CN115170314A - Method, system and device for automatically selecting identification channel in image identification field - Google Patents
Method, system and device for automatically selecting identification channel in image identification field Download PDFInfo
- Publication number
- CN115170314A CN115170314A CN202210816474.8A CN202210816474A CN115170314A CN 115170314 A CN115170314 A CN 115170314A CN 202210816474 A CN202210816474 A CN 202210816474A CN 115170314 A CN115170314 A CN 115170314A
- Authority
- CN
- China
- Prior art keywords
- identification
- weight
- channel
- identification channel
- heartbeat
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 45
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 43
- 238000012545 processing Methods 0.000 claims description 45
- 238000005070 sampling Methods 0.000 claims description 38
- 230000004044 response Effects 0.000 claims description 31
- 230000002159 abnormal effect Effects 0.000 claims description 25
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 6
- 230000000977 initiatory effect Effects 0.000 claims description 6
- 238000011084 recovery Methods 0.000 claims description 5
- 230000000694 effects Effects 0.000 abstract description 2
- 238000012015 optical character recognition Methods 0.000 description 5
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000013499 data model Methods 0.000 description 2
- 230000003247 decreasing effect Effects 0.000 description 2
- 239000000523 sample Substances 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012797 qualification Methods 0.000 description 1
- 238000010187 selection method Methods 0.000 description 1
- 238000010129 solution processing Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/04—Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06311—Scheduling, planning or task assignment for a person or group
- G06Q10/063114—Status monitoring or status determination for a person or group
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06312—Adjustment or analysis of established resource schedule, e.g. resource or task levelling, or dynamic rescheduling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- Finance (AREA)
- Educational Administration (AREA)
- Accounting & Taxation (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Technology Law (AREA)
- Multimedia (AREA)
- Debugging And Monitoring (AREA)
Abstract
The invention relates to the technical field of computer software application, and discloses a method, a system and a device for automatically selecting an identification channel in the field of image identification, wherein the technical scheme comprises the following key points: the weight setting module is used for setting initial weight for each identification channel; the weight adjusting module: executing a weight dynamic adjustment algorithm on all identification channels when an identification transaction request is received after a first time period; the selection execution module is used for operating a channel selection algorithm to select the identification channels according to the initial weight of each identification channel when the identification transaction request is received for the first time; and selecting the identification channel by operating the channel selection algorithm according to the states and weights of all the identification channels updated after the weight dynamic adjustment algorithm is executed, so that the real-time adjustment according to the operating states of all the identification channels is realized, and the effect that the system can select the optimal identification channel is ensured.
Description
Technical Field
The invention relates to the technical field of computer software application, in particular to a method, a system and a device for automatically selecting an identification channel in the field of image identification.
Background
In a bank scene, along with the continuous expansion of the bank business scale and customers, when the qualification of the customers or the reimbursement of staff are checked, the customers are required to upload invoice images (airline tickets, value-added tax special tickets, value-added tax general tickets, train tickets, taxi tickets and quota invoices), and then background business personnel manually input the ticket face information, the office efficiency is very low, so that the bank can introduce an OCR (optical character recognition) recognition system, the invoice images are transmitted to OCR recognition channel manufacturers, and the text information on the invoice images is returned to the bank after the OCR recognition channel manufacturers recognize and analyze the invoice images.
Considering the factors of stability of system operation, operation cost, response timeliness and the like, a bank generally accesses a plurality of OCR recognition channel manufacturers, and acceptance weight backgrounds of the recognition channel manufacturers can be matched, so that the transaction amount is distributed to the manufacturers in proportion.
Through data analysis after production and problem and solution processing in the operation process, the problems and disadvantages of the existing identification channel selection method are summarized as follows:
(1) The online users of each manufacturer are different in size, and particularly at the end of a month, the response aging is slow, and the use experience of the bank is seriously influenced.
(2) When a manufacturer has a problem, the problem cannot be monitored in real time and cannot be found in time.
(3) When the vendor route needs to be switched, the weight of switching needs to be configured manually.
(4) The identification channel routing is difficult to continuously monitor the availability of the identification channel due to insufficient sampling data or no sampling data at all, and the weight of the identification channel of each manufacturer is difficult to adjust.
Disclosure of Invention
The invention aims to provide a method, a system and a device for automatically selecting an identification channel in the field of image identification, which can achieve the effect of adjusting in real time according to the running state of each identification channel and ensuring that the system can select the best identification channel.
The technical purpose of the invention is realized by the following technical scheme: a method for automatically selecting an identification channel in the field of image identification comprises the following steps:
setting initial weight for each identification channel;
when an identification transaction request is received for the first time, operating a channel selection algorithm to select an identification channel according to the initial weight of each identification channel;
executing a weight dynamic adjustment algorithm on all identification channels when an identification transaction request is received after a first time period;
and operating a channel selection algorithm to select the identification channel according to the states and the weights of all the identification channels updated after the weight dynamic adjustment algorithm is executed.
As a preferred technical solution of the present invention, according to the identification channel cost data, an initial weight is set for each identification channel, and the principle of setting the initial weight is as follows: low cost weight of identification channel > high cost weight of identification channel.
As a preferred technical solution of the present invention, the content of the weight dynamic adjustment algorithm includes that asynchronously executed:
weight adjustment calculation of the normal state identification channel;
heartbeat probing processing of an abnormal state identification channel and a part of normal state identification channels:
weight recovery processing of the identified channels with reduced weight.
As a preferred technical solution of the present invention, the weight adjustment calculation of the normal state identification channel includes:
for the identification channel in the normal state, acquiring and calculating the transaction number of the current period, the identification success rate of the current period and the average response time of the current period in the identification channel at intervals of a first time period;
if the current period transaction number of the identification channel is greater than the minimum sampling number, sending an identification channel weight adjusting instruction; if the current period identification success rate of the identification channel is less than the first identification success rate, the heartbeat probing processing is initiated on the identification channel; if the current period identification success rate is larger than the first identification success rate or the current period transaction quantity is zero, the weight of the identification channel is not adjusted;
acquiring real-time operation data of the identification channel according to an identification channel weight adjusting instruction, if the current period identification success rate is zero, initiating heartbeat probing processing, if the current period identification success rate is not zero, judging whether the conditions that the current period identification success rate is larger than or equal to a second identification success rate and the current period average response time is smaller than or equal to a first response time are met at the same time, and if not, reducing the current weight of the identification channel; if yes, increasing the current weight of the identification channel;
and executing a weight range limiting algorithm on the adjusted weight data to obtain the operable weight of the identification channel and update the configuration.
As a preferred technical solution of the present invention, the weight range limiting algorithm includes:
when the identification success rate of the current period is less than the second identification success rate and/or the average response time of the current period is more than the first response time, reducing the current weight of the identification channel: if the reduced weight of the identification channel is more than or equal to the weight threshold, updating the reduced weight to a cache; if the reduced identification channel weight is less than the weight threshold value, updating the running state of the identification channel to be abnormal;
when the identification success rate of the current period is larger than or equal to the second identification success rate and the average response time of the current period is smaller than or equal to the first response time, increasing the current weight of the identification channel: if the increased identification channel weight is less than or equal to the maximum weight value, updating the increased weight to a cache; and if the added identification channel weight is larger than the maximum weight value, updating the maximum weight value to a cache.
As a preferred technical solution of the present invention, the heartbeat probing process comprises:
acquiring a first amount of sampling data in an identification channel, wherein the first amount of sampling data in the identification channel is derived from historical transaction data or manually preset sampling data before a first time period;
and sequentially carrying out heartbeat trial transaction on the identification channel through the first quantity of sampling data in the identification channel, and counting the success rate and the average response time of the trial transaction.
As a preferred embodiment of the present invention, the object of heartbeat heuristic processing includes: initiating heartbeat probing processing on the identification channel in the abnormal state; the heartbeat probing processing is initiated on the normal state identification channel which does not meet the minimum sampling quantity in the first time period and has the condition that the identification success rate of the current period of the identification channel is greater than the first identification success rate; in the first identification period, the transaction quantity of the identification channel in the current period is larger than the lowest sampling quantity, and the heartbeat probing processing is initiated to the identification channel in the normal state meeting the condition that the identification success rate of the current period is zero;
when the heartbeat heuristic processing is initiated on a normal state identification channel which does not meet the condition that the lowest sampling quantity is not met in a first time period and the identification success rate of the identification channel in the current period is larger than the first identification success rate condition, judging whether all heartbeat heuristic transactions fail, if so, updating the running state of the identification channel to be abnormal, and if not, not adjusting the weight of the identification channel;
when the heartbeat probing processing is initiated on the normal state identification channel which has the condition that the current period transaction quantity of the identification channel is greater than the lowest sampling quantity and the current period identification success rate is zero in the first identification period: if the heartbeat tentative transactions are successful, the weight of the identification channel is not adjusted; if the heartbeat heuristic transaction is not completely successful, reducing the weight of the identification channel; if all heartbeat probes fail, judging the state of the identification channel through an availability algorithm;
when the identification channel in the abnormal state is initiated with heartbeat tentative processing, judging whether the number of successful heartbeat tentative transactions reaches a heartbeat tentative transaction success threshold value, and if so, updating the operation state of the identification channel to be normal, and setting the weight of the identification channel to be a first proportion of the original weight of the identification channel; and if not, executing the independent timing task.
As a preferred technical solution of the present invention, the availability algorithm is: and judging whether the result of (the total number of successful transactions + the number of successful heartbeat trial transactions) plus the minimum transaction number-the total number of failed transactions is greater than 0, if so, judging that the state of the identification channel is normal, and if not, judging that the state of the identification channel is abnormal.
As a preferred technical solution of the present invention, the weight recovery processing includes executing an independent timing task, wherein the independent timing task includes: for the normal state identification channel with the reduced weight, if at least M transaction processes are executed in N periods after the period with the reduced weight is executed and all the transaction processes are successful, the weight is recovered; and if the abnormal channel which fails in heartbeat heuristic processing executes at least M heartbeat heuristic transactions in N periods after the heartbeat heuristic processing fails and all the heartbeat heuristic transactions succeed, the state of the abnormal channel is changed to be normal and the weight is restored.
As a preferred technical solution of the present invention, the operating channel selection algorithm includes:
selecting all identification channels with normal operation states;
generating a numerical range set according to the weights of all the identification channels and the sum of the weights of all the identification channels, wherein the weight of each identification channel in the numerical range set has a continuous numerical range subset without intersection;
generating a random number in the set of value ranges;
selecting an identification channel corresponding to a subset of the continuous range of values containing the random number to conduct a transaction.
As a preferred embodiment of the present invention, when the weight of the identification channel is increased, the adjusted weight F is used Rear end ＝F Front part + b, when the weight of the identified channel is decreased, the adjusted weight F Rear end ＝F Front side B, wherein F Front side The identified channel weights for the previous cycle, and b is the weight adjustment cardinality.
A system for automatically selecting a recognition channel in the field of image recognition, comprising:
the weight setting module is used for setting initial weight for each identification channel;
the weight adjusting module: executing a weight dynamic adjustment algorithm on all identification channels when an identification transaction request is received after a first time period;
the selection execution module is used for operating a channel selection algorithm to select the identification channels according to the initial weight of each identification channel when the identification transaction request is received for the first time; and operating a channel selection algorithm to select the identification channel according to the states and the weights of all the identification channels updated after the weight dynamic adjustment algorithm is executed.
An apparatus for automatically selecting a recognition channel in the field of image recognition, comprising: a processor and a memory, said memory storing a computer program executable by said processor, said processor implementing the above method when executing said computer program.
In conclusion, the invention has the following beneficial effects: 1. under the condition that service response of one manufacturer identification channel is bad, manual parameter configuration modification is not needed, switching to other identification channels is fast performed, and user experience is improved.
2. The dynamic adjustment can be carried out in real time according to the running state of each identification channel, and the system can be ensured to select the optimal identification channel.
3. The identification channel can be set according to the actual weight factor when the initial weight of the identification channel is set, for example, the cost is set, and the identification channel can be selected in practice by emphasizing the weight of the low-cost identification channel, so that the efficient operation of the identification work is ensured, the cost control is continued, and the identification cost is saved to the maximum extent.
4. The method has a complete identification channel weight adjusting scheme, is completely adjusted and selected by the system, is high in adjusting and selecting speed and more accurate, and ensures high-efficiency operation of actual identification.
Drawings
FIG. 1 is a flow chart of the method of the present invention.
FIG. 2 is a schematic diagram of the weight adjustment calculation of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings.
As shown in fig. 1 and 2, the present invention provides a method for automatically selecting an identification channel in the field of image identification, and the method steps are executed by corresponding system modules, and the specific steps are as follows:
s1, setting initial weights for all identification channels through a weight setting module; specifically, according to the identification channel cost data, an initial weight is set for each identification channel, and the principle of setting the initial weight is as follows: low cost weight of identification channel > high cost weight of identification channel. Since the subsequent weight adjustment is adjusted according to the recognition efficiency, the weight of the tendency is set according to the cost from the beginning, so that the cost and the efficiency can be selected and used at the same time.
S2, through the selection execution module, when an identification transaction request is received for the first time, operating a channel selection algorithm to select an identification channel according to the initial weight of each identification channel; since there is no actual operation data, at the beginning of operation, the states of all the identification channels are defaulted to be normal, and all the identification channels are taken as selection objects.
S3, a weight adjusting module executes a weight dynamic adjusting algorithm on all the identification channels when receiving the identification transaction request after the first time period;
the content of the weight dynamic adjustment algorithm comprises that the following steps are executed asynchronously: A. weight adjustment calculation of the normal state identification channel; B. heartbeat probing processing of the abnormal state identification channel and the partial normal state identification channel; C. the weight recovery process for the identified channels with reduced weights.
A. The weight adjustment calculation of the normal state identification channel comprises the following steps:
a1, acquiring and calculating transaction data in an identification channel at intervals of a first time period, namely request time, response ending time and a response code of each transaction, for the identification channel in a normal state; the response code comprises success and failure and is used for indicating whether the transaction is successful or not, so that the transaction number in the current period, the identification success rate in the current period and the average response time in the current period can be calculated;
examples are: the obtained transaction data model is shown in the following table 1:
TABLE 1
A2, if the current period transaction number of the identification channel is greater than the minimum sampling number, sending an identification channel weight adjusting instruction; if the current period identification success rate of the identification channel is less than the first identification success rate, heartbeat probing processing is initiated on the identification channel; if the current period identification success rate is larger than the first identification success rate or the current period transaction number is zero, the weight of the identification channel is not adjusted, wherein if the transaction number in the period is 0, the current time period is a low-peak working period, heartbeat probing processing is not needed, a default channel is available, and the weight adjustment is not needed due to insufficient sampling number;
a3, obtaining real-time operation data of the identification channel according to an identification channel weight adjusting instruction, if the current period identification success rate is zero, initiating heartbeat probing processing, if the current period identification success rate is not zero, judging whether the conditions that the current period identification success rate is larger than or equal to a second identification success rate and the current period average response time is smaller than or equal to a first response time are met at the same time, and if not, reducing the current weight of the identification channel; if yes, increasing the current weight of the identification channel;
and A4, executing a weight range limiting algorithm on the adjusted weight data to obtain the operable weight of the identification channel and updating the configuration.
Wherein, the weight range limiting algorithm comprises:
when the identification success rate of the current period is less than the second identification success rate and/or the average response time of the current period is more than the first response time, reducing the current weight of the identification channel: if the reduced weight of the identification channel is more than or equal to the weight threshold, updating the reduced weight to a cache; if the reduced identification channel weight is less than the weight threshold value, updating the running state of the identification channel to be abnormal;
when the identification success rate of the current period is larger than or equal to the second identification success rate and the average response time of the current period is smaller than or equal to the first response time, increasing the current weight of the identification channel: if the increased identification channel weight is less than or equal to the maximum weight value, updating the increased weight to a cache; and if the added identification channel weight is larger than the maximum weight value, updating the maximum weight value to a cache.
Specifically, when the weight of the identified channel is increased, the adjusted weight F is Rear end ＝F Front side + b, when the weight of the identified channel is decreased, the adjusted weight F Rear end ＝F Front side B, wherein F Front side The identified channel weights for the previous cycle, and b is the weight adjustment cardinality. For example, b is set to 5, each time the channel weight is increased, F Rear end ＝F Front side +5, each time the channel weight is reduced, F Rear end ＝F Front side -5。
B. Heartbeat heuristic processing of the abnormal state identification channel and the partial normal state identification channel:
specifically, the heartbeat heuristic processing process includes:
acquiring a first amount of sampling data in an identification channel, wherein the first amount of sampling data in the identification channel is derived from historical transaction data or manually preset sampling data before a first time period;
examples are as follows: sampling data collection is carried out according to the day, any 3 sampling data successfully identified yesterday are stored, the sampling data comprise image data and identification return results, and meanwhile, it needs to be noted that each channel needs to store 3 successfully identified data.
Specifically, heartbeat sampling data is updated through a single daily timing task, and the time is 00 points per day;
the sampling data acquisition rule is as follows:
the method comprises the steps that (1) online is carried out for the first time, all channels sample data, and the data can be preset manually, namely background sampling is developed;
if the channel has no request data yesterday, sampling data are taken from the identification channel with the maximum transaction amount;
if all channels have no request data yesterday, taking data of one week from the channel with the largest transaction amount as collected data;
and if all the channels have no request data for one week, early warning by mails and short messages and asking for manual intervention.
After the sampling data are obtained, heartbeat tentative transactions are carried out on the identification channel sequentially through the first quantity of sampling data in the identification channel, and the success rate and the average response time of the tentative transactions are counted.
Objects of heartbeat heuristic processing include: initiating heartbeat probing processing on the identification channel in the abnormal state; the heartbeat probing processing is initiated on the normal state identification channel which does not meet the minimum sampling quantity in the first time period and has the condition that the identification success rate of the current period of the identification channel is greater than the first identification success rate; in the first identification period, the normal state identification channel which has the current period transaction number larger than the lowest sampling number and meets the condition that the current period identification success rate is zero initiates heartbeat probing processing;
when the heartbeat heuristic processing is initiated on a normal state identification channel which does not meet the condition that the lowest sampling quantity is not met in a first time period and the identification success rate of the identification channel in the current period is larger than the first identification success rate condition, judging whether all heartbeat heuristic transactions fail, if so, updating the running state of the identification channel to be abnormal, and if not, not adjusting the weight of the identification channel;
when the normal state identification channel which is in the first identification period and has the current period transaction number larger than the lowest sampling number and meets the condition that the current period identification success rate is zero initiates heartbeat probing processing: if the heartbeat heuristic transactions are successful, the weight of the identification channel is not adjusted; if the heartbeat heuristic transaction is not completely successful, reducing the weight of the identification channel; if the heartbeat probing fails, judging the state of the identification channel through an availability algorithm;
when the identification channel in the abnormal state is initiated with heartbeat probing processing, judging whether the number of successful heartbeat probing transactions reaches a heartbeat probing transaction success threshold value or not, and if so, updating the operation state of the identification channel to be normal, and setting the weight of the identification channel to be a first proportion of the original weight of the identification channel; and if not, executing the independent timing task. Wherein the first ratio may be one-half.
The usability algorithm is: and judging whether the result of (the total number of successful transactions + the number of successful heartbeat trial transactions) plus the minimum transaction number-the total number of failed transactions is greater than 0, if so, judging that the state of the identification channel is normal, and if not, judging that the state of the identification channel is abnormal. It is noted that prior to the dynamic adjustment of the weights, there is also a determination regarding the status of the identified channels, and that if the status of the identified channels is ambiguous or the point in time of status update is reached, the status of the identified channels is updated using an availability algorithm.
Specific examples are:
taking an image identification channel identifier a as an example, request time period data is constructed:
1) Starting from 00;
2) 00-00;
3) Recording: the method comprises the following steps that an image recognition channel identifier A, the total times of transaction recognition in a time period, the total times of transaction failure in the time period and the total times of transaction success in the time period are identified;
4) Similarly, 1 5-minute data is recorded again;
5) The service data under the scene are respectively recorded as follows:
channel A1, 10 requests, 1 success, 9 failures;
channel A1, 10 requests, all failures (different identification types);
channel A1, 1 request, all failures (different identification types);
the data model is shown in table 2 below:
TABLE 2
C. Weight recovery processing of the identified channels with reduced weight.
The weight restoration process includes executing independent timing tasks, wherein the independent timing tasks include: for the normal state identification channel with the reduced weight, if at least M times of transaction processing are executed in N periods after the period for reducing the weight is executed and the transaction processing is successful, the weight is recovered; for example, if the first time period is five minutes, 6 first time periods are selected for statistics, and if 3 transactions are successful within 30 minutes, indicating that the identified channel can perform normal efficient transactions, the weight of the identified channel is restored.
For the abnormal channel which fails in heartbeat heuristic processing, if at least M times of heartbeat heuristic transactions are executed in N periods after the heartbeat heuristic processing fails and all the transactions are successful, the state of the abnormal channel is converted into normal and the weight is recovered.
And S4, running a channel selection algorithm to select the identification channel according to the states and weights of all the identification channels updated after the weight dynamic adjustment algorithm is executed by the selection execution module.
The operating channel selection algorithm comprises:
selecting all identification channels with normal operation states;
generating a numerical range set according to the weights of all the identification channels and the sum of the weights of all the identification channels, wherein the weight of each identification channel in the numerical range set has a continuous numerical range subset without intersection;
generating a random number in the set of value ranges;
an identification channel corresponding to a subset of the continuous range of values containing the random number is selected for the transaction.
Examples are as follows: assuming three channels A, B and C, the configured weight ratio is x: y: z respectively,
generating a random number rand (x + y + z) before 0 to x + y + z when a transaction request is received;
when rand (x + y + z) < x, go A channel;
when x = < rand (x + y + z) < x + y, go B-channel;
when x + y = < rand (x + y + z) < x + y + z, then the C channel is followed.
Corresponding to the method and the system, the invention also provides a device for automatically selecting the identification channel in the field of image identification, which comprises the following steps: a processor and a memory, the memory storing a computer program executable by the processor, the processor implementing the method when executing the computer program.
The invention has the advantages that: under the condition that the service response of one manufacturer channel is poor, manual parameter configuration modification is not needed, and the channel can be quickly switched to other channels, so that the user experience is improved. The dynamic adjustment can be carried out in real time according to the running state of each channel, and the system can be ensured to select the optimal channel. When the initial weight of the channel is set, the channel can be set according to the actual weight factor, for example, the cost is set, and when the channel is actually selected, the efficient operation of the identification work is ensured, and meanwhile, the cost control can be continued, so that the identification cost is saved to the maximum extent. The method has a complete channel weight adjusting scheme, the system performs adjustment and channel selection completely, the adjusting and selecting speed is high, the accuracy is high, and the high-efficiency operation of actual identification is ensured.
The above description is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments, and all technical solutions belonging to the idea of the present invention belong to the protection scope of the present invention. It should be noted that modifications and adaptations to those skilled in the art without departing from the principles of the present invention should also be considered as within the scope of the present invention.
Claims (10)
1. A method for automatically selecting an identification channel in the field of image identification is characterized in that: the method comprises the following steps:
setting initial weight for each identification channel;
when an identification transaction request is received for the first time, operating a channel selection algorithm to select an identification channel according to the initial weight of each identification channel;
when an identification transaction request is received after a first time period, executing a weight dynamic adjustment algorithm on all identification channels;
and operating a channel selection algorithm to select the identification channel according to the states and the weights of all the identification channels updated after the weight dynamic adjustment algorithm is executed.
2. The method of claim 1, wherein the method comprises the following steps: according to the identification channel cost data, setting an initial weight for each identification channel, wherein the principle of setting the initial weight is as follows: low cost identification channel weight > high cost identification channel weight.
3. The method of claim 1, wherein the method comprises the following steps: the content of the weight dynamic adjustment algorithm comprises that the following steps are executed asynchronously:
weight adjustment calculation of the normal state identification channel;
heartbeat heuristic processing of the abnormal state identification channel and the partial normal state identification channel:
weight recovery processing of the identified channels with reduced weight.
4. The method of claim 3, wherein the method comprises the following steps: the weight adjustment calculation of the normal state identification channel comprises the following steps:
for the identification channel in the normal state, acquiring and calculating the transaction number of the current period, the identification success rate of the current period and the average response time of the current period in the identification channel every other first time period;
if the current period transaction number of the identification channel is greater than the minimum sampling number, sending an identification channel weight adjusting instruction; if the current period identification success rate of the identification channel is less than the first identification success rate, the heartbeat probing processing is initiated on the identification channel; if the current period identification success rate is larger than the first identification success rate or the current period transaction quantity is zero, the weight of the identification channel is not adjusted;
acquiring real-time operation data of the identification channel according to an identification channel weight adjusting instruction, initiating heartbeat probing processing if the current period identification success rate is zero, judging whether the conditions that the current period identification success rate is larger than or equal to a second identification success rate and the current period average response time is smaller than or equal to a first response time are met simultaneously if the current period identification success rate is not zero, and reducing the current weight of the identification channel if the conditions are not met; if yes, increasing the current weight of the identification channel;
and executing a weight range limiting algorithm on the adjusted weight data to obtain the operable weight of the identification channel and update the configuration.
5. The method of claim 4, wherein the method comprises the following steps: the weight range limiting algorithm comprises:
when the identification success rate of the current period is less than the second identification success rate and/or the average response time of the current period is more than the first response time, reducing the current weight of the identification channel: if the reduced weight of the identification channel is more than or equal to the weight threshold, updating the reduced weight to a cache; if the reduced weight of the identification channel is less than the weight threshold value, updating the running state of the identification channel to be abnormal;
when the identification success rate of the current period is larger than or equal to the second identification success rate and the average response time of the current period is smaller than or equal to the first response time, increasing the current weight of the identification channel: if the increased identification channel weight is less than or equal to the maximum weight value, updating the increased weight to a cache; and if the increased weight of the identification channel is greater than the maximum weight value, updating the maximum weight value to a cache.
6. The method of claim 4, wherein the method comprises the following steps: the heartbeat heuristic process comprises the following steps:
acquiring a first amount of sampling data in an identification channel, wherein the first amount of sampling data in the identification channel is derived from historical transaction data or manually preset sampling data before a first time period;
conducting heartbeat trial transaction on the identification channel sequentially through the first quantity of sampling data in the identification channel, and counting the success rate and the average response time of the trial transaction;
objects of heartbeat heuristic processing include: initiating heartbeat probing processing on the identification channel in the abnormal state; the heartbeat probing processing is initiated on the normal state identification channel which does not meet the minimum sampling quantity in the first time period and has the condition that the identification success rate of the current period of the identification channel is greater than the first identification success rate; in the first identification period, the transaction quantity of the identification channel in the current period is larger than the lowest sampling quantity, and the heartbeat probing processing is initiated to the identification channel in the normal state meeting the condition that the identification success rate of the current period is zero;
when the heartbeat heuristic processing is initiated on a normal state identification channel which does not meet the condition that the lowest sampling quantity is not met in a first time period and the identification success rate of the identification channel in the current period is larger than the first identification success rate condition, judging whether all heartbeat heuristic transactions fail, if so, updating the running state of the identification channel to be abnormal, and if not, not adjusting the weight of the identification channel;
when the normal state identification channel which is in the first identification period and has the current period transaction number larger than the lowest sampling number and meets the condition that the current period identification success rate is zero initiates heartbeat probing processing: if the heartbeat tentative transactions are successful, the weight of the identification channel is not adjusted; if the heartbeat heuristic transaction is not completely successful, reducing the weight of the identification channel; if the heartbeat probing fails, judging the state of the identification channel through an availability algorithm;
when the identification channel in the abnormal state is initiated with heartbeat tentative processing, judging whether the number of successful heartbeat tentative transactions reaches a heartbeat tentative transaction success threshold value, and if so, updating the operation state of the identification channel to be normal, and setting the weight of the identification channel to be a first proportion of the original weight of the identification channel; if not, executing an independent timing task on the mobile terminal;
wherein, the usability algorithm is as follows: and judging whether the result of (the total number of successful transactions + the number of successful heartbeat trial transactions) the minimum transaction number-the total number of failed transactions is greater than 0, if so, judging that the state of the identification channel is normal, and if not, judging that the state of the identification channel is abnormal.
7. The method of claim 6, wherein the method comprises: the weight restoration process includes executing independent timing tasks, wherein the independent timing tasks include: for the normal state identification channel with the reduced weight, if at least M transaction processes are executed in N periods after the period with the reduced weight is executed and all the transaction processes are successful, the weight is recovered; and if the abnormal channel which fails in heartbeat heuristic processing executes at least M heartbeat heuristic transactions in N periods after the heartbeat heuristic processing fails and all the heartbeat heuristic transactions succeed, the state of the abnormal channel is changed to be normal and the weight is restored.
8. The method of claim 1, wherein the method comprises the following steps: the operating channel selection algorithm includes:
selecting all identification channels with normal operation states;
generating a numerical range set according to the weights of all the identification channels and the sum of the weights of all the identification channels, wherein the weight of each identification channel in the numerical range set has a continuous numerical range subset without intersection;
generating a random number in the set of value ranges;
selecting an identification channel corresponding to a subset of the continuous range of values containing the random number to conduct a transaction.
9. A system for automatically selecting an identification channel in the field of image identification is characterized in that: the method comprises the following steps:
the weight setting module is used for setting initial weights for all the identification channels;
a weight adjustment module: when an identification transaction request is received after a first time period, executing a weight dynamic adjustment algorithm on all identification channels;
the selection execution module is used for operating a channel selection algorithm to select the identification channel according to the initial weight of each identification channel when the identification transaction request is received for the first time; and operating a channel selection algorithm to select the identification channel according to the states and weights of all the identification channels updated after the weight dynamic adjustment algorithm is executed.
10. A device for automatically selecting an identification channel in the field of image identification is characterized in that: the method comprises the following steps: a processor and a memory, the memory storing a computer program executable by the processor, the processor implementing the method of any one of claims 1-8 when executing the computer program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210816474.8A CN115170314A (en) | 2022-07-12 | 2022-07-12 | Method, system and device for automatically selecting identification channel in image identification field |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210816474.8A CN115170314A (en) | 2022-07-12 | 2022-07-12 | Method, system and device for automatically selecting identification channel in image identification field |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115170314A true CN115170314A (en) | 2022-10-11 |
Family
ID=83493765
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210816474.8A Pending CN115170314A (en) | 2022-07-12 | 2022-07-12 | Method, system and device for automatically selecting identification channel in image identification field |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115170314A (en) |
-
2022
- 2022-07-12 CN CN202210816474.8A patent/CN115170314A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110086666B (en) | Alarm method, device and system | |
| CN109104336B (en) | Service request processing method and device, computer equipment and storage medium | |
| CN111917878A (en) | Message processing method, device, equipment and storage medium | |
| CN111160329A (en) | Root cause analysis method and device | |
| CN111625527A (en) | Out-of-order data processing method, device and equipment and readable storage medium | |
| CN111400352B (en) | Workflow engine capable of processing data in batches | |
| CN113268403A (en) | Time series analysis and prediction method, device, equipment and storage medium | |
| CN115174686B (en) | Method and device for dynamically adjusting weights of multiple service channels based on service efficiency | |
| CN114579654B (en) | Unified operation management method and platform system for multi-payment settlement system of bank | |
| CN111291054A (en) | Data processing method and device, computer equipment and storage medium | |
| CN110191182A (en) | Distributed document batch processing method, device, equipment and readable storage medium storing program for executing | |
| CN113626495A (en) | Operation data acquisition method of energy management system | |
| CN113610516A (en) | Payment channel display method and device, electronic equipment and storage medium | |
| CN115170314A (en) | Method, system and device for automatically selecting identification channel in image identification field | |
| CN115098238B (en) | Application program task scheduling method and device | |
| CN111506455A (en) | Checking method and device for service release result | |
| CN113657817B (en) | Transaction processing method and device, electronic equipment and readable storage medium | |
| CN116501468A (en) | Batch job processing method and device and electronic equipment | |
| US11200547B2 (en) | Payment collection control method and device, server, and computer-readable storage medium | |
| CN111586100B (en) | Target object message sending device and method | |
| CN114676177A (en) | Financial index determination method, device, equipment, medium and product | |
| CN114201369A (en) | Server cluster management method and device, electronic equipment and storage medium | |
| CN113159941A (en) | Intelligent streaming transaction execution method and device | |
| CN112329883A (en) | Model training system, method, device and storage medium | |
| CN113052674A (en) | Quota management method, electronic device, computer device, and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information |
Country or region after: China Address after: No.4 building, Hexi Financial City, Jianye District, Nanjing City, Jiangsu Province, 210000 Applicant after: Jiangsu Sushang Bank Co.,Ltd. Address before: No.4 building, Hexi Financial City, Jianye District, Nanjing City, Jiangsu Province, 210000 Applicant before: JIANGSU SUNING BANK Co.,Ltd. Country or region before: China |
|
| CB02 | Change of applicant information |