Multi-sensor fusion automatic production line quality detection method

Abstract

The invention provides a multi-sensor fusion automatic production line quality detection method, and relates to the technical field of automatic production line quality detection, and the method comprises the following steps: selecting proper sensors according to the product characteristics and quality detection requirements of an automatic production line, and reasonably arranging the sensors; related data of a product is collected in real time through each sensor, and the data is transmitted to a data processing center for processing; fusing the data from different sensors by adopting a proper data fusion algorithm, and determining the weight of the data of each sensor according to the quality detection requirements of the product and the characteristics of the sensors; based on the fused data, establishing a quality detection model, and detecting and judging the quality of the product; and the quality detection system is optimized and adjusted regularly, so that the performance and the stability of the system are improved. The production efficiency and the product quality can be effectively improved, the production cost can be reduced, and remarkable economic benefits and social benefits are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of automated production line quality detection, and in particular to an automated production line quality detection method using multi-sensor fusion. Background Art

[0002] In modern industrial production, the application of automated production lines is becoming more and more widespread. In order to ensure product quality, it is necessary to conduct effective quality inspections on the products on the production line. And with the development of the manufacturing industry, the requirements for product quality are getting higher and higher.

[0003] Traditional manual inspection methods are inefficient, prone to errors, and cannot meet the needs of modern industrial production. In recent years, automated inspection systems have been widely used. A single type of sensor is difficult to fully and accurately detect quality problems of complex products. Secondly, some multi-sensor fusion inspection methods cannot be fully adapted, which affects the accuracy and reliability of the inspection results, thereby reducing the quality of the product. Therefore, it is of great significance to develop a method that can integrate multiple sensor data to achieve efficient and accurate automated production line quality inspection. Summary of the invention

[0004] 1. Technical issues to be solved

[0005] In view of the deficiencies in the prior art, the present invention provides a multi-sensor fusion automated production line quality inspection method, which solves the problems raised by the above-mentioned background technology.

[0006] (II) Technical solution

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: a multi-sensor fusion automated production line quality detection method, comprising the following steps:

[0008] Step 1: Sensor selection and layout:

[0009] According to the product characteristics and quality inspection requirements of the automated production line, select appropriate sensors and arrange them reasonably to ensure that all key parts and parameters of the product can be fully inspected;

[0010] Step 2: Data collection and preprocessing:

[0011] Collect relevant data of products in real time through various sensors and transmit the data to the data processing center for processing;

[0012] Step 3: Multi-sensor data fusion:

[0013] Use appropriate data fusion algorithms to fuse data from different sensors, and determine the weight of each sensor data based on product quality inspection requirements and sensor characteristics;

[0014] Step 4: Quality inspection and judgment:

[0015] Based on the fused data, a quality inspection model is established to inspect and judge the quality of the product, and whether the product is qualified is determined based on the output results of the quality inspection model;

[0016] Step 5: System optimization and maintenance:

[0017] The quality inspection system is optimized and adjusted regularly to improve the performance and stability of the system. At the same time, the quality inspection results are fed back to production managers to optimize and improve the production process.

[0018] Preferably, in step one, the sensor includes a visual sensor, a force sensor, a temperature sensor, a humidity sensor and a sound sensor.

[0019] Preferably, the visual sensor is installed at a key position of the production line to detect appearance defects of the product; the force sensor is installed at a key position of the production equipment to detect the force condition of the product; the temperature sensor is installed near the equipment or product that needs to monitor the temperature to detect the temperature change of the product; the humidity sensor is installed in the production link that is sensitive to humidity to detect the impact of ambient humidity on product quality; the sound sensor is installed at a key position of the equipment operation to detect the equipment operation status and abnormal sound.

[0020] Preferably, in step 2, the data processing center preprocesses the collected data, including data filtering, denoising and normalization; specifically, the data filtering uses a low-pass filter to remove high-frequency noise and ensure the smoothness of the data; the denoising process uses a median filtering algorithm to eliminate random noise in the data; the normalization process scales the data to a uniform numerical range to facilitate subsequent data fusion and analysis.

[0021] Preferably, in step three, the data fusion algorithm includes weighted averaging method, Kalman filtering method and Bayesian estimation method; specifically, the weighted averaging method is used to fuse numerical data from different sensors, and the weighted average value is calculated by assigning corresponding weights to different sensor data; the Kalman filtering method is suitable for fusing data with time series characteristics, and can effectively reduce the impact of noise on data; the Bayesian estimation method uses prior knowledge and current observation data to perform probabilistic inference on product quality, thereby improving the accuracy of detection.

[0022] Preferably, in step four, the quality detection model provides technical support through a machine learning algorithm, which includes a support vector machine and a neural network; specifically, the support vector machine is used to process linearly separable data, and divides the data into two categories: qualified and unqualified by finding the optimal hyperplane; the neural network, through a multi-layer perceptron structure, can handle nonlinear problems, and automatically extract features and make quality judgments by learning a large amount of sample data.

[0023] Preferably, in step 4, if the product is qualified, the next process is continued; if the product is unqualified, corresponding processing is performed, including rework and scrapping.

[0024] Preferably, in step five, the system optimization and maintenance also includes calibration and updating of sensors to ensure the detection accuracy and reliability of the sensors; at the same time, the algorithms of the data processing center are regularly upgraded to adapt to new situations and new requirements that may arise in the production process.

[0025] (III) Beneficial effects

[0026] The present invention provides a multi-sensor fusion automated production line quality detection method, which has the following beneficial effects:

[0027] 1. By integrating data from multiple sensors, the quality status of the product can be more comprehensively reflected, thereby improving detection accuracy and reliability. System optimization and maintenance ensure that sensors and data processing algorithms can adapt to new situations and new requirements that may arise in the production process, making the quality inspection system have good adaptability and scalability.

[0028] 2. Automated quality inspection methods reduce manual intervention, improve the operating efficiency of the production line, shorten the time from product production to completion of inspection, and through timely feedback of quality inspection results, production management personnel can quickly adjust the production process, reduce the generation of defective products, and thus reduce production costs.

[0029] In summary, the multi-sensor fusion automated production line quality inspection method of the present invention can not only effectively improve production efficiency and product quality, but also reduce production costs, and has significant economic and social benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a flow chart of the quality detection method of the present invention. DETAILED DESCRIPTION

[0031] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0032] Embodiment 1:

[0033] like Figure 1 As shown, an embodiment of the present invention provides a multi-sensor fusion automated production line quality detection method, comprising the following steps:

[0034] Step 1: Sensor selection and layout:

[0035] According to the product characteristics and quality inspection requirements of the automated production line, select appropriate sensors and arrange them reasonably to ensure that all key parts and parameters of the product can be fully inspected;

[0036] Step 2: Data collection and preprocessing:

[0037] Collect relevant data of products in real time through various sensors and transmit the data to the data processing center for processing;

[0038] Step 3: Multi-sensor data fusion:

[0039] Use appropriate data fusion algorithms to fuse data from different sensors, and determine the weight of each sensor data based on product quality inspection requirements and sensor characteristics;

[0040] Step 4: Quality inspection and judgment:

[0041] Based on the fused data, a quality inspection model is established to inspect and judge the quality of the product, and whether the product is qualified is determined based on the output results of the quality inspection model;

[0042] Step 5: System optimization and maintenance:

[0043] The quality inspection system is optimized and adjusted regularly to improve the performance and stability of the system. At the same time, the quality inspection results are fed back to production managers to optimize and improve the production process.

[0044] Embodiment 2:

[0045] like Figure 1 As shown, an embodiment of the present invention provides a multi-sensor fusion automated production line quality detection method, comprising the following steps:

[0046] Step 1: Sensor selection and layout:

[0047] According to the product characteristics and quality inspection requirements of the automated production line, select appropriate sensors and arrange the sensors reasonably to ensure that all key parts and parameters of the product can be fully inspected.

[0048] In this embodiment, the selection and arrangement of sensors are based on the specific product characteristics and quality inspection requirements of the production line. For example, for the electronic product assembly line, visual sensors are installed at key locations to detect the placement accuracy of components and the quality of solder joints; force sensors are installed at key locations of automatic placement machines and welding equipment to monitor whether the pressure applied by the equipment meets the process requirements; temperature sensors are installed near the reflow oven to ensure that the temperature during the welding process is controlled within the specified range; humidity sensors are installed in areas where the ambient humidity needs to be controlled to prevent component damage caused by excessively high or low humidity; sound sensors are installed at the operating locations of key equipment to monitor the equipment operating status and detect abnormal sounds in a timely manner.

[0049] Step 2: Data collection and preprocessing:

[0050] The relevant data of the product is collected in real time through various sensors and transmitted to the data processing center for processing.

[0051] In this embodiment, each sensor collects relevant data of the product in real time and transmits the data to the data processing center. The data processing center preprocesses the collected data, including data filtering, denoising and normalization. Data filtering uses a low-pass filter to remove high-frequency noise and ensure the smoothness of the data; denoising processing uses a median filter algorithm to eliminate random noise in the data; normalization processing scales the data to a uniform numerical range to facilitate subsequent data fusion and analysis.

[0052] Step 3: Multi-sensor data fusion:

[0053] Use appropriate data fusion algorithms to fuse data from different sensors, and determine the weight of each sensor data based on the product quality inspection requirements and the characteristics of the sensor.

[0054] In this embodiment, the weighted average method is used to fuse the numerical data from different sensors. By assigning corresponding weights to different sensor data, the weighted average is calculated. In addition, for data with time series characteristics, the Kalman filter method is used for fusion to effectively reduce the impact of noise on the data. The Bayesian estimation rule uses prior knowledge and current observation data to make probabilistic inferences on product quality, thereby improving the accuracy of detection.

[0055] Step 4: Quality inspection and judgment:

[0056] Based on the fused data, a quality inspection model is established to inspect and judge the quality of the product, and whether the product is qualified is determined based on the output results of the quality inspection model.

[0057] In this embodiment, the quality inspection model provides technical support through machine learning algorithms, including support vector machines and neural networks. Support vector machines are used to process linearly separable data, and divide data into qualified and unqualified categories by finding the optimal hyperplane; neural networks can handle nonlinear problems through multi-layer perceptron structures, and automatically extract features and make quality judgments by learning a large amount of sample data. If the product is qualified, it will continue to the next process; if the product is unqualified, it will be processed accordingly, including rework and scrapping.

[0058] Step 5: System optimization and maintenance:

[0059] The quality inspection system is optimized and adjusted regularly to improve the performance and stability of the system. At the same time, the quality inspection results are fed back to production managers to optimize and improve the production process.

[0060] In this embodiment, system optimization and maintenance also includes calibration and updating of sensors to ensure the detection accuracy and reliability of sensors; at the same time, the algorithm of the data processing center is regularly upgraded to adapt to new situations and new requirements that may arise in the production process. In addition, the quality inspection results will be fed back to the production management personnel to optimize and improve the production process, reduce the generation of defective products, and thus reduce production costs.

[0061] In summary, the multi-sensor fusion automated production line quality inspection method of the present invention can not only effectively improve production efficiency and product quality, but also reduce production costs, and has significant economic and social benefits.

[0062] Experimental example: Quality inspection of the automated production line of automotive parts, as follows:

[0063] Sensor selection and layout: In the automated production line of automotive parts, visual sensors, force sensors and temperature sensors are selected. Visual sensors are installed at key locations of the production line to detect appearance defects of parts; force sensors are installed on stamping equipment to detect the stress of parts; temperature sensors are installed on welding equipment to detect temperature changes during welding.

[0064] Data collection and preprocessing: Each sensor collects relevant data of components in real time and transmits the data to the data processing center. The collected data is filtered, denoised and normalized to improve the quality and reliability of the data.

[0065] Multi-sensor data fusion: The data from the visual sensor, force sensor, and temperature sensor are fused using the weighted average method. The weight of each sensor data is determined based on the accuracy and reliability of the sensor.

[0066] Quality inspection and judgment: Based on the fused data, a support vector machine quality inspection model is established to inspect and judge the quality of automotive parts. If the parts are qualified, the next process will be continued; if the parts are unqualified, they will be reworked or scrapped.

[0067] System optimization and maintenance: Regularly optimize and adjust the quality inspection system, adjust the sensor layout and parameters according to actual production conditions, and optimize the data fusion algorithm and quality inspection model. At the same time, regularly maintain and calibrate the sensors to ensure their accuracy and reliability.

[0068] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. The multi-sensor fusion automated production line quality inspection method is characterized by: The following steps are involved: Step 1: Sensor selection and layout: According to the product characteristics and quality inspection requirements of the automated production line, select appropriate sensors and arrange them reasonably to ensure that all key parts and parameters of the product can be fully inspected; Step 2: Data collection and preprocessing: Collect relevant data of products in real time through various sensors and transmit the data to the data processing center for processing; Step 3: Multi-sensor data fusion: Use appropriate data fusion algorithms to fuse data from different sensors, and determine the weight of each sensor data based on product quality inspection requirements and sensor characteristics; Step 4: Quality inspection and judgment: Based on the fused data, a quality inspection model is established to inspect and judge the quality of the product, and whether the product is qualified is determined based on the output results of the quality inspection model; Step 5: System optimization and maintenance: The quality inspection system is optimized and adjusted regularly to improve the performance and stability of the system. At the same time, the quality inspection results are fed back to production managers to optimize and improve the production process.

2. The multi-sensor fusion automated production line quality detection method according to claim 1, characterized in that: In step one, the sensors include a visual sensor, a force sensor, a temperature sensor, a humidity sensor and a sound sensor.

3. The multi-sensor fusion automated production line quality detection method according to claim 2, characterized in that: The visual sensor is installed at a key position of the production line to detect appearance defects of the product; the force sensor is installed at a key position of the production equipment to detect the force condition of the product; the temperature sensor is installed near the equipment or product that needs to monitor the temperature to detect the temperature change of the product; the humidity sensor is installed in the production link that is sensitive to humidity to detect the impact of ambient humidity on product quality; the sound sensor is installed at a key position of the equipment operation to detect the equipment operation status and abnormal sound.

4. The multi-sensor fusion automated production line quality detection method according to claim 1, characterized in that: In step 2, the data processing center preprocesses the collected data, including data filtering, denoising and normalization; specifically, the data filtering uses a low-pass filter to remove high-frequency noise and ensure the smoothness of the data; the denoising process uses a median filter algorithm to eliminate random noise in the data; the normalization process scales the data to a uniform numerical range to facilitate subsequent data fusion and analysis.

5. The multi-sensor fusion automated production line quality detection method according to claim 1, characterized in that: In step 3, the data fusion algorithm includes weighted average method, Kalman filter method and Bayesian estimation method; specifically, the weighted average method is used to fuse numerical data from different sensors, and the weighted average value is calculated by assigning corresponding weights to different sensor data; the Kalman filter method is suitable for fusing data with time series characteristics, and can effectively reduce the impact of noise on data; The Bayesian estimation rule utilizes prior knowledge and current observation data to make probabilistic inferences about product quality, thereby improving the accuracy of detection.

6. The multi-sensor fusion automated production line quality detection method according to claim 1, characterized in that: In step four, the quality detection model provides technical support through a machine learning algorithm, which includes a support vector machine and a neural network; specifically, the support vector machine is used to process linearly separable data, and divides the data into two categories: qualified and unqualified by finding the optimal hyperplane; the neural network, through a multi-layer perceptron structure, can handle nonlinear problems, and automatically extract features and make quality judgments by learning a large amount of sample data.

7. The multi-sensor fusion automated production line quality inspection method according to claim 1, characterized in that: In step 4, if the product is qualified, the next process will be carried out; if the product is unqualified, corresponding processing will be carried out, including rework and scrapping.

8. The multi-sensor fusion automated production line quality detection method according to claim 1, characterized in that: In step five, the system optimization and maintenance also includes calibration and updating of sensors to ensure the detection accuracy and reliability of the sensors; at the same time, the algorithms of the data processing center are regularly upgraded to adapt to new situations and new requirements that may arise in the production process.

Images