WO2020133954A1

WO2020133954A1 - Digital performance analysis method and device

Info

Publication number: WO2020133954A1
Application number: PCT/CN2019/091610
Authority: WO
Inventors: 徐东; 吴正
Original assignee: 帝沃检测技术(上海)有限公司
Priority date: 2018-12-27
Filing date: 2019-06-17
Publication date: 2020-07-02
Also published as: CN109613108B; CN109613108A

Abstract

Disclosed are a digital performance analysis method and device. The method comprises: acquiring a set of sample data of products containing no metal and at least one set of sample data of products containing metal; according to the distribution law of the sample data of the products containing no metal and the sample data of the products containing metal, calculating a product fall-out ratio set corresponding to the sample data of the products containing no metal and an omission ratio set corresponding to each set of sample data of the products containing metal; selecting a corresponding value as a threshold according to the product fall-out ratio set, the omission ratio set corresponding to each set of sample data of the products containing metal and the requirements of the fall-out ratio and omission ratio. By means of the method and device, the selection of a threshold is supported by effective quantized data, and a selected threshold is more in tune with actual usage requirements, thereby reducing the fall-out ratio, ensuring the omission ratio and improving the monitoring of product quality.

Description

Digital performance analysis method and equipment

Technical field

The invention relates to the field of metal detectors, in particular to a digital performance analysis method and equipment.

Background technique

Now many fields (for example: food, pharmaceutical, rubber, chemical, etc.) will monitor the metal foreign objects that may be mixed into the product during the production process. The principle is that when a certain size of metal foreign objects are mixed in the product, the metal detector will produce A larger signal is used by the user to set the different signal thresholds to allow the metal detector to monitor and judge metal foreign objects.

For example, the signal value of a normal product may jump within 100, when mixed with a certain size of metal foreign objects (such as 1.5mm), it will produce a larger signal that may be 300 or even 500. At this time, if the set threshold is 300, then any product whose signal value exceeds 300 is considered unqualified, that is, the metal detector determines that it has metal; at the same time, due to various reasons, even if the same foreign object passes repeatedly For metal detectors, the signal value generated each time is also different.

The current threshold setting is usually set according to the user's experience, but this threshold selected based on experience is different due to the uncertainty of the metal detector, the use environment and the feeling of each person. The influence of factors that are difficult to quantify, there is great instability, there is no clear and quantifiable basis, and it cannot meet the needs of high-quality monitoring with low false detection rate and low missed detection rate.

Summary of the invention

The purpose of the present invention is to provide a digital performance analysis method and equipment, which selects a threshold based on reliable quantitative data, which not only meets the needs of low false detection rate but also realizes scientific monitoring of low missed detection rate.

The technical solutions provided by the present invention are as follows:

A digital performance analysis method includes: acquiring a set of metal-free product sampling data and at least one set of metal-containing product sampling data; based on the metal-free product sampling data and the metal-containing product sampling data Distribution rule, calculate the product misdetection rate group corresponding to the metal-free product sampling data and the missing detection rate group corresponding to each group of metal-containing product sampling data; according to the product misdetection rate group, each group contains Metal product sampling data corresponding to the requirements of the missed detection rate group and false detection missed detection rate, select the corresponding value as the threshold.

In the above technical solutions, according to the distribution rules of sampling data of metal-free products and sampling data of metal-containing products, appropriate thresholds are selected according to the requirements of false detection rate and missed detection rate, so that the selection of thresholds is supported by effective quantitative data , The selected threshold is more in line with the actual use requirements, reduce the false detection rate, missed detection rate, improve the monitoring efficiency of product quality and the quality of quality management work.

Further, according to the distribution rule of the metal-free product sampling data and the metal-containing product sampling data, the product misdetection rate group corresponding to the metal-free product sampling data and each group of metal-containing product sampling are calculated The missing detection rate group corresponding to each data includes: when the sampling data of the metal-free product and the sampling data of the metal-containing product conform to the normal distribution law, according to a preset normal formula, a preset limit parameter, the Sampling data of metal-free products and sampling data of each of the metal-containing products to calculate the product misdetection rate group corresponding to the metal-free product sampling data and the corresponding missing detection of each group of metal-containing product sampling data Rate group.

In the above technical solution, when the normal distribution law is met, a preset normal formula may be used to calculate the missed detection rate group and the false detection rate group.

Further, according to the preset normal formula, the preset limit parameter, the metal-free product sampling data and each of the metal-containing product sampling data, the corresponding data of the metal-free product sampling data is calculated The product false detection rate group and the corresponding missing detection rate group of each group of metal-containing product sampling data include: according to the metal-free product sampling data, calculate the corresponding product (signal) average and product standard deviation; For each group of metal-containing product sampling data, calculate the metal (signal) average and metal standard deviation corresponding to each group of metal-containing product sampling data; according to the product average, the product standard deviation, The preset limit parameter and the preset normal formula to calculate the product false detection rate group corresponding to the metal-free product sampling data; according to the preset limit parameter, the preset normal formula and each group of the The metal mean value and metal standard deviation corresponding to the metal-containing product sampling data are respectively calculated to obtain the missing detection rate group corresponding to each group of metal-containing product sampling data.

In the above technical solution, the detailed calculation method when each sampled data conforms to the normal distribution law is clarified, which ensures the accuracy of the data and further ensures the rationality of the threshold.

Further, the preset limit parameters include: a minimum threshold, a maximum threshold, and a threshold interval value.

In the above technical solution, the setting of the maximum/small threshold can eliminate unnecessary data calculation and reduce the amount of calculation.

Further, according to the product misdetection rate group, each group of metal-containing product sampling data corresponding to the respective missed detection rate group and false detection missed detection rate requirements, selecting the corresponding value as the threshold includes: according to each group containing Metal product sampling data corresponding to the missed detection rate group and the product misdetection rate group, draw a corresponding graph; according to the false detection missed detection rate requirements, select the corresponding value from the graph as a threshold.

In the above technical solution, it is presented in the form of a graph, so that the user can understand the reason of threshold selection more intuitively and clearly.

Further, the step S1 obtaining a set of metal-free product sampling data and at least one set of metal-containing product sampling data includes: when a certain time interval is reached, obtaining a set of metal-free product sampling data and at least one set of Metal product sampling data; Step S2 calculates the product false detection rate corresponding to the metal-free product sampling data according to the distribution rule of the metal-free product sampling data and the metal-containing product sampling data The missing detection rate group corresponding to the sampling data of each group and each group of metal-containing product sampling data includes: according to the distribution rule of the sampling data of the metal-free product and the sampling data of the metal-containing product obtained at a certain time interval, the calculated Describe the product false detection rate group corresponding to the metal-free product sampling data and the corresponding missing detection rate group for each group of metal-containing product sampling data.

In the above technical solution, the sampling data is periodically obtained, and the threshold value is periodically monitored, so that the threshold value of the metal detector meets actual production requirements.

Further, the step S3 selects the corresponding value as the threshold according to the requirements of the product misdetection rate group, each group of metal-containing product sampling data corresponding to the missed detection rate group and the false detection missed detection rate, and further includes: The set threshold meets the requirements of the false detection and miss detection rate, this threshold is still used; when the set threshold does not meet the requirements of the false detection and miss detection rate, according to the product false detection rate group, each group contains metal The product sampling data corresponding to the requirements of the missed detection rate group and the false detection missed detection rate, re-select the corresponding value to update or prompt to update the threshold.

In the above technical solution, when the set threshold does not meet the current situation, it can be adjusted so that the detection result of the metal detector always meets the production requirements.

The present invention also provides a digital performance analysis device, including: an acquisition module for acquiring a set of metal-free product sampling data and at least one set of metal-containing product sampling data; a calculation module for calculating according to the metal-free product Product sampling data and the distribution law of the metal-containing product sampling data, the product misdetection rate group corresponding to the metal-free product sampling data and the corresponding missing detection rate of each group of metal-containing product sampling data are calculated Group; a selection module, used to select the corresponding value as the threshold according to the requirements of the product false detection rate group, the corresponding leakage detection rate group and the false detection leakage detection rate of each group of metal-containing product sampling data.

In the above technical solution, according to the distribution rules of the sampling data of metal-free products and the sampling data of metal-containing products, an appropriate threshold is selected according to the requirements of false detection and missed detection rate, so that the selection of the threshold is supported by effective quantitative data. The threshold is more in line with the actual use requirements, reducing the false detection rate and missed detection rate, and improving the monitoring of product quality.

Further, the calculation module is configured to calculate a product false detection rate group corresponding to the metal-free product sampling data according to the distribution rule of the metal-free product sampling data and the metal-containing product sampling data The missing detection rate group corresponding to each group of metal-containing product sampling data includes: the calculation module, when the metal-free product sampling data and the metal-containing product sampling data conform to the normal distribution law, according to A preset normal formula, a preset limit parameter, the sampling data of the metal-free product and the sampling data of each metal-containing product, and the product misdetection rate group corresponding to the sampling data of the metal-free product is calculated Each group of metal-containing product sampling data corresponds to the corresponding missed detection rate group.

Further, the calculation module calculates the metal-free product sampling data according to a preset normal formula, a preset limit parameter, the metal-free product sampling data and each metal-containing product sampling data The corresponding product false detection rate group and the corresponding missing detection rate group for each group of metal-containing product sampling data include: the calculation module, calculating the corresponding product average and product standard based on the metal-free product sampling data Difference; and, according to each group of the metal-containing product sampling data, calculate the metal mean and metal standard deviation corresponding to each group of the metal-containing product sampling data; and, according to the product average, the product The standard deviation, the preset limit parameter and the preset normal formula to calculate the product false detection rate group corresponding to the metal-free product sampling data; and, according to the preset limit parameter and the preset normal formula The metal mean value and metal standard deviation corresponding to the sampling data of each group of metal-containing products are respectively calculated, and the missed detection rate group corresponding to the sampling data of each group of metal-containing products is calculated respectively.

Further, the selection module is configured to select the corresponding value as the threshold according to the requirements of the product false detection rate group, the corresponding missing detection rate group and the false detection leakage detection rate of each group of metal-containing product sampling data, including: The drawing sub-module is used to draw the corresponding curve graph according to the corresponding missing detection rate group and the product false detection rate group of each group of metal-containing product sampling data; It is required to select the corresponding value from the graph as the threshold.

Further, it also includes: a camera module for real-time acquisition of the working state of the metal detector; the acquisition module is further used for identifying a set of metal-free product sampling data based on the real-time acquired working state of the metal detector and Sampling data for at least one set of metal-containing products.

In the above technical solution, the working state of the metal detector is obtained through the camera module to obtain the sampling data, the application range is wider, and the compatibility is higher.

Further, the acquisition module is further configured to acquire a set of product sampling data containing no metal and at least one set of product sampling data containing metal when a certain time interval is reached.

Further, the selection module is further configured to use the threshold when the set threshold meets the requirements of the false detection and miss detection rate; and, when the set threshold does not meet the requirements of the false detection and miss detection rate, Then, according to the requirements of the product misdetection rate group, each group of metal-containing product sampling data corresponding to the missed detection rate group and the false detection missed detection rate, the corresponding value is reselected to update or prompt to update the threshold.

Compared with the prior art, the beneficial effects of the digital performance analysis method and device of the present invention are:

The invention selects a suitable threshold value according to the requirements of the false detection rate and the missed detection rate according to the distribution rules of the metal-free product sampling data and the metal-containing product sampling data, so that the selection of the threshold is supported by effective quantitative data, and the selected threshold It is more in line with the actual use requirements, reduces the false detection rate, controls the missed detection rate, and improves product quality monitoring.

BRIEF DESCRIPTION

In the following, the preferred embodiments will be described in a clear and easy-to-understand manner with reference to the drawings, and the above-mentioned characteristics, technical characteristics, advantages and implementation of a digital performance analysis method and device will be further described.

1 is a flowchart of an embodiment of the digital performance analysis method of the present invention;

2 is a flowchart of another embodiment of the digital performance analysis method of the present invention;

FIG. 3 is an interface display diagram of an embodiment of sampling data and preset limit parameters of the present invention;

4 is a graph of the present invention drawn according to the product misdetection rate group and the missed detection rate group corresponding to three sets of metal-containing product sampling data;

5 is a schematic diagram of an embodiment of the lowest limit value when selecting a threshold;

6 is a schematic diagram of the highest limit value when selecting the threshold value in FIG. 5;

7 is a schematic structural diagram of an embodiment of a digital performance analysis device of the present invention;

8 is a schematic structural diagram of another embodiment of the digital performance analysis device of the present invention;

9 is a flowchart of another embodiment of the digital performance analysis method of the present invention.

Description of Drawing Symbols:

10. Acquisition module, 20. Calculation module, 30. Selection module, 31. Drawing submodule, 32. Selection submodule, 40. Camera module.

detailed description

In order to more clearly explain the embodiments of the present invention or the technical solutions in the prior art, the specific embodiments of the present invention will be described below with reference to the drawings. Obviously, the drawings in the following description are only some embodiments of the present invention. For a person of ordinary skill in the art, without paying any creative labor, other drawings can be obtained from these drawings and obtained Other implementations.

In order to make the drawings simple, the figures only show the parts related to the present invention, they do not represent the actual structure of the product. In addition, in order to make the drawings concise and easy to understand, in some drawings, components having the same structure or function are only schematically illustrated, or only one of them is marked. In this article, "one" not only means "only this one", but also means "more than one".

The working principle of the metal detector is electromagnetic induction technology. There is a specially designed electromagnetic coil inside the metal detector. When the metal ball passes, it will cause changes in the electromagnetic field. The receiving circuit part detects this weak signal change, combined with a certain technology Means, through certain analysis and processing, the final determination of whether there is metal foreign matter in the product, and the threshold setting is the basis for the metal detector judgment. The false detection rate refers to the rate at which normal products are judged to contain metals when the threshold is too low. The missed detection rate is the rate at which products containing metals are judged to be normal products when the threshold is too high, thus failing to detect. Therefore, the reasonable setting of the threshold is very important.

With the requirement of higher and higher sensitivity, the set threshold is getting lower and lower. This threshold has been approaching the normal signal value area of the product, such as 200, 190, 180, 130, 120, 100, 90, etc. Threshold After a gradual decrease, it will lead to an increase in false detections. Excessively high false detection rates will bring greater economic losses. The selection of thresholds based on experience alone cannot meet production needs and improve quality.

The digital performance analysis method is equivalent to an arithmetic program, which can be installed on the metal detector, or can be installed on another independent device to serve each metal detector.

FIG. 1 shows an embodiment of the present invention. A digital performance analysis method includes:

S101 Obtain a set of sampling data of products containing no metal and at least one set of sampling data of products containing metal.

Specifically, the sampling data of products without metal refers to the data obtained by the product itself passing through the metal detector.

The choice of threshold is closely related to the application. Therefore, the number of samples of metal-containing product sampling data depends on the number of metals to be detected. Sampling data of metal-containing products refers to the data contained in the products containing related metal-over-metal detectors.

For example: if you want to detect the presence of iron in the product, you can obtain a set of sampling data of the product containing iron, the specific method is that the product contains an iron ball through the metal detector; or, you want to detect whether there is iron or non-iron in the product Metal, stainless steel, you need to obtain three sets of metal-containing product sampling data, respectively ferrous, non-ferrous metals and stainless steel, the implementation method is to put three different metals into the product to test the corresponding groups of metal-containing Product sampling data.

The number of sampled data in each group is set according to the actual accuracy, and can be any number. For example, as shown in Figure 3, there are 4 sets of sampling data, one of which is the product sampling data of metal-free products, and the other three are the sampling data of metal-containing products, each of which has 100 data.

Generally, when testing the sampling data of products containing metals, metal balls (for example: 1.5mm diameter iron beads, 2.0mm non-ferrous metal beads, 2.5mm stainless steel beads) are put into the product when they pass the metal detection Due to various complicated and unavoidable objective factors, such as speed when passing, relative machine position, running track, conveyor belt, vibration, product temperature, ambient temperature, power supply voltage, atmospheric pressure, static electricity, workshop The complex electromagnetic environment, etc., causes the signal to be almost the same every time the metal ball passes, therefore, the number of sampled data should be as much as possible, for example: set 200, 150, etc. in each group to improve Accuracy.

It should be noted that due to the current technical reasons, the detection of metal detectors will be different due to the different shapes of metal foreign objects. For example: a wire with a diameter of 1mm and a length of 10mm may be detected at an angle , But it may not be detected after rotating a certain angle (for example: 90 degrees), so it is recommended to use the sphere as a standard.

The sampled data is obtained by the metal detector. When the program corresponding to the digital performance analysis method runs on the metal detector, it can be called directly; when this program is run on equipment other than the metal detector, it can be sent by the metal detector. It can also be identified by the camera module during the monitoring process of the metal detector. The specific acquisition method of sampling data is not limited here.

S102 calculates the product misdetection rate group corresponding to the metal-free product sampling data and the missing detection corresponding to each group of metal-containing product sampling data according to the distribution rules of the metal-free product sampling data and the metal-containing product sampling data Rate group.

Specifically, after obtaining the sampling data of metal-free products and the sampling data of each group of metal-containing products, analyze and determine the distribution rule of the sampling data, and then select the appropriate formula to calculate the product misdetection rate group and the missed detection rate group. Choose the appropriate threshold to lay the foundation.

The product misdetection rate group is composed of the product misdetection rate corresponding to the sampling data of products containing no metal at different thresholds; a group of missed detection rate group refers to the corresponding missed detection of a group of metal-containing product sampling data at different thresholds Rate composition.

S103 selects the corresponding value as the threshold according to the product misdetection rate group, the corresponding misdetection rate group corresponding to each group of metal-containing product sampling data, and the actual misdetection miss detection rate requirements.

Specifically, the false detection and missed detection rate requirements are determined according to the actual situation, for example: the false detection rate is required to be less than 0.1%, and the missed detection rate is also required to be less than 0.1%. Choose the value that satisfies less than 0.1% at the same time as the threshold.

Most of the time, it is a value range that meets the requirements for the false detection rate and the missed detection rate. At this time, a value in this value range can be arbitrarily selected as the threshold.

Preferably, S103 selects the corresponding value as the threshold according to the requirements of the product misdetection rate group, the corresponding misdetection rate group and the misdetection miss rate of each metal-containing product sampling data, including:

Draw the corresponding graphs according to the corresponding missed detection rate group and product misdetection rate group of each group of metal-containing product sampling data; according to the false detection missed detection rate requirements, select the corresponding value from the graph as the threshold.

Specifically, the calculated missed detection rate groups and product misdetection rate groups are drawn into corresponding graphs, so that users can intuitively understand the false detection rate and missed detection rate corresponding to each value, and understand the threshold value conveniently and quickly. Choose a reason.

As shown in Figures 5 and 6, the product false detection rate group is drawn to the left curve, and the three groups of metal-containing product sampling data corresponding to the missing detection rate group are drawn to the right three curves, when the false detection rate is required to be below 1‰ It can be seen from Figures 5 and 6 that as long as the t value is set between 233 and 327, the product false detection rate of less than 1‰ and the metal leakage detection rate can be ensured at the same time to ensure good production. When considering quality as the first factor, it should be as close as possible to 233, which not only ensures that the false detection rate of the product will not exceed 1‰, but also ensures the maximum detection capability of the metal detector; when the production is due to interference, temperature, site When the electromagnetic environment begins to slowly affect the device and the signal increases, the threshold can be increased appropriately while ensuring the safe detection rate of the metal, as long as it does not exceed 327, that is, the adjustable range of the site is allowed to be 233-327, rather than relying on people The sense that the quantified numerical range improves quality control standards and quality.

In this embodiment, according to the distribution rules of sampling data of metal-free products and sampling data of metal-containing products, an appropriate threshold is selected according to the requirements of the false detection rate and the missing detection rate, so that the selection of the threshold is supported by effective quantitative data. The threshold value is more in line with the actual use requirements, reducing the false detection rate and missed detection rate, and improving the more accurate monitoring of product quality.

FIG. 2 shows another embodiment of the present invention. A digital performance analysis method includes:

S201 acquires a set of sampling data of products containing no metal and at least one set of sampling data of products containing metal.

S202 calculates the product false detection rate group corresponding to the metal-free product sampling data and the missed detection corresponding to each group of metal-containing product sampling data according to the distribution rules of the metal-free product sampling data and the metal-containing product sampling data Rate groups include:

S212 When the sampling data of the product containing no metal and the sampling data of the product containing metal conform to the normal distribution law, according to the preset normal formula and the preset limit parameters (the preset limit parameters include: minimum threshold, maximum threshold and threshold interval value) ), metal-free product sampling data and each metal-containing product sampling data, calculate the product misdetection rate group corresponding to the metal-free product sampling data and the corresponding missing detection rate group for each group of metal-containing product sampling data .

Specifically, the preset limit parameter refers to the threshold range of the metal detector, and an appropriate minimum threshold, maximum threshold, and threshold interval value are set according to actual detection requirements.

For example, if the metal to be detected is stainless steel and iron, as shown in FIG. 3, the maximum threshold can be set to 1500, the minimum threshold is 5, and the threshold interval value is 1. Of course, these data can be flexibly changed, for example: the maximum threshold is set to 1400, the minimum threshold is 10, the threshold interval value is 3 and so on. If the maximum threshold exceeds the value corresponding to the largest metal block, it is meaningless, and only unnecessary calculations are done. Therefore, the corresponding preset limit parameters can be set according to the actual metal detection requirements.

The normal distribution law has its corresponding statistical formula, which can be set as a preset normal formula for subsequent calculations.

Preferably, S212 calculates according to a preset normal formula, a preset limit parameter (the preset limit parameter includes: a minimum threshold, a maximum threshold, and a threshold interval value), sampling data of metal-free products and sampling data of each metal-containing product The product misdetection rate group corresponding to the metal-free product sampling data and the corresponding missing detection rate group for each group of metal-containing product sampling data include:

S2121 calculates the corresponding product average and product standard deviation based on the sampling data of products that do not contain metals;

S2122 calculates the metal average and metal standard deviation corresponding to each group of metal sampling data according to the metal sampling data of each group.

Specifically, the average value is the sum of the sampled data divided by the total number.

For example, if there are 200 data in a set of sample data of metal-free products, add the 200 data and divide by 200 to get the average value of the product. Similarly, the average value of metal corresponding to the sampling data of each group of metal-containing products.

The standard deviation is calculated according to the following formula:

σ is the standard deviation, μ is the average value, μ _i is the ith data in a group of product sampling data, and N is the total number of a group of product sampling data.

S2123 calculates the product false detection rate group corresponding to the metal-free product sampling data based on the product average, product standard deviation, preset limit parameters and preset normal formula;

S2124 calculates the leak detection rate group corresponding to each group of metal-containing product sampling data according to the preset limit parameters, the preset normal formula, and the respective metal average and metal standard deviation of each group of metal-containing product sampling data.

Specifically, the preset normal formula is:

σ is the standard deviation, μ is the average value, x is the threshold, the minimum threshold ≤ x ≤ the maximum threshold, the threshold interval value is used for incremental calculation, and F is the calculated probability.

Substituting different values of x into the above formula can calculate the sum of the probabilities less than or equal to the value of x. For products, values below x are normal, and values above x are false detections; for metals, values above x The value belongs to detection, and the value lower than x belongs to missing detection.

Assuming that there are 3 sets of metal-containing product sampling data (groups a, b, c) and a set of metal-free product sampling data (group d), the metal-free product sampling data, product standard deviation, product average and The preset limit parameters are substituted into the data set of the product probability sum calculated by the above formula one by one as Fd[x], and the product false detection rate group is 1-Fd[x]. For example: when x=500, Fd[x]=0.9 represents the normal passing probability of the product, and the corresponding product false detection rate is 1-0.9=0.1; when x=1000, Fd[x] =0.98, the corresponding product false detection rate is 1-0.98=0.02......

Substitute the sampling data, metal standard deviation, metal average value and preset limit parameters of group a metal-containing products into the sum of the probability calculated by the above formula one by one as Fa[x], that is, the sampling data of group a metal-containing products The missed detection rate group of the same, the same as the b and c group of metal-containing product sampling data corresponding to the missed detection rate group Fb[x] and Fc[x], according to 1-Fd[x], Fa[x], Fb The graphs drawn by [x] and Fc[x] are shown in Figure 4.

It should be noted that although this embodiment describes the sequence of S2121-S2124, it is not limited to this sequence, as long as the product misdetection rate group and the missed detection rate corresponding to each group of metal-containing product sampling data can be calculated Group only.

S203 selects the corresponding value as the threshold according to the product misdetection rate group, the corresponding misdetection rate group of each metal-containing product sampling data, and the actual misdetection leak detection rate requirements.

Preferably, S203 selects the corresponding value as the threshold according to the product misdetection rate group, the corresponding misdetection rate group of each metal-containing product sampling data, and the false detection miss detection rate requirements, including:

S213 draws the corresponding curve chart according to the corresponding missing detection rate group and product false detection rate group of each group of metal-containing product sampling data; S223 selects the corresponding value from the curve chart as the threshold according to the requirements of the false detection leakage detection rate .

The specific explanation is the same as in the above embodiment, and will not be repeated here.

In this embodiment, a specific probability calculation method is given for the sampling data that conforms to the normal distribution law, and quantitative data support is obtained, so that the selected value as the threshold value is more in line with actual monitoring needs.

In yet another embodiment of the present invention, as shown in FIG. 9, a digital performance analysis method includes:

S301 When a certain time interval is reached, acquire (preferably, the closest to the current time) a set of metal-free product sampling data and at least one set of metal-containing product sampling data.

S302 calculates the product false detection rate group corresponding to the metal-free product sampling data and each group of metal-containing product sampling data according to the distribution rules of the metal-free product sampling data and the metal-containing product sampling data obtained at a certain time interval The corresponding missed detection rate group.

S303 selects the corresponding value as the threshold according to the requirements of the product misdetection rate group, the corresponding misdetection rate group and the misdetection miss rate of each metal-containing product sampling data.

Specifically, the threshold value selection in this embodiment may be periodically performed when actually used, and when a certain time interval is set sufficiently small, real-time monitoring may be achieved.

When the corresponding program is run on the metal detector, the metal detector selects a set of metal-free product sampling data and at least one set of metal-containing product sampling data based on the sampling data obtained in real time, and performs real-time/periodical calculations. Choose reasonable data as the threshold.

When other equipment is used to monitor the threshold of the metal detector, it can monitor the metal detector in real time, obtain each sampling data in real time, select a set of product sampling data without metal and at least one set of product sampling data with metal for real-time /Periodic calculation, select reasonable data as threshold for metal detector.

Preferably, when the amount of each sampled data is updated, a certain amount of data closest to the current time is taken for analysis, so that the calculation result is more in line with the current situation.

S303 According to the requirements of the product misdetection rate group, each group of metal-containing product sampling data corresponding to the missed detection rate group and the false detection missed detection rate, selecting the corresponding value as the threshold includes:

S313 When the set threshold meets the requirements of false detection and miss detection rate, this threshold is still adopted;

S323 When the set threshold does not meet the requirements of false detection and miss detection rate, then re-select according to the requirements of the product false detection rate group, each group of metal-containing product sampling data corresponding to the corresponding false detection rate group and false detection leak detection rate The corresponding value is updated or the threshold is updated.

Specifically, in real-time/periodical monitoring, it means that the metal detector has been set with a threshold (may be a value set based on experience or a value selected based on the result calculated in real time), and put into actual product monitoring In the actual use process, due to interference, temperature, and the on-site electromagnetic environment may slowly affect the metal detector and cause the signal to increase, the currently set threshold may not meet the low miss detection rate of the metal and the low false detection rate of the product.

Therefore, you can confirm whether the set threshold needs to be adjusted according to the latest calculated product false detection rate group, the corresponding leakage detection rate group and the false detection leakage detection rate requirements of each group of metal-containing product sampling data, if it meets the requirements The requirements for false detection and missed detection rate do not need to be adjusted. If it does not meet the requirement, you need to re-select a data as a new threshold (the specific implementation process can automatically select a new threshold, or send a prompt to let the staff know Need to adjust) to meet production requirements.

Optionally, according to the distribution rule of the metal-free product sampling data and the metal-containing product sampling data, the product false detection rate group corresponding to the metal-free product sampling data is calculated and each group of metal-containing product sampling data corresponds to each The missed detection rate group includes:

When the sampling data of metal-free products and the sampling data of metal-containing products conform to the normal distribution law, according to the preset normal formula, the preset limit parameters, the sampling data of metal-free products and the sampling data of each metal-containing product, The product misdetection rate group corresponding to the metal-free product sampling data and the missing detection rate group corresponding to each group of metal-containing product sampling data are calculated.

Optionally, according to the preset normal formula, the preset limit parameters, the metal-free product sampling data and the metal-containing product sampling data, the product false detection rate group and each corresponding to the metal-free product sampling data are calculated Groups of metal-containing product sampling data corresponding to the missing detection rate group include:

Calculate the corresponding product average and product standard deviation based on the sampling data of products that do not contain metals;

According to the sampling data of each group of metal-containing products, calculate the corresponding metal average and metal standard deviation of each group of metal-containing product sampling data;

According to the product average value, product standard deviation, preset limit parameters (the preset limit parameters include: minimum threshold, maximum threshold and threshold interval value) and the preset normal formula, the product error corresponding to the sample data of metal-free products is calculated Inspection rate group;

According to the preset limit parameters, the preset normal formula, and the metal average and metal standard deviation corresponding to the sample data of each group of metal-containing products, respectively, the leak detection rate group corresponding to the sample data of each group of metal-containing products is calculated.

Optionally, according to the requirements of the product false detection rate group, the corresponding missing detection rate group and the false detection leakage detection rate of each group of metal-containing product sampling data, the selection of the corresponding value as the threshold includes:

Draw the corresponding graph according to the corresponding missing detection rate group and product false detection rate group of each group of metal-containing product sampling data;

According to the requirement of false detection miss detection rate, select the corresponding value from the graph as the threshold.

The same explanation part of this embodiment as the above method embodiments is not repeated here, please refer to the above method embodiments.

In this embodiment, periodic/real-time monitoring is performed on the selection of the threshold to ensure that the threshold set by the metal detector can meet the actual production needs.

FIG. 7 shows an embodiment of a digital performance analysis device of the present invention, including:

The obtaining module 10 is used to obtain a set of sampling data of products containing no metal and at least one set of sampling data of products containing metal.

The number of sampled data in each group is set according to the actual accuracy, and can be any number. For example, as shown in Fig. 3, there are 4 sets of sampling data, one of which is the product sampling data of metal-free products, and the other three are the sampling data of metal-containing products. Each group has 100 data.

Generally, when testing the sampling data of products containing metals, metal balls (for example: 1.5mm diameter iron beads, 2.0mm non-ferrous metal beads, 2.5mm stainless steel beads.) are put into the product and let them pass the metal detection器取。 Obtained. Due to the influence of various complex and inevitable objective factors, such as speed when passing, relative machine position, running track, conveyor belt, vibration, product temperature, ambient temperature, power supply voltage, atmospheric pressure, static electricity, complex electromagnetic in the workshop The environment and so on cause the signal of the metal ball to be almost the same every time it passes. Therefore, the number of sampled data should be as much as possible, for example: set 200, 150, etc. in each group to improve accuracy.

Sampling data is obtained through metal detector monitoring. When the program corresponding to the digital performance analysis method runs on the metal detector, it can be directly called; when this program runs on equipment other than the metal detector, it can be sent by the metal detector It can also be identified by the camera module during the monitoring process of the metal detector. The specific acquisition method of sampling data is not limited here.

When the program corresponding to the digital performance analysis method does not run on the metal detector, optionally, the digital performance analysis device further includes: a camera module for real-time acquisition of the working state of the metal detector; the acquisition module 10 is further used for The working status of the metal detector acquired in real time identifies a set of product sampling data without metal and at least one set of product sampling data with metal.

Specifically, the camera module is composed of several camera devices, and each camera device is used to monitor one or more metal detectors, and it photographs the working state of the metal detector (for example: test results when testing products and products containing metals) Recognition and interpretation by the acquisition module 10 (for example, the conversion of pictures to values according to the captured pictures), to obtain a set of product sampling data without metal and at least one set of product sampling data with metal.

The functions of the acquisition module 10, the calculation module 20, and the selection module 30 can be performed by one computer/server. The camera device is connected to the computer/server (for example: each camera transmits the captured video signal to the computer/server via a network device through a high-speed data communication line. The high-speed data communication line can be Ethernet, corporate LAN, or the Internet , Can also be various high-speed communication lines including USB) to ensure data transmission between the two. Of course, the camera device can also work independently, and it automatically reminds when it finds that the false detection rate and the missing detection rate of the metal detector need to be adjusted.

The calculation module 20 is used to calculate the product misdetection rate group corresponding to the metal-free product sampling data and each group of metal-containing product sampling data according to the distribution rule of the metal-free product sampling data and the metal-containing product sampling data The corresponding missed detection rate group.

The selection module 30 is used to select the corresponding value as the threshold according to the requirements of the product false detection rate group, the corresponding missing detection rate group of each group of metal-containing product sampling data, and the actual false detection leakage detection rate.

Most of the time, it is a value range that meets the requirements for the false detection rate and the missed detection rate. You can arbitrarily choose one of the values in this value range as the threshold.

Preferably, the selection module 30 is configured to select the corresponding value as the threshold according to the product false detection rate group, the corresponding missing detection rate group of each group of metal-containing product sampling data, and the actual false detection missing detection rate requirements:

The drawing sub-module 31 is used to draw a corresponding graph according to the corresponding missing detection rate group and product false detection rate group of each group of metal-containing product sampling data;

The selection sub-module 32 is used to select the corresponding value from the graph as the threshold according to the false detection and miss detection rate requirements.

Specifically, the calculated missed detection rate groups and product misdetection rate groups are drawn into corresponding graphs, so that users can intuitively understand the false detection rate and missed detection rate corresponding to each value, and understand the threshold value conveniently and quickly. Choose a reason. For specific examples, please refer to the corresponding method embodiments, which will not be repeated here.

In this embodiment, according to the distribution rules of sampling data of metal-free products and sampling data of metal-containing products, an appropriate threshold is selected according to the requirements of the false detection rate and the missing detection rate, so that the selection of the threshold is supported by effective quantitative data. The threshold is more in line with the actual use requirements, reducing the false detection rate and the missed detection rate, and improving the monitoring of product quality.

FIG. 8 shows another embodiment of a digital performance analysis device, including:

The obtaining module 10 is configured to obtain a set of product sampling data without metal and at least one set of product sampling data with metal.

Optionally, when the program corresponding to the digital performance analysis method does not run on the metal detector, optionally, the digital performance analysis device further includes: a camera module 40 for acquiring the working state of the metal detector in real time; the acquisition module 10 It is further used to identify a set of metal-free product sampling data and at least one set of metal-containing product sampling data based on the working status of the metal detector acquired in real time.

The specific explanation is the same as the above device embodiment, and will not be repeated here.

The calculation module 20 is used to calculate the product misdetection rate group corresponding to the metal-free product sampling data and each group of metal-containing product sampling data according to the distribution rule of the metal-free product sampling data and the metal-containing product sampling data The corresponding missed detection rate groups include:

The calculation module 20, when the sampling data of the product containing no metal and the sampling data of the product containing metal conform to the normal distribution law, according to the preset normal formula and the preset limit parameters (the preset limit parameters include: the minimum threshold, the maximum threshold and (Threshold interval value), metal-free product sampling data and each metal-containing product sampling data, the product misdetection rate group corresponding to the metal-free product sampling data and the corresponding leakage of each group of metal-containing product sampling data are calculated. Inspection rate group.

For example, if the metal to be detected is stainless steel and iron, as shown in FIG. 3, the maximum threshold can be set to 1500, the minimum threshold is 5, and the threshold interval value is 1. Of course, these data can be flexibly changed, for example: the maximum threshold is set to 1400, the minimum threshold is 10, the threshold interval value is 3 and so on.

Preferably, the calculation module 20 calculates the product false detection rate corresponding to the metal-free product sampling data according to the preset normal formula, the preset limit parameters, the metal-free product sampling data and the metal-containing product sampling data The missing detection rate groups corresponding to the sampling data of each group and each group of metal-containing products include:

The calculation module 20 calculates the corresponding product average and product standard deviation based on the sampling data of the metal-free products; and, based on the sampling data of each group of metal-containing products, calculates the corresponding metal average of each group of metal-containing product sampling data Value and metal standard deviation; and, based on the product average, product standard deviation, preset limit parameters and preset normal formula, the product misdetection rate group corresponding to the product sampling data is calculated; and, according to the preset limit parameters, Suppose the normal formula and the metal mean value and metal standard deviation corresponding to the sampling data of each group of metal-containing products, respectively, and calculate the missing detection rate group corresponding to each group of metal-containing product sampling data.

Specifically, the average value is the sum of the sampled data divided by the total number. For example: there are 150 data in a set of product sampling data, then add the 150 data and divide by 150 to get the average value of the product. Similarly, the average value of the metal corresponding to each group of metal sampling data.

The standard deviation is calculated according to the formula in the corresponding method embodiment described above. Please refer to the corresponding method embodiment, which will not be repeated here.

For the preset normal formula, please refer to the corresponding method embodiment. Substituting different values of x into the preset normal formula can calculate the sum of the probabilities less than or equal to the value of x. For products, below the value of x is normal, high The x value is a false detection; for metals, a value higher than x is detected, and a value lower than x is missed. For specific examples, please refer to the corresponding method embodiments, which will not be repeated here.

The selection module 30 is used to select the corresponding value as the threshold according to the product misdetection rate group, the corresponding misdetection rate group and the misdetection miss detection rate requirements of each group of metal-containing product sampling data.

Preferably, the selection module 30 is configured to select the corresponding value as the threshold based on the product misdetection rate group, the corresponding misdetection rate group and the misdetection miss rate requirement of each metal-containing product sampling data, including:

The specific implementation process is the same as the specific implementation process in the above device embodiment, and will not be described in detail here.

When the procedure corresponding to the threshold selection method is executed by a non-metal detector device, the metal detector can be monitored by cameras and other devices to obtain sampling data, and the implementation methods are diversified to meet the application requirements of different occasions.

In yet another device embodiment of the present invention, a digital performance analysis device includes:

The acquiring module 10 is configured to acquire (preferably, the closest to the current time) a set of metal-free product sampling data and at least one set of metal-containing product sampling data when a certain time interval is reached;

The calculation module 20 is used to calculate the product misdetection rate group corresponding to the metal-free product sampling data and each group of metal-containing product sampling data according to the distribution rule of the metal-free product sampling data and the metal-containing product sampling data The corresponding missing detection rate group;

The selection module 30 is used to select the corresponding value as the threshold according to the requirements of the product misdetection rate group, the corresponding misdetection rate group and the misdetection miss rate of each metal-containing product sampling data.

When other equipment is used to monitor the threshold of the metal detector, it can monitor the metal detector in real time, obtain each sampling data in real time, select a set of product sampling data without metal and at least one set of product sampling data with metal for real-time /Periodic calculation, giving reasonable suggestions for metal detectors as thresholds.

The selection module 30 is used to select the corresponding value as the threshold according to the requirements of the product false detection rate group, the corresponding leakage detection rate group and the false detection leakage detection rate of each group of metal-containing product sampling data, including:

The selection module 30 is used when the set threshold meets the requirements of false detection and miss detection rate, and the threshold is still used; and, when the set threshold does not meet the requirements of false detection and miss detection rate, according to the product misdetection rate group, Each group of metal-containing product sampling data corresponds to the requirements of the missed detection rate group and the false detection missed detection rate, and re-select the corresponding value to update or prompt to update the threshold.

Optionally, the calculation module 20 is configured to calculate the product false detection rate group corresponding to the metal-free product sampling data and the metal-containing product sampling data distribution product corresponding to the metal-free product sampling data and each group of metal-containing products The corresponding missed detection rate groups of product sampling data include:

Calculation module 20, when the sampling data of metal-free products and the sampling data of metal-containing products conform to the normal distribution law, according to the preset normal formula, the preset limit parameters, the sampling data of metal-free products and the metal-containing product For product sampling data, calculate the product misdetection rate group corresponding to the metal-free product sampling data and the missing detection rate group corresponding to each group of metal-containing product sampling data.

Optionally, the calculation module 20 calculates the product misdetection corresponding to the metal-free product sampling data according to the preset normal formula, the preset limit parameters, the metal-free product sampling data and the metal-containing product sampling data The rate group and each group of metal-containing product sampling data corresponding to the missing detection rate group include:

The calculation module 20 calculates the corresponding product average and product standard deviation based on the sampling data of the metal-free products; and, based on the sampling data of each group of metal-containing products, calculates the corresponding metal average of each group of metal-containing product sampling data Value and metal standard deviation; and, based on the product average, product standard deviation, preset limit parameters (the preset limit parameters include: minimum threshold, maximum threshold, and threshold interval value) and the preset normal formula, the calculated metal-free The product false detection rate group corresponding to the product sampling data of the product; and, according to the preset limit parameters, the preset normal formula and the metal average and metal standard deviation corresponding to each group of metal-containing product sampling data, each group is calculated separately Missing detection rate group corresponding to metal-containing product sampling data.

Optionally, the selection module 30 is configured to select the corresponding value as the threshold according to the requirements of the product misdetection rate group, the corresponding misdetection rate group and the misdetection miss rate of each metal-containing product sampling data, including:

The selection sub-module 32 is used to select the corresponding value from the graph as the threshold according to the requirement of false detection and miss detection rate.

Optionally, when the digital performance analysis device is a non-metal detector device, the digital performance analysis device may further include: a camera module 40 for acquiring the working state of the metal detector in real time; The obtained working state of the metal detector identifies a set of sampling data of products without metal and at least one set of sampling data of products with metal.

The same explanation part of this embodiment as the above device embodiments is not repeated here, please refer to the above method embodiments.

It should be noted that the above embodiments can be freely combined as needed. The above is only the preferred embodiment of the present invention. It should be pointed out that for those of ordinary skill in the art, without departing from the principles of the present invention, several improvements and retouches can be made. These improvements and retouches also It should be regarded as the protection scope of the present invention.

Claims

A digital performance analysis method, characterized in that it includes the following steps:

S1 Obtain a set of sample data of metal-free products and at least one set of sample data of metal-containing products;

S2 calculates the product misdetection rate group corresponding to the metal-free product sampling data and each group of metal-containing product sampling according to the distribution rule of the metal-free product sampling data and the metal-containing product sampling data The missing detection rate group corresponding to the data;

S3 selects the corresponding value as the threshold according to the requirements of the product misdetection rate group, the corresponding misdetection rate group and the misdetection miss rate of each metal-containing product sampling data.
The digital performance analysis method according to claim 1, characterized in that, based on the distribution rule of the sampling data of the metal-free product and the sampling data of the metal-containing product, the metal-free product is calculated The product false detection rate group corresponding to the product sampling data and the corresponding missing detection rate group for each group of metal-containing product sampling data include:

When the sampling data of the metal-free product and the sampling data of the metal-containing product conform to the normal distribution law, according to a preset normal formula, a preset limit parameter, the sample data of the metal-free product and each Based on the sampling data of metal-containing products, the product misdetection rate group corresponding to the sampling data of metal-free products and the missing detection rate group corresponding to each group of metal-containing product sampling data are calculated.
The digital performance analysis method according to claim 2, characterized in that, according to a preset normal formula, a preset limit parameter, the sampling data of the metal-free product and the sampling data of each of the metal-containing products Calculate that the product false detection rate group corresponding to the metal-free product sampling data and the corresponding missing detection rate group of each group of metal-containing product sampling data include:

Calculate the corresponding product average and product standard deviation based on the sampling data of the metal-free products;

According to the sampling data of each group of metal-containing products, calculate the respective metal average and metal standard deviation corresponding to the sampling data of each group of metal-containing products;

Calculate the product misdetection rate group corresponding to the metal-free product sampling data according to the product average value, the product standard deviation, the preset limit parameter and the preset normal formula;

According to the preset limit parameter, the preset normal formula and the metal average and metal standard deviation corresponding to each group of the metal-containing product sampling data, respectively calculate the leak detection corresponding to each group of metal-containing product sampling data Rate group.
The digital performance analysis method according to claim 2, wherein the preset limit parameters include: a minimum threshold, a maximum threshold, and a threshold interval value.
The digital performance analysis method according to claim 1, characterized in that, according to the product misdetection rate group, each group of metal-containing product sampling data corresponds to the respective misdetection rate group and misdetection miss rate As required, selecting the corresponding value as the threshold includes:

Draw the corresponding graph according to the corresponding missing detection rate group and the product false detection rate group of each group of metal-containing product sampling data;

According to the requirement of false detection miss detection rate, the corresponding value is selected from the graph as the threshold.
The digital performance analysis method according to any one of claims 1 to 5, wherein:

The step S1 of acquiring a set of metal-free product sampling data and at least one set of metal-containing product sampling data includes:

When a certain time interval is reached, obtain a set of sample data of metal-free products and at least one set of sample data of metal-containing products;

Step S2 calculates the product misdetection rate group corresponding to the metal-free product sampling data and each group of metal-containing products according to the distribution rule of the metal-free product sampling data and the metal-containing product sampling data The corresponding missed detection rate groups of product sampling data include:

According to the distribution rules of the metal-free product sampling data and the metal-containing product sampling data acquired at a certain time interval, the product misdetection rate group corresponding to the metal-free product sampling data and each group containing Metal product sampling data corresponding to the missing detection rate group.
The digital performance analysis method according to claim 6, wherein the step S3 is based on the product misdetection rate group, the respective misdetection rate group and misdetection miss detection rate of each group of metal-containing product sampling data Requirements, selecting the corresponding value as the threshold further includes:

When the set threshold meets the requirements of the false detection and miss detection rate, this threshold is still adopted;

When the set threshold does not meet the requirements of the false detection and detection rate, according to the product false detection rate group, each group of metal-containing product sampling data corresponding to the corresponding detection rate group and false detection and detection rate requirements , Reselect the corresponding value to update or prompt to update the threshold.
A digital performance analysis device, characterized in that it includes:

An acquisition module for acquiring a set of sample data of products containing no metal and at least one set of sample data of products containing metal;

The calculation module is used to calculate the product misdetection rate group corresponding to the metal-free product sampling data and each group containing Metal product sampling data corresponding to the missing detection rate group;

The selection module is used to select the corresponding value as the threshold value according to the requirements of the product false detection rate group, the corresponding leakage detection rate group and the false detection leakage detection rate of each group of metal-containing product sampling data.
The digital performance analysis device according to claim 8, wherein the calculation module is configured to calculate the distribution according to the distribution rule of the sampling data of the metal-free product and the sampling data of the metal-containing product The product false detection rate group corresponding to the metal-free product sampling data and the corresponding missing detection rate group for each group of metal-containing product sampling data include:

The calculation module, when the sampling data of the metal-free product and the sampling data of the metal-containing product conform to the normal distribution law, according to a preset normal formula, a preset limit parameter, the metal-free product Sampling data and each of the metal-containing product sampling data are calculated to obtain a product misdetection rate group corresponding to the metal-free product sampling data and a missing detection rate group corresponding to each group of metal-containing product sampling data.
The digital performance analysis device according to claim 9, wherein the calculation module is based on a preset normal formula, a preset limit parameter, sampling data of the metal-free product and each of the metal-containing products Sampling data, and calculating the product misdetection rate group corresponding to the metal-free product sampling data and the missing detection rate group corresponding to each group of metal-containing product sampling data include:

The calculation module calculates the corresponding product average and product standard deviation based on the sampling data of the metal-free product; and calculates each group of the metal-containing product based on each group of the metal-containing product sampling data The metal average value and metal standard deviation corresponding to the sampled data; and, based on the product average value, the product standard deviation, the preset limit parameter, and the preset normal formula, calculate the metal-free product The product false detection rate group corresponding to the sampled data; and, based on the preset limit parameter, the preset normal formula, and the metal mean and metal standard deviation corresponding to each group of the metal-containing product sample data, respectively, calculated The missing detection rate group corresponding to the sampling data of each metal-containing product.
The digital performance analysis device according to claim 8, characterized in that the selection module is configured to, based on the product false detection rate group, each group of metal-containing product sampling data, corresponding to the respective missing detection rate group and false detection For the requirement of missed detection rate, the selection of the corresponding value as the threshold includes:

A drawing sub-module, used to draw a corresponding graph according to the missing detection rate group and the product false detection rate group corresponding to each group of metal-containing product sampling data;

A selection sub-module is used to select a corresponding value from the graph as a threshold according to the requirement of false detection and miss detection rate.
The digital performance analysis device according to claim 8, further comprising:

The camera module is used to obtain the working status of the metal detector in real time;

The acquisition module is further used to identify a set of metal-free product sampling data and at least one set of metal-containing product sampling data based on the working status of the metal detector acquired in real time.
The digital performance analysis device according to any one of claims 8 to 12, wherein the acquisition module is further configured to acquire a set of metal-free product sampling data and at least one set of Metal product sampling data.
The digital performance analysis device according to claim 13, wherein the selection module is further configured to use the threshold value when the set threshold value meets the requirements of the false detection and missed detection rate; and, when it has been set The threshold value does not meet the requirements of the false detection and leakage detection rate, and then re-select according to the requirements of the product false detection rate group, the corresponding leakage detection rate group and the false detection leakage detection rate of each group of metal-containing product sampling data The corresponding value is updated or prompted to update the threshold.