Detection equipment interval parameter adjusting method and system and medium

Abstract

The invention discloses a detection equipment interval parameter adjustment method and system and a medium, and the method comprises the steps: obtaining historical accumulated data, and forming a defect library list through the calibration of the historical accumulated data; constructing an interval tree according to the optimal screening ratio sequence by adopting a preset algorithm, and outputting a result; obtaining the result by adopting a data calling algorithm; and carrying out online verification on the result. According to the method, historical accumulated big data is utilized, and the over-detection rate is considered on the basis of guaranteeing the level of the over-detection rate under the limitation of a parameter interval; moreover, the method does not need to depend on deep experience, and can achieve the quick operation and iteration.

Description

Technical Field

[0001] The present invention relates to the technical field of parameter adjustment of automatic optical detection equipment, and particularly relates to a method, a system and a medium for adjusting interval parameters of detection equipment. Background Art

[0002] The parameter adjustment of equipment such as detection instruments (mainly automatic optical detection equipment, for example: the automatic optical detection equipment (AOI) configured on a wafer inspection machine or the automatic optical detection equipment (AOI) configured in the front process of a polarizer) is often an interval, for example: [a, +∞] (or (a, +∞) (which can also be written as ≥a or >a), a left-closed and right-open interval; [-∞, b] (or [-∞, b]) (which can also be written as ≤b or <b), a left-open and right-closed interval; [a, b] (or (a, b], [a, b)), a left-and-right closed interval; The interval-type parameters have long relied on the rich experience of debuggers. Insufficient experience often leads to poor debugging effects; and such parameters have the problem of "easy to tighten, difficult to relax". The parameter adjustment is generally to simulate the parameters of the previously missed-inspected products through a simulator; and then add the tightened parameters to mass production. However, when passing the inspection, it is impossible to grasp the degree of relaxation, and it is very easy to have missed-inspected situations for the adjusted parameters; and because the parameter adjustment requires deep experience and is basically "tightened" each time, the actual iteration speed is very slow, and it often takes a month for one iteration.

[0003] The parameter adjustment solutions in the prior art mainly have the following defects:

[0004] 1. Only ensuring missed inspection and difficult to take into account passing inspection, that is, misjudging good products as bad products:

[0005] The existing mode uses the simulation function to perform "tightening" adjustment on the missed-inspected products; but there is currently no effective method to adjust the parameters of the passed-inspected products. Generally, the parameter adjustment is one-way, only "tightening" and no "relaxing": the missed-inspection rate can be guaranteed, but the passing-inspection rate often deteriorates.

[0006] 2. Slow iteration speed:

[0007] Since the "tightening" adjustment often leads to out-of-control passing-inspection rate. Unless there are major customer complaints, the parameters are generally not adjusted. The iteration often takes months or quarters as the cycle; it is difficult to quickly converge to the ideal state of low missed inspection and low passing inspection.

[0008] 3. Dependence on personnel experience:

[0009] The existing mode cannot effectively utilize daily data, and the parameter adjustment depends on personnel experience. It takes a long time to train a qualified engineer and has extremely high requirements for the quality of parameter adjusters. Summary of the Invention

[0010] The main purpose of the present invention is to propose a method, system and medium for adjusting the interval parameters of detection equipment, aiming to utilize the big data accumulated historically to ensure the level of missed detection rate while taking into account the pass rate under the restriction of parameter interval; and it does not need to rely on deep experience, and can be quickly started and iterated.

[0011] To achieve the above object, the present invention proposes a method for adjusting interval parameters of a detection device, the method comprising the following steps:

[0012] Step S10, acquiring historical accumulated data, and forming a defect library list by calibrating the historical accumulated data;

[0013] Step S20, constructing an interval tree in the order of the best screening ratio using a preset algorithm, and outputting the result;

[0014] Step S30, using a data retrieval algorithm to obtain the result;

[0015] Step S40, verifying the result online.

[0016] A further technical solution of the present invention is that step S10 comprises:

[0017] Step S101, data collection: extract the detection data that needs to be optimized and place them in a unified folder;

[0018] Step S102, calibrate the defect image and place it in a corresponding folder;

[0019] Step S103: Corresponding the image category calibration and the defect items to form the defect library list.

[0020] A further technical solution of the present invention is that step S102 comprises:

[0021] Traverse the defective images and classify them into corresponding A-product folders or B-product folders.

[0022] A further technical solution of the present invention is that step S103 comprises:

[0023] The raw data is traversed through the program and a preliminary defect list is formed in a unified format.

[0024] A further technical solution of the present invention is that the step of traversing the original data through a program and forming a preliminary defect list in a unified format includes:

[0025] The data in the folder of product A is traversed by the program through the image names, and matched with the defect list, and a new "category" column is added, marked A; the data in the folder of product B is traversed by the program through the image names, and matched with the defect list, and a new "category" column is added, marked B;

[0026] Save and complete the defect library checklist.

[0027] A further technical solution of the present invention is that step S20 comprises:

[0028] Define the left and right ends of the interval of the item with problems or defects, as well as the B-product screening rate, A-product screening rate, screening efficiency ratio, and the step speed of interval reduction. Among them, the B-product screening rate refers to the number of items classified as Class B in a certain item interval / the total number of Class B in the current defect list, the A-product screening rate refers to the number of items classified as Class A in a certain item interval / the total number of Class A in the current defect list, and the screening efficiency ratio refers to the ratio of the B-product screening rate to the A-product screening rate;

[0029] Determine whether the column items meet the indicators;

[0030] If not satisfied, output the result;

[0031] If it is met, the interval will be shortened item by item based on the indicators being met;

[0032] Select the best items and intervals;

[0033] Filter the items according to the best item and interval, and remove the item column from the data;

[0034] Return to the step of determining whether the column item satisfies the indicator.

[0035] A further technical solution of the present invention is that step S30 comprises:

[0036] A new empty table M1 is created according to the defect library list, and table M1 is updated with new values ​​according to the optimal project interval algorithm. The updated items include: project name, left end value of interval, right end value of interval, B product screening rate value, A product screening rate value, and screening efficiency ratio.

[0037] A further technical solution of the present invention is that step S40 comprises:

[0038] If table M1 is empty, it means that no better indicators that meet the requirements can be obtained based on the existing defect library list; in this case, the defect library list needs to be further expanded to increase the balance and diversity of A / B categories;

[0039] If the M1 table is not empty, update the parameters according to the corresponding items and verify the effect online.

[0040] To achieve the above objectives, the present invention also proposes a detection device interval parameter adjustment system, the system comprising a memory, a processor and a detection device interval parameter adjustment program stored on the processor, the detection device interval parameter adjustment program executes the steps of the method described above when run by the processor.

[0041] To achieve the above object, the present invention further proposes a computer-readable storage medium, wherein the computer-readable storage medium stores a detection device interval parameter adjustment program, and the detection device interval parameter adjustment program executes the steps of the above method when executed by a processor.

[0042] The beneficial effects of the detection equipment interval parameter adjustment method, system and medium of the present invention are:

[0043] The present invention utilizes the big data accumulated in history through the above technical scheme, and takes into account the pass rate while ensuring the missed detection rate level under the restriction of the parameter range. It does not need to rely on profound experience, and can be quickly learned and iterated. It accelerates the monthly or quarterly iteration to a daily cycle, and even completes one iteration in one day, which is far faster than the traditional iteration speed. It simplifies the parameter adjustment process that is "easy to be strict and difficult to be loose" and heavily relies on experience to a process that can be implemented by simple calibration. The adjusted interval parameters take into account both missed detection and pass rate, achieving a win-win situation. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying creative work.

[0045] Figure 1 It is a flow chart of a preferred embodiment of the method for adjusting interval parameters of a detection device of the present invention;

[0046] Figure 2 It is a detailed flow chart of step S20 in the present invention;

[0047] Figure 3 is a schematic flow chart of step S10;

[0048] Figure 4 It is a schematic diagram showing that interval operations satisfy the commutative law;

[0049] Figure 5 This is a diagram showing that interval operations satisfy the associative law.

[0050] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0051] 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.

[0052] The present invention proposes a method for adjusting the interval parameters of a detection device. The present invention utilizes the big data accumulated historically, and takes into account the pass rate while ensuring the missed detection rate level under the restriction of the parameter interval. The method does not require reliance on deep experience, and can be quickly learned and iterated, and even one iteration can be completed in one day, which is far faster than the traditional iteration speed.

[0053] The application scenarios of this aspect are: equipment or instruments in mass production situations (when parameters can be called repeatedly) and parameters are adjusted into intervals (left open and right closed, left closed and right open can be generalized into a particularly large generalized interval).

[0054] For details, please refer to Figure 1 The preferred embodiment of the detection device interval parameter adjustment method of the present invention comprises the following steps:

[0055] Step S10, acquiring historical accumulated data, and forming a defect library list by calibrating the historical accumulated data;

[0056] Step S20, constructing an interval tree in the order of the best screening ratio using a preset algorithm, and outputting the result;

[0057] Step S30, using a data retrieval algorithm to obtain the result;

[0058] Step S40, verifying the result online.

[0059] Specifically, the step S10 includes:

[0060] Step S101, data collection: extract the detection data that needs to be optimized and place them in a unified folder;

[0061] Step S102, calibrate the defect image and place it in a corresponding folder;

[0062] Step S103: Corresponding the image category calibration and the defect items to form the defect library list.

[0063] Wherein, the step S102 specifically includes:

[0064] Traverse the defective images and classify them into corresponding A-product folders or B-product folders.

[0065] The step S103 specifically includes:

[0066] The raw data is traversed through the program and a preliminary defect list is formed in a unified format.

[0067] In this embodiment, the step of traversing the original data through a program and forming a preliminary defect list in a unified format includes:

[0068] The data in the folder of product A is traversed by the program through the image names, and matched with the defect list, and a new "category" column is added, marked A; the data in the folder of product B is traversed by the program through the image names, and matched with the defect list, and a new "category" column is added, marked B;

[0069] Save and complete the defect library checklist.

[0070] Please refer to Figure 2 In this embodiment, step S20 includes:

[0071] Define the left and right ends of the interval of the item with problems or defects, as well as the B-product screening rate, A-product screening rate, screening efficiency ratio, and the step speed of interval reduction. Among them, the B-product screening rate refers to the number of items classified as Class B in a certain item interval / the total number of Class B in the current defect list, the A-product screening rate refers to the number of items classified as Class A in a certain item interval / the total number of Class A in the current defect list, and the screening efficiency ratio refers to the ratio of the B-product screening rate to the A-product screening rate;

[0072] Determine whether the column items meet the indicators;

[0073] If not satisfied, output the result;

[0074] If it is met, the interval will be shortened item by item based on the indicators being met;

[0075] Select the best items and intervals;

[0076] Filter the items according to the best item and interval, and remove the item column from the data;

[0077] Return to the step of determining whether the column item satisfies the indicator.

[0078] In this embodiment, step S30 includes:

[0079] A new empty table M1 is created according to the defect library list, and table M1 is updated with new values ​​according to the optimal project interval algorithm. The updated items include: project name, left end value of interval, right end value of interval, B product screening rate value, A product screening rate value, and screening efficiency ratio.

[0080] In this embodiment, the step S40 includes:

[0081] If table M1 is empty, it means that no better indicators that meet the requirements can be obtained based on the existing defect library list; in this case, the defect library list needs to be further expanded to increase the balance and diversity of A / B categories;

[0082] If the M1 table is not empty, update the parameters according to the corresponding items and verify the effect online.

[0083] The following combination Figures 1 to 5 The detection equipment interval parameter adjustment method of the present invention is further described in detail.

[0084] 1. Make full use of historical accumulated data and form a defect library list by calibrating historical data:

[0085] Inspection equipment such as chip sorting machines will output a folder per chip, which contains a defect detail text and a corresponding defect thumbnail. This format of data is scattered and cannot form "big data"; it needs to be sorted. The specific process is as follows: Figure 3 shown.

[0086] 1. Collect data:

[0087] Extract the test data that needs to be optimized and place them in a unified folder;

[0088] 2. Calibrate the defect image and place it in the corresponding folder;

[0089] a. Traverse defect images and transfer them to a unified folder;

[0090] b. Personnel visually classify defect images into corresponding A-product or B-product folders based on experience

[0091] 3. Align the image category with the defect items to form a defect library list;

[0092] a. Traverse the original data through the program and form a preliminary defect list in a unified format;

[0093] b. The data of product library A is traversed through the image name (unique code) through the program, and matched with the defect list, and a new "category" column is added, marked with A (other distinguishable codes are also acceptable);

[0094] c. The data of product library B is traversed through the image name (unique code) through the program; it is matched with the defect list, and a new "category" column is added, marked with B (other distinguishable codes are also acceptable);

[0095] d. Save and complete the defect library list

[0096] 2. Design and run the algorithm to construct the interval tree in the order of the best screening ratio:

[0097] 1. Define indicators:

[0098] Item: A problem or defect often has multi - dimensional data, and each dimension is an item.

[0099] For example: Defects usually have dimensional data such as size, length, width, average gray level, contrast, etc. Then size, length, width, average gray level, and contrast are all items.

[0100] Left end of the interval: Parameter adjustment is often an interval. For example: [a, +∞] (or (a, +∞) (which can also be written as ≥a or >a), a left - closed and right - open interval, and a becomes the left end; [-∞, b] (or [-∞, b]) (which can also be written as ≤b or <b; in actual operation, there is no infinity (∞), so at this time, select the minimum value of this item in the defect library list as the left end of the interval), a left - open and right - closed interval; [a, b] (or (a, b], [a, b)), a left - and - right - closed interval, then a is recorded as the left end of the interval.

[0101] Right end of the interval: Parameter adjustment is often an interval. For example: [a, +∞] (or (a, +∞) (which can also be written as ≥a or >a), a left - closed and right - open interval, in actual operation, there is no infinity (∞), so at this time, select the maximum value of this item in the defect library list as the left end; [-∞, b] (or [-∞, b]) (which can also be written as ≤b or <b); b is recorded as the right end, and at this time, select the minimum value of this item in the defect library list as the left end, a left - open and right - closed interval; [a, b] (or (a, b], [a, b)), a left - and - right - closed interval, then b is recorded as the right end.

[0102] Screening rate (Grade B): The number of items classified as Grade B within a certain item interval / The total number of Grade B in the current defect list. In the detection equipment, the most important purpose of parameter adjustment is to improve the screening rate (Grade B) level;

[0103] Screening rate (Grade A): The number of items classified as Grade A within a certain item interval / The total number of Grade A in the current defect list. In the detection equipment, the higher the screening rate (Grade A), the heavier the passing inspection. On the basis of ensuring the screening rate (Grade B) level, it is necessary to尽可能 reduce the screening rate (Grade A) level,

[0104] Screening efficiency ratio: Screening rate (Grade B) / Screening rate (Grade A) should be as high as possible. The higher the screening efficiency ratio, the stronger the screening ability; in actual operation, on the basis of ensuring the screening rate (Grade B) level, improve the screening efficiency ratio as much as possible.

[0105] Stepping speed: The speed at which the interval shrinks left and right, divided into fast and slow speeds. For example: The data distribution interval of item 1 is [a, b]; the minimum interval of data distribution is k1, then the fast stepping speed can be set to k1 / 2; the slow stepping speed can be set to 250*k1.

[0106] 2. Algorithm principle:

[0107] a. Interval operations satisfy the commutative law, such as Figure 4 shown.

[0108] b. Interval operations satisfy the associative law, such as Figure 5 shown.

[0109] Since interval operations satisfy the commutative law and the associative law, if there is an interval combination that meets the screening rate (B product) requirements, it can be combined in order of screening efficiency ratio.

[0110] 3. Subroutine logic or algorithm:

[0111] a. Quasi-normalization:

[0112] Different from traditional normalization which uses the entire column data range for calculation, quasi-normalization uses the data interval classified as product B as the range of variation. For example, the overall data range fluctuates between -5 and 10, and the data interval classified as product B is [2, 7]. Suppose the specific value of a row of this item is X0, then after mapping, X1=(X0-2) / (7-2).

[0113] The advantage of this algorithm is that the interval calculation has a clear starting point:

[0114] The left end of the converted interval is 0.

[0115] The right end of the converted interval is 1.

[0116] b. Step span:

[0117] Due to the inconsistency of project dimensions and distribution ranges, even after quasi-normalization, the distribution sparsity of each project is still different. If a unified step span is used, omissions are likely to occur. In order not to miss the optimal solution, the step span needs to be processed:

[0118] a) De-duplication of normalized data;

[0119] b) Sort the deduplicated data from small to large;

[0120] c) Perform list difference operation (subtract the previous digit from the last digit in the list) and generate a new list

[0121] d) Get the minimum value of the new list;

[0122] e) In addition: fast step span = 250*minimum value;

[0123] Slow step span = 1*minimum value;

[0124] 250 can be set to any value greater than 1; if the value is set too large (small), it may reduce the system operation efficiency, so it is recommended to use the default value; 1 can be set to any value not greater than 1, otherwise it may not be the optimal solution.

[0125] c. Interval division algorithm:

[0126] The main purpose of the interval partitioning algorithm is to perform an interval approximation step for a certain project.

[0127] If Flag == 0:

[0128] Use fast interval span

[0129] Left indent:

[0130] Left indent 1 step

[0131] Calculate the new interval screening rate (B product)

[0132] Calculate the new interval screening rate (Product A)

[0133] Calculate the new interval screening efficiency ratio

[0134] Right indent:

[0135] Indent the right end by 1 step (leave the left end unchanged)

[0136] Calculate the new interval screening rate (B product)

[0137] Calculate the new interval screening rate (Product A)

[0138] Calculate the new interval screening efficiency ratio

[0139] Left and right indents:

[0140] Indent both left and right

[0141] Calculate the new interval screening rate (B product)

[0142] Calculate the new interval screening rate (Product A)

[0143] Calculate the new interval screening efficiency ratio

[0144] If the left and right indented screening rate (B product) & screening efficiency ratio meet the requirements:

[0145] Update the left and right ends of the corresponding interval

[0146] Corresponding output update screening rate (B product)

[0147] Corresponding output update screening rate (A product)

[0148] Corresponding output update screening efficiency ratio

[0149] elif The left indented screening rate (B product) & screening efficiency ratio meet the requirements:

[0150] Update the corresponding left end of the interval

[0151] Corresponding output update screening rate (B product)

[0152] Corresponding output update screening rate (A product)

[0153] Corresponding output update screening efficiency ratio

[0154] elif The screening rate (B product) & screening efficiency ratio of the right indentation meets the requirements:

[0155] Update the right end of the corresponding interval

[0156] Corresponding output update screening rate (B product)

[0157] Corresponding output update screening rate (A product)

[0158] Corresponding output update screening efficiency ratio

[0159] else:

[0160] Flag==1

[0161] Continue

[0162] else:

[0163] Use slow interval span

[0164] Left indent:

[0165] Left indent 1 step

[0166] Calculate the new interval screening rate (B product)

[0167] Calculate the new interval screening rate (Product A)

[0168] Calculate the new interval screening efficiency ratio

[0169] Right indent:

[0170] Indent the right end by 1 step (leave the left end unchanged)

[0171] Calculate the new interval screening rate (B product)

[0172] Calculate the new interval screening rate (Product A)

[0173] Calculate the new interval screening efficiency ratio

[0174] Left and right indents:

[0175] Indent both left and right

[0176] Calculate the new interval screening rate (B product)

[0177] Calculate the new interval screening rate (Product A)

[0178] Calculate the new interval screening efficiency ratio

[0179] If the left and right indented screening rate (B product) & screening efficiency ratio meet the requirements:

[0180] Then update the corresponding interval left and right ends corresponding output update screening rate (B product)

[0181] Corresponding output update screening rate (A product)

[0182] Corresponding output update screening efficiency ratio

[0183] elif The left indented screening rate (B product) & screening efficiency ratio meet the requirements:

[0184] Update the corresponding left end of the interval

[0185] Corresponding output update screening rate (B product)

[0186] Corresponding output update screening rate (A product)

[0187] Corresponding output update screening efficiency ratio

[0188] elif The screening rate (B product) & screening efficiency ratio of the right indentation meets the requirements:

[0189] Update the right end of the corresponding interval

[0190] Corresponding output update screening rate (B product)

[0191] Corresponding output update screening rate (A product)

[0192] Corresponding output update screening efficiency ratio

[0193] else:

[0194] Flag1==0

[0195] Output list [item, interval left end, interval right end, screening rate (B product), screening rate (A product), screening efficiency ratio]

[0196] Break

[0197] d. Best Item Interval Algorithm:

[0198] The purpose of the optimal project interval algorithm is: a. Calculate the optimal interval of a certain project; b. Find the project and its interval that best meets the requirements:

[0199] Requirements:

[0200] p: Screening rate (B product) needs to be ≥ p

[0201] q: Screening efficiency ratio needs to be ≥q

[0202] Initialize the indicator:

[0203] Flag1=1

[0204] Item list (items in the existing defect library list)

[0205] Defect Entries (Number of entries in the existing defect library list)

[0206] Create a new empty table M0, column name:

[0207] Item name, interval left end value, interval right end value, screening rate (B product) value, screening rate (A product) value, screening efficiency ratio value

[0208] For project in project list:

[0209] Normalize this item column

[0210] Output the maximum value of category B in this item column (denoted as n_max)

[0211] Output the minimum value of category B in this item column (denoted as n_min)

[0212] initialization:

[0213] Flag=0,Flag1=0,

[0214] The left end of the interval = 0,

[0215] The right end of the interval = 1,

[0216] Screening rate (B product) = 0,

[0217] Screening rate (A product) = 2000,

[0218] Screening efficiency ratio = 0

[0219] while Flag1>0:

[0220] Interval partitioning algorithm

[0221] if (screening rate (B product) ≥ p) & (screening efficiency ratio ≥ q):

[0222] New items are added to the M0 project, including [project, interval left end value, interval right end value, screening rate (B product), screening rate (A product), screening efficiency ratio] of the interval division output

[0223] Correction corresponding to the left end value of the interval = left end value of the interval * (n_max-n_min) + n_min

[0224] Correction corresponding to the right end value of the interval = right end value of the interval * (n_max_n_min) + n_min

[0225] else:

[0226] break

[0227] lf len(M0)>0:

[0228] Select the item with the largest screening efficiency ratio;

[0229] If the number of entries > 1:

[0230] Select the maximum screening rate (B product)

[0231] If the number of entries > 1:

[0232] The first one

[0233] Output the best index [item, interval left end value, interval right end value, screening rate (B product), screening rate (A product), screening efficiency ratio]

[0234] Update the defect database list: filter the data by the interval of the best project and delete the column where the project is located

[0235] else:

[0236] Flag2=0

[0237] 3. Use data retrieval algorithm to obtain results:

[0238] Existing data: defect library list

[0239] Initialization indicator: Flag2 = 1

[0240] Create a new empty table M1, column names: Project name, Interval left end value, Interval right end value, Screening rate (B product) value, Screening rate (A product) value, Screening efficiency ratio value

[0241] While (len(item list)>0)&(len(defective item)>0)&(Flag2>0):

[0242] Best Item Interval Algorithm

[0243] Update M1 and update the newly added values ​​according to the optimal project interval algorithm: [project name, interval left end value, interval right end value, screening rate (B product) value, screening rate (A product) value, screening efficiency ratio value.

[0244] 4. Utilization results:

[0245] If the M1 table is empty, it means that no better indicators that meet the requirements can be obtained based on the existing defect library list. In this case, the defect library list needs to be further expanded to increase the balance and diversity of A / B categories.

[0246] If the M1 table is not empty, update the parameters according to the corresponding items and verify the effect online.

[0247] The beneficial effects of the detection equipment interval parameter adjustment method of the present invention are:

[0248] The present invention utilizes the big data accumulated in history through the above technical scheme, and takes into account the pass rate while ensuring the missed detection rate level under the restriction of the parameter range. It does not need to rely on profound experience, and can be quickly learned and iterated. It accelerates the monthly or quarterly iteration to a daily cycle, and even completes one iteration in one day, which is far faster than the traditional iteration speed. It simplifies the parameter adjustment process that is "easy to be strict and difficult to be loose" and heavily relies on experience to a process that can be implemented by simple calibration. The adjusted interval parameters take into account both missed detection and pass rate, achieving a win-win situation.

[0249] To achieve the above-mentioned purpose, the present invention also proposes a detection device interval parameter adjustment system, the system comprising a memory, a processor and a detection device interval parameter adjustment program stored on the processor, the detection device interval parameter adjustment program is executed by the processor to perform the steps of the method described in the above embodiment, which will not be repeated here.

[0250] To achieve the above objectives, the present invention also proposes a computer-readable storage medium, which stores a detection device interval parameter adjustment program. When the detection device interval parameter adjustment program is executed by a processor, the steps of the method described in the above embodiment are executed, which will not be repeated here.

[0251] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. All equivalent structural changes made by using the contents of the present invention specification and drawings under the concept of the present invention, or directly / indirectly applied in other related technical fields are included in the patent protection scope of the present invention.

Claims

1. A method for adjusting interval parameters of a detection device, characterized in that: The method comprises the following steps: Step S10, acquiring historical accumulated data, and forming a defect library list by calibrating the historical accumulated data; Step S20, constructing an interval tree in the order of the best screening ratio using a preset algorithm, and outputting the result; Step S30, using a data retrieval algorithm to obtain the result; Step S40, verifying the result online.

2. The detection equipment interval parameter adjustment method according to claim 1, characterized in that: The step S10 comprises: Step S101, data collection: extract the detection data that needs to be optimized and place them in a unified folder; Step S102, calibrate the defect image and place it in a corresponding folder; Step S103: Corresponding the image category calibration and the defect items to form the defect library list.

3. The detection equipment interval parameter adjustment method according to claim 2, characterized in that: The step S102 includes: Traverse the defective images and classify them into corresponding A-product folders or B-product folders.

4. The detection equipment interval parameter adjustment method according to claim 3, characterized in that: The step S103 includes: The raw data is traversed through the program and a preliminary defect list is formed in a unified format.

5. The detection equipment interval parameter adjustment method according to claim 4, characterized in that: The step of traversing the original data through a program and forming a preliminary defect list in a unified format includes: The data in the folder of product A is traversed by the program through the image names, and matched with the defect list, and a new "category" column is added, marked as A; the data in the folder of product B is traversed by the program through the image names, and matched with the defect list, and a new "category" column is added, marked as B; Save and complete the defect library checklist.

6. The detection equipment interval parameter adjustment method according to claim 5, characterized in that: The step S20 comprises: Define the left and right ends of the interval of the item with problems or defects, as well as the B-product screening rate, A-product screening rate, screening efficiency ratio, and the step speed of interval reduction. Among them, the B-product screening rate refers to the number of items classified as Class B in a certain item interval / the total number of Class B in the current defect list, the A-product screening rate refers to the number of items classified as Class A in a certain item interval / the total number of Class A in the current defect list, and the screening efficiency ratio refers to the ratio of the B-product screening rate to the A-product screening rate; Determine whether the column items meet the indicators; If not satisfied, output the result; If it is met, the interval will be shortened item by item based on the indicators being met; Select the best items and intervals; Filter the items according to the best item and interval, and remove the item column from the data; Return to the step of determining whether the column item satisfies the indicator.

7. The detection equipment interval parameter adjustment method according to claim 6, characterized in that: The step S30 comprises: A new empty table M1 is created according to the defect library list, and table M1 is updated with new values ​​according to the optimal project interval algorithm. The updated items include: project name, left end value of interval, right end value of interval, B product screening rate value, A product screening rate value, and screening efficiency ratio.

8. The detection equipment interval parameter adjustment method according to claim 7, characterized in that: The step S40 comprises: If table M1 is empty, it means that no better indicators that meet the requirements can be obtained based on the existing defect library list; in this case, the defect library list needs to be further expanded to increase the balance and diversity of A / B categories; If the M1 table is not empty, update the parameters according to the corresponding items and verify the effect online.

9. A detection equipment interval parameter adjustment system, characterized in that: The system includes a memory, a processor, and a detection device interval parameter adjustment program stored on the processor, and the detection device interval parameter adjustment program executes the steps of the method according to any one of claims 1 to 8 when executed by the processor.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a detection device interval parameter adjustment program, and the detection device interval parameter adjustment program, when executed by a processor, performs the steps of the method according to any one of claims 1 to 8.

Images