quality assessment is not accurate and comprehensive enough and the sorting control efficiency is low, making it difficult to meet the needs of efficient and accurate production and inspection of water level sensors.

[0005] In view of the above problems, the present application provides an intelligent quality detection and sorting system for water level sensors.

[0006] The present application provides an intelligent quality detection and sorting system for water level sensors, the system comprising:

[0007] A water level sensor detection program calling module, which is used to perform a power-on reset operation on the water level sensor intelligent quality detection and sorting equipment, and call the water level sensor detection program according to the water level sensor intelligent quality detection and sorting equipment; a multi-dimensional quality detection data stream acquisition module, which is used to place the water level sensor to be detected into a fixture, activate the water level sensor detection program to perform a linkage workstation quality detection on the water level sensor to be detected, and obtain the sensor multi-dimensional quality detection data stream; a quality detection fusion feature set acquisition module, which is used to construct a quality detection fusion feature set acquisition module A data processing multi-channel, wherein the data processing multi-channel is used to process and fuse the multi-dimensional quality detection data stream of the sensor to obtain a sensor quality detection fusion feature set; a quality assessment parameter acquisition module, wherein the quality assessment parameter acquisition module is used to build a sensor quality adaptive evaluator, and based on the sensor quality adaptive evaluator, the sensor quality detection fusion feature set is subjected to quality assessment to obtain sensor quality assessment parameters; a qualified product sorting control module, wherein the qualified product sorting control module is used to set a qualified quality threshold according to a sensor quality application standard, and perform qualified product sorting control on the sensor quality assessment parameters based on the qualified quality threshold.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] The water level sensor detection program calling module is used to perform a power-on reset operation on the water level sensor intelligent quality detection and sorting equipment, and call the water level sensor detection program; the multi-dimensional quality detection data stream acquisition module is used to put the water level sensor to be detected into the fixture, activate the water level sensor detection program to perform linkage station quality detection, and obtain the sensor multi-dimensional quality detection data stream; the quality detection fusion feature set acquisition module builds a data processing multi-channel, processes and fuses the sensor multi-dimensional quality detection data stream, and obtains the sensor quality detection fusion feature set; the quality assessment parameter acquisition module builds a sensor quality adaptive evaluator, and performs quality assessment on the sensor quality detection fusion feature set based on the sensor quality adaptive evaluator to obtain the sensor quality assessment parameters; the qualified product sorting control module is used to set the qualified quality threshold according to the sensor quality application standard, and perform qualified product sorting control on the sensor quality assessment parameters based on the qualified quality threshold. The technical effect of improving the accuracy and efficiency of water level sensor detection and sorting, ensuring product quality, and improving the automation and intelligence level of the production process is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0011] Figure 1 A structural schematic diagram of an intelligent quality detection and sorting system for a water level sensor is provided for an embodiment of the present application;

[0012] Figure 2 A schematic diagram of a process for obtaining a multi-dimensional quality detection data stream of a sensor in an intelligent quality detection and sorting system for a water level sensor is provided for an embodiment of the present application.

[0013] Explanation of the reference numerals: water level sensor detection program calling module 10 , multi-dimensional quality detection data stream acquisition module 20 , quality detection fusion feature set acquisition module 30 , quality assessment parameter acquisition module 40 , qualified product sorting control module 50 . DETAILED DESCRIPTION

[0014] The present application provides an intelligent quality inspection and sorting system for water level sensors, which is used to solve the technical problems in the prior art that the quality assessment is not accurate and comprehensive enough and the sorting control efficiency is low, making it difficult to meet the needs of efficient and accurate production and inspection of water level sensors.

[0015] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0016] Embodiment 1, as Figure 1 As shown, the present application provides an intelligent quality detection and sorting system for water level sensors, the system comprising:

[0017] The water level sensor detection program calling module 10 is used to perform a power-on reset operation on the water level sensor intelligent quality detection and sorting equipment, and call the water level sensor detection program according to the water level sensor intelligent quality detection and sorting equipment.

[0018] Specifically, when the water level sensor detection program calling module 10 starts operating, it will send a power-on reset command to the water level sensor intelligent quality detection and sorting equipment. This operation is similar to the power-on restart of a computer system. Its purpose is to initialize the various circuits, chips and related components inside the equipment, so that they return to a stable and known initial state, ensuring that subsequent detection work can be carried out in a unified and reliable basic environment. After completing the power-on reset, relying on its in-depth understanding of the hardware architecture and functional characteristics of the device, it accurately selects and calls the water level sensor detection program adapted to the device from the system storage. This program is the core instruction set of the entire detection process, which contains a series of algorithms, logics and operation steps for detecting various performance indicators and quality characteristics of the water level sensor. Once correctly called, it will start a series of subsequent complex and orderly detection work, laying a solid foundation for comprehensive and accurate evaluation of the quality of the water level sensor.

[0019] The multi-dimensional quality detection data stream acquisition module 20 is used to place the water level sensor to be detected into a fixture, activate the water level sensor detection program to perform a linkage workstation quality detection on the water level sensor to be detected, and obtain the sensor multi-dimensional quality detection data stream.

[0020] Specifically, through a precise mechanical operating mechanism, the water level sensor to be tested is placed in a fixture with a high-precision positioning function in a preset posture and position. The fixture adopts advanced mechanical structure design and materials to ensure that the sensor can be stably fixed without causing any physical damage or performance impact to it, and at the same time provides a unified reference position for subsequent testing. Subsequently, a start instruction is sent to the called water level sensor detection program to trigger the multi-threaded or multi-process task scheduling mechanism in the detection program. According to the preset detection process and logic, the detection program activates multiple linkage stations in sequence. These stations are equipped with high-precision sensors and detection equipment, such as high-precision resistance measuring instruments, high-precision pressure sensors, high-resolution optical imaging equipment, etc., to test different performance indicators of water level sensors. At the electrical performance detection station, a series of precise electrical signal excitations are applied to the sensor using a precise electronic measurement circuit. At the same time, the output signal of the sensor is collected in real time through a high-speed analog-to-digital converter (ADC), and pre-processing operations such as filtering, amplification, and sampling are performed on it to obtain the response data of the sensor under different electrical signal inputs, including electrical characteristic parameters such as sensitivity, linearity, and hysteresis. The mechanical structure inspection station uses high-precision three-dimensional laser scanning technology or contact profile measuring instruments to scan and measure the sensor's housing, pins and other mechanical parts to obtain data such as its geometric dimensions, shape accuracy, surface roughness, and detect whether there are defects such as cracks, deformations, and burrs through image analysis algorithms and geometric calculation models. For the sealing test of the sensor, a specially designed sealing test cavity is used to apply a specific pressure or vacuum environment to the cavity. The pressure changes in the cavity are monitored in real time through a high-precision pressure sensor, and combined with gas flow measurement technology, it is determined whether the sensor has leakage and obtains data related to its sealing performance. During the inspection process, each station encapsulates the collected real-time data in accordance with a strictly defined data format and communication protocol, and transmits the data to the module multi-dimensional quality inspection data stream acquisition module 20 through a high-speed data bus (such as industrial Ethernet, high-speed serial bus, etc.). The multi-dimensional quality detection data stream acquisition module 20 utilizes an efficient data caching and management mechanism to integrate and synchronize information from different workstations, different time points, and different data types, and ultimately forms a sensor multi-dimensional quality detection data stream containing multi-dimensional quality information, providing a comprehensive and accurate data source for subsequent data analysis and processing.

[0021] The quality detection fusion feature set acquisition module 30 is used to construct a data processing multi-channel, and use the data processing multi-channel to process and fuse the sensor multi-dimensional quality detection data stream to obtain the sensor quality detection fusion feature set.

[0022] Specifically, the quality detection fusion feature set acquisition module 30 undertakes the data processing and feature extraction tasks in the entire water level sensor quality detection system. Its workflow begins with the construction of a multi-channel for data processing, and uses advanced software architecture design concepts and hardware resource allocation strategies to create multiple parallel data processing paths. These paths are logically independent of each other but work together, similar to building an efficient multi-lane data highway. After receiving the sensor multi-dimensional quality detection data stream from the multi-dimensional quality detection data stream acquisition module, the module parses and splits the data stream according to the pre-set data format and transmission protocol, and accurately distributes data of different dimensions and types to the corresponding data processing channels. In each data processing channel, data preprocessing operations are first performed, and digital signal processing techniques, such as filtering algorithms, are used to remove noise interference in the data, and data normalization operations map the data to a specific numerical range to improve the stability and comparability of the data. Then, a variety of feature extraction algorithms are used for deep mining of the preprocessed data. For example, for the data related to the electrical signal of the sensor, the frequency characteristics are extracted by the spectrum analysis algorithm, and the amplitude change characteristics of the signal are obtained by time domain analysis; for the data related to the mechanical structure, the geometric features such as edge features and texture features are extracted from the optical imaging data by using the image feature extraction algorithm. After the feature extraction of each channel is completed, the features from different channels are fused based on the preset data fusion rules and algorithms, such as the weighted average fusion algorithm and principal component analysis (PCA). By calculating the weight coefficient of each feature and comprehensively considering the contribution of each feature to the overall quality assessment, multiple feature vectors are merged into a comprehensive sensor quality detection fusion feature set. This fusion feature set highly condenses the key information in the multi-dimensional quality detection data stream of the sensor, providing a more accurate and effective data basis for subsequent quality assessment, so that the judgment of the quality of the water level sensor can be based on more comprehensive and representative feature information, thereby improving the accuracy and reliability of the entire quality detection system.

[0023] The quality assessment parameter acquisition module 40 is used to build a sensor quality adaptive evaluator, and perform quality assessment on the sensor quality detection fusion feature set based on the sensor quality adaptive evaluator to obtain sensor quality assessment parameters.

[0024] Specifically, the quality assessment parameter acquisition module 40 plays a key role in the evaluation and decision support in the water level sensor quality detection system. Its primary task is to build a sensor quality adaptive evaluator. In the process of building, machine learning and artificial intelligence technology are fully utilized. First, a large amount of quality detection data of historical water level sensors is collected, including sensor quality detection fusion feature sets and corresponding actual quality result data, which constitute the training data set. The neural network algorithm is used to build the basic architecture of the evaluation model, in which the number of input layer nodes matches the number of features in the sensor quality detection fusion feature set, and the output layer is set according to the quality assessment target, such as outputting a numerical value representing the quality level or a Boolean value indicating whether the quality is qualified or not. In the model training stage, by inputting the fusion feature set in the training data set into the evaluation model, the model is calculated based on the internal algorithm logic, and compared with the actual quality result data to calculate the prediction error. Then, the back propagation algorithm is used to adjust the weights and parameters of the model according to the error. After multiple iterative trainings, the model is continuously optimized until the model reaches the preset accuracy and stability standards, thereby completing the construction of the sensor quality adaptive evaluator. When the quality detection fusion feature set of the sensor to be evaluated is received, it is input into the trained sensor quality adaptive evaluator. The algorithm inside the evaluator will perform a series of complex calculations and analyses on the input feature set, and comprehensively evaluate the sensor quality based on the relationship pattern between the features and quality learned during the training process. Finally, a set of sensor quality evaluation parameters are output. These parameters can quantitatively reflect the performance of the sensor in multiple quality dimensions, such as numerical indicators of reliability, accuracy, stability, etc., providing an accurate judgment basis for subsequent qualified product sorting control, ensuring the scientificity and accuracy of the entire quality detection and sorting process.

[0025] The qualified product sorting control module 50 is used to set a qualified quality threshold according to the sensor quality application standard, and perform qualified product sorting control on the sensor quality assessment parameters based on the qualified quality threshold.

[0026] Specifically, the qualified product sorting control module 50, as the key execution unit of the water level sensor intelligent quality detection and sorting system, is mainly responsible for classifying and screening the detected sensors according to the precisely set standards. At the beginning of its work, it will set the qualified quality threshold through a rigorous calculation and analysis process based on the sensor quality application standards determined by various factors such as industry specifications, product design requirements and actual application scenarios. The determination of this threshold is not arbitrary, but takes into account many factors such as the importance weight of various performance indicators of the sensor in actual use and the degree of influence on the operation of the overall system to ensure that it can accurately distinguish qualified products from unqualified products. After receiving the sensor quality evaluation parameters from the quality evaluation parameter acquisition module, the sorting control process is quickly started. It compares the quality evaluation parameters of each sensor with the pre-set qualified quality threshold one by one. If the quality evaluation parameters of the sensor fully meet the requirements of the qualified quality threshold, it means that the sensor has reached or exceeded the qualified standards in all quality indicators, and it will be judged as a qualified product, and the corresponding qualified product logo will be marked for it, and the corresponding control instructions will be generated at the same time to guide the unloading station to accurately grasp the sensor and place it on the qualified product conveyor belt for subsequent packaging and other processes. On the contrary, if any of the quality assessment parameters of the sensor does not reach the qualified quality threshold, it will be judged as a defective product, assigned a defective product label, and the unloading station will be directed to grab the sensor to a specially designed defective product area, such as a defective product box or a defective product conveyor belt, so as to achieve accurate sorting of the sensors, ensure the quality reliability of the final product, and maintain the stability and efficiency of the entire production and application links.

[0027] In one possible implementation, Figure 2 As shown, the method of obtaining the multi-dimensional quality detection data stream of the sensor includes:

[0028] Based on the water level sensor detection program, the sensor detection station set is started, and the sensor detection station set includes a painting station, a laser marking station, a CCD detection station, a bolt height detection station, and a finished product detection station. According to the sensor quality application standard, the appearance design information and performance requirement targets of the water level sensor to be detected are obtained. Based on the appearance design information and performance requirement targets, the sensor detection station set is analyzed for detection parameters in turn to obtain a station detection parameter set. The sensor detection station set performs linkage station quality detection on the water level sensor to be detected according to the station detection parameter set, and obtains the multi-dimensional quality detection data stream of the sensor.

[0029] Specifically, when the sensor detection station set is started based on the water level sensor detection program, the entire detection process works in coordination. The first station to be started is the paint spotting station. On the basis of ensuring that the surface of the water level sensor is clean and flat, the station uses high-precision paint spraying equipment to spray the paint evenly and accurately on the designated area of ​​the sensor housing according to the preset paint type, color and process standards. This not only protects the sensor housing from environmental erosion, but also makes the sensor appearance meet the design specifications, while avoiding any negative impact of the paint on the sensor performance. Then the laser marking station starts to operate. With the help of a high-power, high-stability laser emitting device, according to the predetermined coding rules and identification format, clear, accurate and indelible product information such as model, batch number, production date and other key identifications are permanently engraved at a specific position of the sensor housing for subsequent production management, quality traceability and product identification. Then the CCD detection station is started, using a high-resolution CCD image sensor and an advanced optical imaging system to collect comprehensive appearance images of the water level sensor to be detected from multiple angles. Through sophisticated image analysis algorithms, the collected images are deeply processed to carefully check whether there are scratches, pits, bubbles, uneven colors and other defects on the surface of the sensor housing, as well as whether the assembly between the components is accurate and correct, so as to strictly control the appearance quality of the sensor. The bolt height detection station uses a high-precision laser rangefinder or contact displacement sensor to accurately measure the actual height of the bolts on the water level sensor and compare it with the design standard height in real time. If a deviation in the bolt height is found, the alarm system will be triggered immediately so that unqualified products can be adjusted or removed in time to ensure the accuracy of the bolt height, because even a slight deviation in the bolt height may affect the installation stability and sealing performance of the sensor. The finished product inspection station that is started last is the key checkpoint of the entire inspection process. It uses a variety of advanced testing equipment and technical means to conduct comprehensive performance tests on the sensor. In the sensitivity test, the water level changes are precisely controlled to simulate the tiny water level fluctuations under actual working conditions, and the sensor output signal change amplitude and response speed are monitored using high-precision measuring instruments to ensure that the sensor can keenly capture the subtle changes in the water level; in the accuracy test, the sensor is placed in a standard test environment with different water level heights, and the error between the sensor measurement value and the actual water level value is compared, and its measurement accuracy is strictly required to be within the specified range; the stability test is to let the sensor continuously monitor the fluctuation of its performance parameters under the impact of complex environmental conditions such as temperature, humidity, and pressure for a long time, to ensure that the sensor can still maintain stable working performance in various harsh environments. Only when the sensor successfully passes all tests in all these testing stations can it be judged as a qualified product and enter the next production link. The coordinated operation of the entire sensor testing station set, from appearance to performance, fully ensures that the quality of the water level sensor meets high standards.

[0030] According to the sensor quality application standard, when obtaining the appearance design information and performance requirement targets of the water level sensor to be tested, the system will deeply study and analyze the existing sensor quality application standard documents or database records. In terms of obtaining appearance design information, the dimensional parameters of the water level sensor in the standard are carefully consulted, including specific values ​​such as length, diameter, height, and the proportional relationship between the various parts; for the shape, its contour characteristics are clarified, such as round, square or other special shape requirements; the color specification covers specific color numbers or color ranges such as the main color and logo color; material requirements involve material type, surface treatment process, etc., such as whether a smooth surface, frosted texture, etc. are required; there are also specific regulations for decorative designs such as surface texture and pattern, such as whether there is an anti-slip texture, the style and position of the brand logo pattern, etc. In terms of obtaining performance requirements, we focus on the accuracy of water level measurement and clarify the allowable error range in different water level intervals, such as accurate to millimeter level when measuring low water levels and within a certain percentage when measuring high water levels; in terms of response time, we determine the longest time from when the water level changes to when the sensor outputs the corresponding signal, such as requiring a response within a few seconds; stability requirements involve the fluctuation range of sensor performance under different ambient temperature, humidity, and pressure conditions to ensure that it can work stably in complex environments; repeatability requirements stipulate the maximum deviation value of the measurement results when measuring the same water level multiple times; in terms of durability, we determine the expected service life of the sensor under normal use conditions and the attenuation requirements of various performance indicators during the service life according to the standards. Through the careful sorting of these contents, the system can completely and accurately obtain the appearance design information and performance requirements of the water level sensor to be tested, laying a solid foundation for the subsequent testing process.

[0031] Based on the appearance design information and performance requirements, the process of analyzing the detection parameters of the sensor detection station set to obtain the station detection parameter set is a meticulous and critical link. For the paint spotting station, the specific location of the paint application is accurately determined based on the shape, size and surface decoration requirements of the sensor shell in the appearance design information, such as spraying the marking paint in a specific area of ​​the shell, or evenly covering the protective paint on the entire shell surface. At the same time, combined with the requirements for the thickness and uniformity of the paint layer in the performance requirements, and the consideration of the impact of the paint on the sensor performance, the flow rate of each spraying, the movement speed of the spray gun, the number of spraying times and other parameters are determined through calculation and experience setting to ensure that the paint layer can achieve both aesthetic and protective effects, and will not have an adverse effect on the working performance of the sensor. At the laser marking station, according to the requirements of the position, size, font style and other requirements of the product logo in the appearance design information, and the requirements of the clarity and durability of the mark in the performance requirements, the power, frequency, pulse width and other parameters of the laser marking are analyzed. Ensure that the mark on the sensor housing is clear, accurate, and not easy to wear, which can not only meet the needs of product traceability and identification, but also will not damage the internal structure or performance of the sensor due to the marking process. For the CCD inspection station, according to the appearance quality standards of the sensor in the appearance design information, such as surface flatness, color consistency, texture clarity, etc., and the requirements for detection accuracy and speed in the performance requirements, determine the resolution, shooting angle, exposure time, and image processing algorithm parameters of the CCD camera. Enable the camera to accurately capture the subtle features of the sensor appearance, and quickly and accurately determine whether the sensor appearance meets the quality standards through image analysis. For the bolt height inspection station, according to the specifications, positions, and assembly relationships of the bolts in the appearance design information, combined with the requirements for bolt height accuracy in the performance requirements, analyze the measurement range, measurement accuracy, measurement method and other parameters of the measuring instrument. Ensure that the bolt height can be accurately measured, the height deviation can be discovered in time, and the accuracy of the bolt installation can be ensured, thereby ensuring the stability of the overall structure of the sensor and the reliability of performance. At the finished product inspection station, based on the specific requirements for sensor sensitivity, accuracy, stability and other performance indicators in the performance demand goals, as well as the special needs in different application scenarios, the test range, test accuracy, test frequency, environmental simulation parameters, etc. of the test equipment are analyzed. For example, in the sensitivity test, the minimum detectable amount of water level change and the corresponding sensor output signal change range are determined; in the accuracy test, the allowable error range between the standard water level value and the sensor measurement value is clarified; in the stability test, the range of change of ambient temperature, humidity, pressure and test duration are set. Through the precise analysis of these parameters, a comprehensive, scientific and reasonable detection basis is provided for the finished product inspection station, ensuring that the test results can truly reflect whether the performance of the sensor meets the quality requirements, thereby obtaining an accurate and reliable station detection parameter set, and providing solid technical support for subsequent linkage station quality inspections.

[0032] After receiving the station detection parameter set, the sensor detection station set immediately starts the linkage station quality detection process for the water level sensor to be detected, thereby obtaining the sensor multi-dimensional quality detection data stream. The paint point station accurately controls the spraying position, thickness and uniformity of the paint according to the set parameters to ensure that the appearance of the paint layer meets the design requirements and does not affect the performance of the sensor. During the painting process, the paint flow rate, spray gun pressure and other data are monitored in real time, and the test results such as paint layer thickness and color uniformity are recorded to form a data subset related to the paint layer quality. The laser marking station adjusts the laser power, frequency and marking speed according to the parameters, and clearly and accurately marks the permanent mark at the specified position of the sensor housing. At the same time, the laser energy stability, marking depth consistency, etc. are monitored, and the marking clarity, position accuracy and other information are recorded to form a marking quality data subset. The CCD detection station sets the camera resolution, shooting angle and exposure time according to the parameters to collect all-round images of the sensor appearance. Image processing algorithms are used to analyze images to detect whether there are defects on the surface of the shell and whether the assembly is accurate. Parameters such as defect size and position deviation are quantified to generate a subset of appearance quality inspection data. The bolt height inspection station uses specified measuring instruments and measurement methods to accurately measure the bolt height and compare it with the standard value. The actual value of the bolt height, the deviation, and the instrument stability data during the measurement process are recorded to form a subset of bolt height inspection data. The finished product inspection station uses a variety of test equipment to comprehensively test the performance indicators of the sensor, such as sensitivity, accuracy, and stability, according to the set test range, accuracy, and frequency. Actual working conditions are simulated under different water level conditions, the changes in sensor output signals are monitored, and the sensitivity response curve, measurement error value, and performance fluctuation data during long-term testing are recorded to construct a subset of performance quality inspection data. During the inspection process, each station integrates and synchronizes the data subsets generated by each station according to a unified data format and transmission protocol, and finally converges into a sensor multi-dimensional quality inspection data stream containing multi-dimensional information such as paint layer quality, marking quality, appearance quality, bolt height quality, and performance quality, providing a rich and comprehensive data foundation for subsequent data processing and quality evaluation.

[0033] In a possible implementation, the constructing data processing multiple channels includes:

[0034] According to the water level sensor detection program, a data processing channel architecture is designed, and the number of channels in the data processing channel architecture corresponds one-to-one to the station sequence in the water level sensor detection program; each processing channel in the data processing channel architecture is data structured and classified to obtain multi-channel data structured attributes; based on the multi-channel data structured attributes, a processing program analysis is performed on each processing channel in the data processing channel architecture to obtain a multi-channel data processing node program set; based on the multi-channel data processing node program set, a processing program is configured for each processing channel in the data processing channel architecture to construct the data processing multi-channel.

[0035] Specifically, in the quality inspection fusion feature set acquisition module, the construction and related configuration of the data processing channel architecture are crucial. First, based on the water level sensor detection program, the various station sequences involved are deeply analyzed, including the operation process and data interaction logic of the painting station, laser marking station, CCD detection station, bolt height detection station, and finished product detection station. Based on this, the corresponding data processing channel architecture is carefully designed to ensure that the number of channels fully matches the station sequence, laying the foundation for the orderly processing of subsequent data.

[0036] Then, for each processing channel in the data processing channel architecture, we started to carry out data structured classification. We carefully sorted out the data types involved in the detection process of each channel, such as the paint layer thickness, color and other numerical data generated by the painting station, the marking position, logo content and other character data of the laser marking station, the image data collected by the CCD detection station, the height measurement value of the bolt height detection station, and various performance test values ​​of the finished product detection station. At the same time, we clarified its structured format, such as the pixel format of the image data, the accuracy and range of the numerical data and other related data attributes, so as to obtain comprehensive and accurate multi-channel data structured attributes.

[0037] Based on these multi-channel data structured properties, an in-depth analysis of the processing procedures for each processing channel is carried out. For different data types and structured formats, appropriate processing algorithms and operation logic are determined. For numerical data, programs such as data cleaning, numerical conversion, and statistical analysis are involved; character data focuses on processing such as format verification and encoding conversion; image data requires image preprocessing, feature extraction, and other operations. Through this analysis process, a complete and targeted set of multi-channel data processing node programs is obtained, and each program corresponds to the data processing requirements of a specific channel.

[0038] Finally, the obtained multi-channel data processing node program set is used to perform detailed program configuration for each processing channel in the data processing channel architecture. The corresponding processing program is accurately deployed to the corresponding channel, and the program operation parameters, priorities, and data interaction rules are set. After this series of rigorous operations, an efficient data processing multi-channel is successfully built, so that data of different types and formats from various inspection stations can be scientifically and reasonably processed in their own dedicated channels, providing strong technical support for the subsequent fusion processing of sensor multi-dimensional quality inspection data streams and the acquisition of quality inspection fusion feature sets.

[0039] In a possible implementation, obtaining a multi-channel data processing node assembly includes:

[0040] Define a data channel processing node, the data channel processing node includes a data preprocessing node and a feature extraction node; based on the data preprocessing node, perform preprocessing program analysis on the multi-channel data structured attributes respectively to obtain a multi-channel data preprocessing program set; use the feature extraction node to perform feature extraction program analysis on the multi-channel data structured attributes respectively to obtain a multi-channel data feature extraction program set; based on the multi-channel data preprocessing program set and the multi-channel data feature extraction program set, determine the multi-channel data processing node program set.

[0041] Specifically, the data channel processing node is the core element in the data processing flow, in which the data preprocessing node and the feature extraction node work together to achieve effective data processing and feature mining. The data preprocessing node uses a variety of algorithms and techniques to preprocess the raw data in the structured attributes of multi-channel data. For numerical data, the filtering algorithm (such as Kalman filtering) is used to remove measurement noise, and the data distribution is normalized through data standardization (such as Z-score standardization) to improve data stability; for character data, the encoding format is unified using the encoding conversion algorithm (such as UTF-8 conversion), and the string processing function is used for format verification and cleaning; for image data, the image filtering algorithm (such as median filtering) is used to eliminate noise, and the image contrast is enhanced through histogram equalization. The feature extraction node performs targeted feature extraction according to the data type. For image data (such as images collected by visual inspection stations), a convolutional neural network (CNN) is used. Its convolution layer extracts local features of the image (such as edges and textures), the pooling layer reduces the data dimension, and the fully connected layer integrates feature information to obtain appearance and structural features. For numerical data, algorithms such as principal component analysis (PCA) are used to extract key feature vectors and mine the main change trends and feature information in the data, thereby providing high-quality data features for subsequent data processing and analysis.

[0042] In the process of performing preprocessor analysis on the structured attributes of multi-channel data based on the data preprocessor node to obtain the multi-channel data preprocessor set, the data preprocessor node first conducts an in-depth analysis of the structured attributes of each channel data. For channels containing numerical data, such as the bolt height data generated by the bolt height detection station, analyze its data distribution range, data accuracy requirements, and possible outlier patterns. Based on these analysis results, determine the data cleaning algorithm, such as using an outlier detection method based on statistical principles, setting a reasonable threshold to identify and eliminate data points that are obviously deviated from the normal range; at the same time, select a suitable data normalization or standardization algorithm, such as minimum-maximum normalization to map the data to a specific interval, to ensure that data of different magnitudes are comparable in subsequent processing. For character data channels, such as the marking content data generated by the laser marking station, analyze the character encoding format, string length distribution, and possible character error types. On this basis, a coding conversion strategy is formulated to uniformly convert characters of different codes into a standard coding format, such as converting ASCII code into UTF-8 code; at the same time, a character check and error correction program is designed, and regular expressions and other technologies are used to check whether the string conforms to the predetermined format specification, and the characters that do not conform are corrected or marked to ensure the consistency and accuracy of the data. For the image data channel, such as the image data collected by the CCD detection station, the structural attributes such as image resolution, color space, and image noise characteristics are analyzed. According to these attributes, the image preprocessing algorithm is determined, such as using Gaussian filtering or median filtering algorithms to remove noise interference in the image and improve the image quality; image enhancement technology, such as histogram equalization, is used to adjust the image contrast to make the image features more obvious; it may also include image cropping and scaling algorithms to adjust the image to a suitable size according to the subsequent processing requirements, reducing the amount of data processing while retaining key image information. Through the detailed analysis of the structural attributes of each channel data and the determination of the corresponding preprocessing program, a comprehensive and targeted multi-channel data preprocessing program set is finally integrated to lay a solid foundation for the high-quality processing of subsequent data.

[0043] When using the feature extraction node to perform feature extraction program analysis on the structured attributes of multi-channel data to obtain a multi-channel data feature extraction program set, corresponding advanced technologies and algorithms are used for different types of data channels. Taking the image data collected by the visual inspection station as an example, its data structured attributes include information such as image resolution and color mode. The feature extraction node uses a convolutional neural network (CNN) to perform feature extraction program analysis for these attributes. The convolution layer of CNN performs convolution operations with the image through multiple different convolution kernels, which can effectively extract local features in the image, such as edge, texture and other detail information. The pooling layer downsamples the output of the convolution layer to reduce the amount of data while retaining key features and enhancing the robustness of the features. The fully connected layer integrates the features after multiple convolutions and pooling to obtain the appearance features of the image, such as shape, color distribution, etc., as well as structural features, such as the relative position relationship between components. Through the careful design and adjustment of the CNN network structure, parameter settings and training algorithms, a feature extraction program for image data is constructed. For other types of data channels, such as numerical data channels, the feature extraction node will select a suitable algorithm based on the characteristics of the data. For example, for time-series data, recurrent neural networks (RNN) or long short-term memory networks (LSTM) are used to extract dynamic features in the data, such as changing trends, periodicity, etc. For discrete numerical data, statistical analysis methods are used to extract features, such as calculating statistical quantities such as mean, variance, mode, etc., or clustering algorithms are used to mine potential patterns and features in the data. Through in-depth analysis of the structured attributes of each channel data and the construction of corresponding feature extraction programs, a set of multi-channel data feature extraction programs covering multiple data types and capable of effectively extracting key features is ultimately formed, providing rich and valuable feature information for subsequent data fusion and quality assessment.

[0044] The multi-channel data preprocessing program set is sorted out to clarify the data preprocessing process and algorithm configuration of each channel. For example, for channels containing numerical data, the preprocessing program may include data cleaning algorithms to remove outliers and normalization algorithms to adjust the data range; for character data channels, the preprocessing program covers operations such as encoding conversion and format verification; the preprocessing program of image data channels includes steps such as image filtering, enhancement and size adjustment. These preprocessing programs lay the foundation for the preliminary optimization and standardization of data. Next, the multi-channel data feature extraction program set is analyzed. For different types of data channels, such as the visual inspection station image data channel, convolutional neural network (CNN) is used for feature extraction. Its program includes the configuration and parameter setting of convolution layer, pooling layer and fully connected layer to extract appearance and structural features; numerical data channels use corresponding algorithms according to data characteristics, such as recurrent neural network (RNN) or statistical analysis methods to extract dynamic features or statistical features. These feature extraction programs are designed to mine key information in the data. Then, the order, logical relationship and data dependency of data preprocessing and feature extraction are comprehensively considered. Generally speaking, data is first processed by a preprocessor to improve the data quality before entering the feature extraction program to ensure that the extracted features are accurate and reliable. For example, image data must first be preprocessed by filtering, enhancement, etc. before CNN can better extract effective features. Finally, the multi-channel data preprocessor and feature extraction program are combined and optimized in a reasonable order and logic to determine the data processing node program for each channel. This program set covers the complete process from data input to final feature extraction, ensuring that the data of each channel can be processed efficiently and accurately throughout the entire processing process, providing solid technical support for subsequent data analysis, fusion, and quality assessment, thereby achieving accurate detection and evaluation of the quality of water level sensors.

[0045] In a possible implementation manner, obtaining the sensor quality detection fusion feature set includes:

[0046] The data processing multi-channel is used to map and parallelly process the multidimensional quality detection data stream of the sensor to obtain a multidimensional sensor quality detection feature set; a fusion demand analysis is performed on the multidimensional sensor quality detection feature set to obtain feature fusion demand characteristic information; feature matching is performed based on the feature fusion demand characteristic information and a data fusion algorithm list to obtain a target feature fusion algorithm; feature fusion is performed on the multidimensional sensor quality detection feature set based on the target feature fusion algorithm to obtain the sensor quality detection fusion feature set.

[0047] Specifically, a carefully constructed data processing multi-channel is used to map and parallelize the sensor multi-dimensional quality detection data stream. Each data processing channel independently processes the data stream of the corresponding dimension according to its pre-set program and algorithm. For example, for the data channel from the paint point station, it focuses on the analysis of relevant data such as paint layer thickness and uniformity, and uses image processing technology and numerical analysis methods to extract features related to paint layer quality; the data channel of the laser marking station focuses on processing data such as marking clarity and position accuracy, and obtains features through character recognition and position detection algorithms; the data channel of the CCD detection station conducts in-depth mining of image data to extract various features of the sensor appearance, such as surface defects, component assembly accuracy, etc.; the data channel of the bolt height detection station accurately analyzes the bolt height data and calculates features such as height deviation; the data channel of the finished product inspection station comprehensively processes a variety of performance test data, such as data related to indicators such as sensitivity, accuracy, and stability, and uses statistical analysis and signal processing technology to obtain performance characteristics. Through this mapping parallel processing method, each channel can simultaneously and efficiently process different types of data, give full play to the advantages of multiple channels, and quickly obtain a multi-dimensional sensor quality detection feature set, in which the features of each dimension reflect the quality status of the sensor from a specific angle.

[0048] Next, the acquired multi-dimensional sensor quality detection feature set is subjected to fusion demand analysis. This process deeply studies the intrinsic relationship and mutual influence between each feature and their importance to the overall sensor quality assessment. For example, the potential correlation between the paint quality feature and the long-term stability of the sensor, the matching degree between the marking feature and the product traceability and identification requirements, the influence of the appearance feature on the sealing and reliability of the sensor, the relationship between the bolt height feature and the overall structural stability, and the synergy between the performance features are analyzed. Through such analysis, it is clear which features need to be deeply fused to improve the accuracy of quality assessment, and which features can be used as auxiliary information to provide a more comprehensive quality description, so as to obtain the characteristic information of feature fusion requirements, and provide a basis for the subsequent selection of appropriate fusion algorithms.

[0049] Then, feature matching is performed based on the feature fusion requirement characteristic information and the data fusion algorithm list. The data fusion algorithm list covers a variety of advanced fusion algorithms, such as weighted average fusion algorithm, principal component analysis (PCA) fusion algorithm, Kalman filter fusion algorithm, etc. According to the requirements of feature importance, data type, correlation, etc. in the feature fusion requirement characteristic information, the most suitable target feature fusion algorithm is selected from the algorithm list. For example, if there is a strong linear relationship between the features and the importance of certain key features needs to be highlighted, the weighted average fusion algorithm is selected; if the data dimension needs to be reduced and the main feature information needs to be retained, the PC fusion algorithm is more suitable; for dynamic data and the need to update the fusion results in real time, the Kalman filter fusion algorithm has advantages.

[0050] Finally, the determined target feature fusion algorithm is used to fuse the multi-dimensional sensor quality detection feature set. The algorithm organically integrates the features of each dimension according to the pre-set rules and parameters. For example, the weighted average fusion algorithm performs weighted summation of different features according to the importance weight of the features to obtain a comprehensive feature vector; the PCA fusion algorithm performs eigenvalue decomposition on the feature covariance matrix, selects the main eigenvectors to construct a new feature space, and projects the original features into the space to achieve fusion; the Kalman filter fusion algorithm uses the state equation and the observation equation to recursively estimate and fuse the dynamic features. After feature fusion, the sensor quality detection fusion feature set is obtained. This fusion feature set highly condenses the key information in the sensor multi-dimensional quality detection data stream, can more accurately and comprehensively reflect the quality status of the sensor, and provides a solid data foundation for subsequent quality assessment and sorting control.

[0051] In a possible implementation, the step of building a sensor quality adaptive evaluator includes:

[0052] The sensor quality detection feature data set is acquired through collection and processing, and the sensor quality detection feature data set includes historical sensor quality detection fusion feature data and corresponding sensor quality result data; a quality assessment factor indicator set of the water level sensor to be detected is acquired; based on the sensor quality detection feature data set, data association is performed on each factor indicator in the quality assessment factor indicator set to obtain a sensor factor indicator detection feature data set; branch evaluation training and model integration fusion are performed on the sensor factor indicator detection feature data set to build the sensor quality adaptive evaluator.

[0053] Specifically, the sensor quality detection feature data set is acquired through collection and processing. By collecting and organizing a large amount of past sensor detection data, historical sensor quality detection fusion feature data is obtained. These data have gone through the previous data processing process and integrated the multi-dimensional feature information from each detection station, which fully reflects the quality characteristics of the sensor in different aspects. At the same time, the corresponding sensor quality result data are also recorded. These result data clarify the quality performance of the sensor in actual use or rigorous testing, such as whether it is qualified, specific performance indicator values, etc. The two together constitute the sensor quality detection feature data set, which provides a rich data foundation for the subsequent construction of the evaluator.

[0054] Next, we need to obtain a set of quality assessment factor indicators for the water level sensor to be tested, which requires in-depth research on the technical specifications, performance requirements, and actual application scenarios of the water level sensor. For example, key performance indicators such as water level measurement accuracy, response speed, stability, and durability, as well as physical property indicators such as the sealing and corrosion resistance of the sensor housing, and adaptability indicators under different environmental conditions will all be determined as quality assessment factor indicators. These indicators cover all aspects of the sensor from function to structure, from short-term performance to long-term reliability, forming a comprehensive and targeted set of quality assessment factor indicators.

[0055] Then, based on the sensor quality detection feature data set, data association is performed on each factor indicator in the quality assessment factor indicator set. For each quality assessment factor indicator, closely related data features are mined from the huge sensor quality detection feature data set. For example, for the factor indicator of water level measurement accuracy, the related internal circuit parameter characteristics of the sensor, the physical characteristics of the sensing element, and the measurement data characteristics under different water level conditions are associated. Through this data association operation, the most representative and relevant data subsets are selected for each factor indicator. These subsets together constitute the sensor factor indicator detection feature data set, so that each factor indicator can be accurately reflected and analyzed at the data level.

[0056] Finally, the sensor factor index detection feature data set is subjected to branch evaluation training and model integration fusion to build a sensor quality adaptive evaluator. In the branch evaluation training stage, the detection feature data set of each factor index is independently trained using a suitable machine learning algorithm or statistical analysis method. For example, for numerical feature data, the regression analysis algorithm is used to train the prediction model; for classification-related feature data, the classification algorithm (such as decision tree, support vector machine, etc.) is used for model training. Each factor index has a corresponding evaluation model, which has certain accuracy and reliability in their respective evaluation dimensions. Then, these branch models are organically combined through model integration fusion technology. A weighted average fusion strategy is adopted to assign different weights according to the performance of each model in the training process, and the prediction results of each model are combined to obtain a sensor quality adaptive evaluator that can comprehensively and comprehensively evaluate the sensor quality. The evaluator can accurately output the sensor quality evaluation parameters based on the input sensor quality detection fusion feature data, providing a scientific and accurate decision-making basis for the subsequent qualified product sorting control.

[0057] In a possible implementation, obtaining a sensor factor index detection feature data set includes:

[0058] Perform criticality assessment on each factor indicator in the quality assessment factor indicator set to obtain a factor indicator criticality coefficient set; determine an indicator feature association depth factor set based on the factor indicator criticality coefficient set; perform data association on the sensor quality detection feature data set based on the indicator feature association depth factor set to obtain the sensor factor indicator detection feature data set.

[0059] Specifically, it is a comprehensive and systematic process to conduct a critical evaluation of each factor indicator in the quality assessment factor indicator set to obtain a set of factor indicator critical coefficients. First, it is necessary to deeply analyze the working principle and application scenarios of the water level sensor. For example, the water level sensor relies on specific physical principles (such as pressure difference, ultrasonic reflection, etc.) to measure water level changes. Its measurement accuracy, stability, and response speed are directly related to the accuracy and timeliness of water level monitoring. In practical applications, such as water level monitoring in water conservancy projects and liquid level control in industrial production, small deviations in these indicators may lead to serious consequences. Therefore, they will be given a higher weight in the critical evaluation. From the perspective of the structural characteristics of the sensor, factors such as the sealing of its shell, the stability of the internal circuit, and the assembly accuracy of each component should not be ignored. Poor sealing may cause water to penetrate and damage internal components, affecting the life and performance of the sensor; circuit instability will cause fluctuations and errors in the measurement data; insufficient assembly accuracy of components may affect the overall reliability and consistency of the sensor. The importance of these structural related indicators will be reasonably considered in the evaluation according to the degree of their impact on the sensor function. At the same time, considering the working requirements of the sensor under different environmental conditions, environmental adaptability indicators such as temperature adaptability, humidity tolerance, and anti-electromagnetic interference ability are also key parts of the evaluation. In a high temperature environment, the material properties of the sensor and the characteristics of the electronic components may change, affecting the measurement accuracy; a high humidity environment may accelerate the corrosion of metal parts and the risk of short circuits in the circuit; a strong electromagnetic interference environment will interfere with the signal transmission and processing of the sensor, resulting in measurement errors. In addition, it is necessary to combine industry standards and historical data feedback. The provisions and requirements for various performance indicators of water level sensors in industry standards provide a basic reference framework for evaluation; the problems and causes of failures of sensors recorded in historical data during actual use can reveal which factor indicators are the key factors leading to quality problems. For example, if historical data show that a certain model of sensor frequently increases the measurement error due to temperature changes, then the criticality of the temperature adaptability indicator will be correspondingly increased in the evaluation. Through a comprehensive analysis of these multi-faceted factors, a scientific evaluation method (such as hierarchical analysis method, fuzzy comprehensive evaluation method, etc.) is used to determine a reasonable criticality coefficient for each factor indicator, and finally form a set of factor indicator criticality coefficients, laying a solid foundation for subsequent accurate evaluation and quality control.

[0060] Determining the indicator feature association depth factor set based on the factor indicator criticality coefficient set is a process of reasonably allocating data association resources based on the criticality. After obtaining the factor indicator criticality coefficient set, each coefficient in the coefficient set is first analyzed in detail. For factor indicators with higher criticality coefficients, it means that they play a more critical role in reflecting the quality of water level sensors. For example, if the criticality coefficient of the factor indicator of water level measurement accuracy is at a high level, it means that it is crucial to the quality assessment of the entire sensor. Due to its high criticality, when determining the association depth factor, it is necessary to associate and train more related feature data, including the measurement error data of the sensor under different water level environments, the stability characteristics of the measurement data, and the performance data of the internal components of the sensor related to the measurement accuracy (such as the accuracy characteristics of the sensitive components, the resolution of the signal processing circuit, etc.), and may even involve external environmental factors (such as the impact of temperature and pressure changes on measurement accuracy) and other aspects of feature data. By associating these rich and deep feature data, the characteristics of the factor indicator of water level measurement accuracy can be more comprehensively and accurately characterized, thereby determining a larger association depth factor. On the contrary, for factor indicators with relatively low criticality coefficients, such as certain decorative features of the sensor housing (color uniformity, etc.), although they also have a certain impact on the overall quality of the sensor, it is relatively small. In this case, the associated feature data is relatively small, and may be mainly concentrated in the data directly related to the decorative feature obtained at the appearance inspection station, such as color detection data, surface texture data, etc., and the corresponding correlation depth factor is also small. By determining the correlation depth factor based on the criticality coefficient of the factor indicator, it can be ensured that in the subsequent data association process, more computing resources and data processing energy will be focused on key factor indicators, while also taking into account other factor indicators, thereby constructing a reasonable and efficient sensor factor indicator detection feature data set, providing a solid data foundation for subsequent accurate evaluation and quality control.

[0061] Data association of sensor quality detection feature data sets based on the indicator feature association depth factor set to obtain sensor factor indicator detection feature data sets is a process of accurate screening and deep data mining. First, according to the indicator feature association depth factor set, the depth and breadth of data required to be associated with each factor indicator are clarified. For factor indicators with large association depth factors, such as water level measurement accuracy, the system will go deep into the sensor quality detection feature data set and widely collect various data features related to it, including measurement error data under different water level heights, different ambient temperatures, and humidity conditions, as well as performance change data of key components inside the sensor (such as pressure sensor chips, signal conditioning circuits, etc.) during long-term operation, as well as calibration data and data transmission stability data closely related to measurement accuracy. By associating these multi-faceted and deep-level data features, the characteristics of the factor indicator of water level measurement accuracy can be comprehensively and meticulously characterized. For factor indicators with relatively small association depth factors, such as the color consistency of the sensor housing, the associated data is relatively limited, mainly focusing on color-related data obtained from the appearance inspection station, such as color spectrum data and color deviation data of different parts. In the data association process, data mining technology and algorithms are used to accurately extract data features related to each factor indicator from the huge sensor quality detection feature data set according to pre-set association rules. For data with time series characteristics, such as performance monitoring data of sensors in long-term operation, the temporal nature of the data is also considered to extract features that can reflect its dynamic change law. Finally, for each factor indicator, an exclusive data subset that closely surrounds its characteristics is constructed, namely, the sensor factor indicator detection feature data set. These data sets provide a highly targeted data foundation for subsequent branch evaluation training, so that each factor indicator can be accurately evaluated and model trained in its exclusive data environment, thereby improving the accuracy and reliability of the entire sensor quality evaluation.

[0062] In a possible implementation, the controlling the qualified product sorting of the sensor quality assessment parameter based on the qualified quality threshold includes:

[0063] Based on the qualified quality threshold, the sensor quality assessment parameters are marked for qualified product sorting to obtain a sensor qualified sorting result; according to the sensor qualified sorting result, a sensor sorting control strategy is determined; and a material unloading station is started, and the material unloading station grabs and sorts the water level sensor to be detected according to the sensor sorting control strategy.

[0064] Specifically, the sensor quality assessment parameters are marked for qualified products based on the pre-set qualified quality threshold. The system will carefully compare the sensor quality assessment parameters with the qualified quality threshold one by one. For example, for the measurement accuracy parameter of the water level sensor, if its evaluation value is within the accuracy range specified by the qualified quality threshold, such as the error does not exceed ±0.5%, the sensor is marked as possibly qualified; for the stability parameter, if its output fluctuation does not exceed the set threshold range within the specified test time, it is also included in the category of possible qualification. Through a comprehensive comparison of various quality assessment parameters, the system can accurately identify which sensors preliminarily meet the standards of qualified products and which sensors have quality problems, thereby obtaining the qualified sensor sorting results.

[0065] When the sorting control strategy is determined based on the qualified sorting results of the sensor, for qualified sensors, the sorting control strategy aims to ensure that they can enter the subsequent process smoothly and without damage. If the sensor is determined to be qualified after detection, the unloading station will gently grab it and place it on the unloading conveyor belt packaging table. During the whole process, the movement force and speed of the robot arm are precisely controlled to avoid collision or bumps on the sensor. At the same time, the transmission path of qualified products on the conveyor belt is also optimized to ensure that it flows smoothly to the packaging link, preparing for subsequent packaging and storage. For unqualified sensors, the sorting control strategy focuses on effective isolation and identification. Once the sensor is determined to be unqualified, the unloading station will quickly grab it and place it on the unqualified product recycling conveyor belt to strictly separate it from qualified products to prevent confusion. At the same time, during the entire sorting process, the system will record the unqualified sensors in detail and mark the specific reasons for their unqualifiedness, such as problems in a certain inspection station (painting, laser marking, CCD inspection, bolt height inspection or finished product inspection, etc.), so as to facilitate subsequent quality traceability and problem analysis, thereby helping to improve production processes and improve product quality.

[0066] After starting the unloading station, the unloading station quickly starts the grabbing and sorting of the water level sensor to be detected according to the sensor sorting control strategy. If the sensor sorting control strategy determines that the sensor is a qualified product, the robotic arm of the unloading station will move to the location of the sensor according to the preset precise path and appropriate force, and smoothly grab the sensor to ensure that no damage is caused to the sensor during the grabbing process. Subsequently, the robotic arm places the sensor accurately on the packaging table of the unloading conveyor belt so that it can enter the subsequent packaging process in an orderly manner. If the sensor is judged to be a defective product, the unloading station also responds quickly. The robotic arm grabs the defective sensor in a specific way and then places it on the defective product recycling conveyor belt to achieve effective separation from the qualified products. Throughout the process, the unloading station's actions are precise and efficient, and are strictly executed in accordance with the sorting control strategy, ensuring the smooth progress of the sensor sorting work.

[0067] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0068] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

[0069] This specification and the drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. Water level sensor intelligent quality detection and sorting system, characterized by: The system is connected in communication with a water level sensor intelligent quality detection and sorting device, and the system includes: A water level sensor detection program calling module, the water level sensor detection program calling module is used to perform a power-on reset operation on the water level sensor intelligent quality detection and sorting device, and call a water level sensor detection program according to the water level sensor intelligent quality detection and sorting device; A multi-dimensional quality detection data stream acquisition module, which is used to place the water level sensor to be detected into a fixture, activate the water level sensor detection program to perform linkage station quality detection on the water level sensor to be detected, and obtain the sensor multi-dimensional quality detection data stream; A quality detection fusion feature set acquisition module, wherein the quality detection fusion feature set acquisition module is used to construct a data processing multi-channel, and use the data processing multi-channel to process and fuse the multi-dimensional quality detection data stream of the sensor to obtain a sensor quality detection fusion feature set; A quality assessment parameter acquisition module, wherein the quality assessment parameter acquisition module is used to build a sensor quality adaptive evaluator, and perform quality assessment on the sensor quality detection fusion feature set based on the sensor quality adaptive evaluator to obtain sensor quality assessment parameters; A qualified product sorting control module, the qualified product sorting control module is used to set a qualified quality threshold according to the sensor quality application standard, and perform qualified product sorting control on the sensor quality assessment parameter based on the qualified quality threshold; The step of obtaining a multi-dimensional quality detection data stream of a sensor includes: Based on the water level sensor detection program, a sensor detection station set is started, wherein the sensor detection station set includes a painting station, a laser marking station, a CCD detection station, a bolt height detection station and a finished product detection station; According to the sensor quality application standard, obtaining the appearance design information and performance requirement target of the water level sensor to be detected; Based on the appearance design information and the performance requirement target, the detection parameters of the sensor detection station set are analyzed in sequence to obtain a station detection parameter set; The sensor detection station set performs linkage station quality detection on the water level sensor to be detected according to the station detection parameter set to obtain a multi-dimensional quality detection data stream of the sensor.

2. The water level sensor intelligent quality detection and sorting system according to claim 1, characterized in that: The construction of data processing multiple channels includes: According to the water level sensor detection program, a data processing channel architecture is designed, wherein the number of channels in the data processing channel architecture corresponds one-to-one to the station sequence in the water level sensor detection program; Performing data structured classification on each processing channel in the data processing channel architecture to obtain multi-channel data structured attributes; Performing a processing program analysis on each processing channel in the data processing channel architecture based on the multi-channel data structured attributes to obtain a multi-channel data processing node program set; Based on the multi-channel data processing node program set, each processing channel in the data processing channel architecture is configured with a processing program to construct the data processing multi-channel.

3. The water level sensor intelligent quality detection and sorting system according to claim 2, characterized in that: The step of obtaining a multi-channel data processing node assembly comprises: Defining a data channel processing node, wherein the data channel processing node includes a data preprocessing node and a feature extraction node; Based on the data preprocessing node, respectively perform preprocessing program analysis on the structured attributes of the multi-channel data to obtain a multi-channel data preprocessing program set; Using the feature extraction nodes to perform feature extraction program analysis on the multi-channel data structured attributes respectively to obtain a multi-channel data feature extraction program set; The multi-channel data processing node program set is determined based on the multi-channel data preprocessing program set and the multi-channel data feature extraction program set.

4. The water level sensor intelligent quality detection and sorting system according to claim 1, characterized in that: The obtaining of the sensor quality detection fusion feature set comprises: The data processing multi-channel is used to map and parallel process the multi-dimensional quality detection data stream of the sensor to obtain a multi-dimensional sensor quality detection feature set; Performing fusion demand analysis on the multi-dimensional sensor quality detection feature set to obtain feature fusion demand characteristic information; Based on the feature fusion requirement feature information and the data fusion algorithm list, feature matching is performed to obtain a target feature fusion algorithm; The multi-dimensional sensor quality detection feature set is subjected to feature fusion based on the target feature fusion algorithm to obtain the sensor quality detection fusion feature set.

5. The water level sensor intelligent quality detection and sorting system according to claim 1, characterized in that: The method of constructing a sensor quality adaptive evaluator comprises: Acquire a sensor quality detection feature data set by collecting and processing, wherein the sensor quality detection feature data set includes historical sensor quality detection fusion feature data and corresponding sensor quality result data; Obtain a set of quality assessment factor indicators of the water level sensor to be detected; Based on the sensor quality detection feature data set, data association is performed on each factor indicator in the quality assessment factor indicator set to obtain a sensor factor indicator detection feature data set; The sensor factor index detection feature data set is subjected to branch evaluation training and model integration fusion to build the sensor quality adaptive evaluator.

6. The water level sensor intelligent quality detection and sorting system according to claim 5, characterized in that: The step of obtaining a sensor factor index detection feature data set includes: Performing criticality assessment on each factor indicator in the quality assessment factor indicator set respectively to obtain a factor indicator criticality coefficient set; Determine the indicator feature correlation depth factor set according to the factor indicator criticality coefficient set; The sensor quality detection feature data sets are respectively data associated based on the indicator feature association depth factor set to obtain the sensor factor indicator detection feature data set.

7. The water level sensor intelligent quality detection and sorting system according to claim 1, characterized in that: The controlling of sorting qualified products of the sensor quality assessment parameter based on the qualified quality threshold comprises: Based on the qualified quality threshold, the sensor quality assessment parameter is marked as qualified products to obtain a sensor qualified sorting result; Determining a sensor sorting control strategy according to the qualified sensor sorting result; The unloading station is started, and the unloading station grabs and sorts the water level sensor to be detected according to the sensor sorting control strategy.

Images