CN113791034B - Sample collection and classification supervision system for steel rail flaw detection - Google Patents
Sample collection and classification supervision system for steel rail flaw detection Download PDFInfo
- Publication number
- CN113791034B CN113791034B CN202111161213.9A CN202111161213A CN113791034B CN 113791034 B CN113791034 B CN 113791034B CN 202111161213 A CN202111161213 A CN 202111161213A CN 113791034 B CN113791034 B CN 113791034B
- Authority
- CN
- China
- Prior art keywords
- sample
- data
- loss
- value
- sorting
- 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.)
- Active
Links
- 229910000831 Steel Inorganic materials 0.000 title claims abstract description 145
- 239000010959 steel Substances 0.000 title claims abstract description 145
- 238000001514 detection method Methods 0.000 title claims abstract description 45
- 238000012544 monitoring process Methods 0.000 claims abstract description 13
- 238000012163 sequencing technique Methods 0.000 claims description 75
- 238000012545 processing Methods 0.000 claims description 44
- 238000012360 testing method Methods 0.000 claims description 20
- 238000000034 method Methods 0.000 claims description 17
- 238000005070 sampling Methods 0.000 claims description 10
- 238000003491 array Methods 0.000 claims description 9
- 230000002159 abnormal effect Effects 0.000 description 4
- 230000010354 integration Effects 0.000 description 2
- 241001669679 Eleotris Species 0.000 description 1
- 238000007792 addition Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000006996 mental state Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/01—Arrangements or apparatus for facilitating the optical investigation
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B61—RAILWAYS
- B61K—AUXILIARY EQUIPMENT SPECIALLY ADAPTED FOR RAILWAYS, NOT OTHERWISE PROVIDED FOR
- B61K9/00—Railway vehicle profile gauges; Detecting or indicating overheating of components; Apparatus on locomotives or cars to indicate bad track sections; General design of track recording vehicles
- B61K9/08—Measuring installations for surveying permanent way
- B61K9/10—Measuring installations for surveying permanent way for detecting cracks in rails or welds thereof
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/01—Arrangements or apparatus for facilitating the optical investigation
- G01N2021/0181—Memory or computer-assisted visual determination
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8854—Grading and classifying of flaws
- G01N2021/888—Marking defects
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Analytical Chemistry (AREA)
- General Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Pathology (AREA)
- Biochemistry (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Mechanical Engineering (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention relates to the technical field of sample classification management, in particular to a sample collection and classification supervision system for steel rail flaw detection, which comprises a sample collection unit, a sample recording unit, a sample position unit, a sample management monitoring unit, a sample classification unit and a sample storage and display unit, wherein the sample collection unit is used for collecting and classifying samples; the sample set unit is used for acquiring sample base information related to the steel rail corresponding to the road section and transmitting the sample base information to the sample position unit; the sample recording unit is used for storing sample recording information related to previous sample collection, and the sample recording information comprises steel recording data, sample loss data, sample good data and a sample loss image.
Description
Technical Field
The invention relates to the technical field of sample classification management, in particular to a sample collection and classification supervision system for steel rail flaw detection.
Background
The steel rail is one of railway transportation infrastructures, guides the advancing direction of wheels of a train, directly bears the acting force of the wheels of the train and transmits the acting force to the sleeper, and the steel rail is easy to generate various fatigue cracks under the long-term and repeated action of the wheels. The service state of the steel rail directly influences the transportation safety of the railway, and the detection, identification and judgment of the damage of the steel rail in service of the railway are the premise and the basis for making a line maintenance decision.
Along with the continuous speeding up of the railway, the transportation quantity is continuously increased, the damage of the steel rail is obviously increased, the damage types are also various, the playback workload of analysts is large, the analysis speed is low, the intelligent degree is low, and due to the factors of the physical state, the mental state, the service level and the like of the analysts, the missing judgment and the misjudgment are easily caused, and the potential safety hazard is left for the safe operation of the railway.
However, on the premise of accurately judging the damage, the steel rail needs to be sampled, however, all data are put in a storage unit after the existing sampling, the collected samples are not classified and stored, and meanwhile, the collected samples are not arranged in sequence, so that technicians spend a large amount of time for sorting the stored samples when analyzing the stored samples, and the identification time is easily increased.
Disclosure of Invention
The invention aims to provide a sample collection and classification supervision system for steel rail flaw detection, which is characterized in that corresponding judgment conditions are obtained through the integration processing of new sample data, primary data classification judgment is carried out on related data according to the judgment conditions, secondary division judgment is carried out according to the classification judgment result of the primary data, secondary division judgment is carried out according to the result of the secondary division judgment, and therefore the accurate classification of the sample data is realized, and the sorting and the storage are carried out according to the classification data.
The purpose of the invention can be realized by the following technical scheme: a sample collection and classification supervision system for steel rail flaw detection comprises a sample collection unit, a sample recording unit, a sample position unit, a sample pipe monitoring unit, a sample classification unit and a sample storage and display unit;
the sample set unit is used for acquiring sample base information related to the steel rail corresponding to the road section and transmitting the sample base information to the sample position unit;
the sample recording unit is used for storing sample recording information related to previous sample collection, and the sample recording information comprises steel recording data, sample loss data, sample good data and a sample loss image;
the sample position unit is used for carrying out sample position operation on the steel recording data, the sample loss data, the sample good data and the sample base information to obtain a damage mean value and a sample loss image and first detected steel rail sequencing data or second detected steel rail sequencing data, transmitting the first detected steel rail sequencing data or the second detected steel rail sequencing data to the sample set unit, and transmitting the damage mean value and the sample loss image to the sample pipe monitoring unit;
the sample collection unit carries out data acquisition according to the first detection steel rail sequencing data or the second detection steel rail sequencing data, so as to carry out sample acquisition on the steel rails of the corresponding road section, mark the acquired samples as sample detailed information, and transmit the sample detailed information to the sample management monitoring unit, wherein the sample detailed information comprises sample image data, sample storage data, sample name data and sample division data;
the sample management monitoring unit is used for carrying out sample grade processing on the damaged mean value and the damaged sample image together with the sample image data, the sample storage data, the sample name data and the sample grade data to obtain sample storage sequencing, and transmitting the sample storage sequencing to the sample storage and display unit;
the sample storage and display unit is used for storing and displaying the sample storage and sorting.
Further, the specific operation process of the sample processing is as follows:
selecting a plurality of marking data, sequencing the section steel rails corresponding to the plurality of marking data in sequence, marking the section steel rails corresponding to the plurality of marking data in sequence as sample base sequencing, and performing sample base treatment according to the sample base sequencing to obtain a damage mean SJ and a multiple rule value u 1;
extracting sample base information, marking the steel rail of each road section in the sample base information as steel sampling data, selecting the steel sampling data corresponding to all the road sections, and sequentially arranging the steel sampling data corresponding to all the road sections, wherein the specific steps are as follows: selecting one of the road section starting points, sequentially arranging and marking the road section starting points towards a set end point direction according to the road section starting points, and marking the corresponding steel acquisition data as Gi, wherein i =1,2,3, i.
The method comprises the steps of identifying the total amount i of steel rails corresponding to all road sections in a steel rail mark Gi, selecting steel rails with i =1,2 and 3, marking the steel rails with i =1,2 and 3 as test samples, carrying out flaw detection on the test samples through a steel rail flaw detector, marking a result image detected by the steel rail flaw detector as a test image, and matching the test image with the damage image to obtain first detected steel rail sequencing data and second detected steel rail sequencing data.
Further, the specific processing procedure of sample basis processing according to the sample basis sorting comprises:
marking the sample base sequence with a digital mark Aa, wherein a is a positive integer, marking the sample base sequence according to sample loss data and sample good data, marking the sample loss data and the sample good data with Y-and Y + respectively, and marking the sample base sequence with the digital mark Aa according to the Y-and Y + so as to obtain a mark AaY-or AaY +;
selecting a plurality of sample loss data AaY-corresponding to the steel recording data, listing AaY-one by one to obtain a plurality of AaY-arrays, selecting a numerical value in each AaY-array, counting the total number of each group AaY-, summing the total numbers of a plurality of groups AaY-, calculating the sum of the total number of AaY-, dividing the sum of the AaY-total number by the number of the groups AaY-total number to obtain the average value of the total numbers of a plurality of groups AaY-, and calibrating the average value as the sample loss average value;
the total number of AaY-and the number of corresponding Aa are subjected to proportion calculation by a plurality of groups, the damage proportion is calculated, the average value of a plurality of damage proportions is calculated, the damage average value is calculated, and the damage average value is marked as SJ;
selecting the number of words with the same a data in a plurality of groups AaY-arrays, calibrating the number of words as a heavy sample loss value, selecting the minimum value of the a data in a plurality of groups AaY-arrays, calibrating the minimum value as a minimum sample loss value, selecting the sorting difference value between the key sample loss value and the minimum sample loss value in a plurality of groups, carrying out mean value calculation on the sorting difference value, calculating an interval mean value, and bringing the minimum sample loss value, the key sample loss value and the interval mean value into a sample loss relation calculation formula: dense sample loss value = minimum sample loss value + mean interval value u1 e, where u1 is expressed as a fold-law value and e is expressed as a sample loss influence regulatory factor.
Further, matching the test image with the loss sample image specifically comprises:
when the data which are the same as the test image are matched from the damaged sample image, the damaged sample image is calibrated to be a damaged image to generate a damaged image signal, and when the data which are the same as the test image cannot be matched from the damaged sample image, the damaged sample image is calibrated to be a lossless image to generate a lossless signal;
extracting a damaged map signal and a lossless signal, identifying the damaged map signal and the lossless signal, judging damaged values when the corresponding i =1,2 and 3 are identified when the damaged map signal is identified, selecting the minimum value of the damaged values, calibrating the minimum value as a first detection steel rail, bringing the value of i corresponding to the first detection steel rail into a sample damage relation calculation formula, selecting the interval mean value as 2, calculating the next detection steel rail, and performing analogism calculation in sequence, thereby calculating the sequencing data of the detection steel rail;
when a lossless signal is identified, selecting a numerical value corresponding to the minimum sample loss value, carrying out numerical value replacement according to a sample loss relation calculation formula, converting the minimum sample loss value into a heavy sample loss value, wherein the calculation formula is as follows: and (3) calculating a first detection steel rail according to the minimum sample loss value = the first detection steel rail + the interval mean value u1 e, wherein the interval mean value is 1, calibrating the minimum sample loss value as a second detection steel rail, calculating a sample sequence numerical value with the interval mean value of 2 according to a sample loss relation calculation formula, calibrating the sample sequence numerical value as a third detection steel rail, and repeating the steps to calculate sequencing data of the two detection steel rails.
Further, the specific treatment process of sample fraction treatment is as follows:
selecting sample name data, extracting corresponding sample image data according to the sample name data, matching the sample image data with a loss sample image, and generating a real loss signal and a real no-signal according to the consistency or inconsistency of the matching results;
identifying real loss signals and real no signals, calibrating sample image data into positive track sample data when the real no signals are identified, and calibrating the sample image data into loss track sample data when the real loss signals are identified;
calculating the real loss rate according to the occurrence number of the real loss signals and the real no-signals, comparing the real loss rate with the damage mean value SJ to obtain heavy loss signals and low loss signals, and marking the heavy loss signals and the low loss signals as rate loss signals;
extracting corresponding sample name data and sample division data according to the normal orbit sample data, and carrying out sample division judgment on the sample division data and the rate loss signal, wherein the sample division judgment specifically comprises the following steps: sorting sample data corresponding to the sample name data from large to small according to the positive orbit sample data to obtain positive sorting data, identifying the rate loss signal, storing and sorting the positive sorting data and a plurality of corresponding sample storage data to obtain positive sorting data when the identification result is the heavy loss signal, and sorting and storing the sample name data and the corresponding sample image data according to the positive sorting data;
when the identification result is a low-loss signal, setting a sample preset value M1, matching the sample preset value M1 with the forward sorting data, matching a value corresponding to M1 in the forward sorting data, marking the value as a judgment value, marking the forward sorting data which is arranged in front of the judgment value in the forward sorting data as suspected data, marking the forward sorting data which is arranged behind the judgment value in the forward sorting data as positive data, and processing the suspected data and the positive data according to a processing mode of storage sorting processing, so as to obtain suspected storage sorting data corresponding to the suspected data and positive storage sorting data corresponding to the positive data;
carrying out the same processing on sample name data and sample score data corresponding to the damaged rail sample data according to the positive value sorting data, the suspected storage sorting data and the positive similar storage sorting data to obtain damaged value sorting data, damaged positive storage sorting data and damaged differential storage sorting data;
and extracting final identification data of positive value sorting data, suspected storage sorting data, positive similar storage sorting data, damaged value sorting data, damaged positive storage sorting data and damaged different storage sorting data and marking the identification data as sample storage sorting.
Further, the specific processing procedure of the storage sorting processing is as follows:
sequencing a plurality of sample storage data from large to small to obtain sample storage sequencing data, assigning the sequencing in the sample storage sequencing data, assigning the first-sequenced sample storage sequencing data as F1, assigning the second-sequenced sample storage sequencing data as F1-k1, assigning the third-sequenced sample storage sequencing data as F1-2k1, and assigning the g-sequenced sample storage sequencing data as F1- (g-1) k 1;
assigning ranks in the positive rank ordering data, assigning the rank of the first positive rank ordering data to F2, assigning the rank of the second positive rank ordering data to F2-k2, assigning the rank of the third positive rank ordering data to F2-2k2, and assigning the rank of the g-th positive rank ordering data to F2- (g-1) k 2;
extracting numerical values of sample name data corresponding to each positive orbit sample data in the positive sorting data and the sample storage sorting data, bringing the numerical values into a positive value calculation formula, and calculating a positive value ZZi;
and sorting according to the numerical value of the positive value factor corresponding to the sample name data, thereby obtaining positive value sorting data.
Further, the positive value calculation formula is specifically:
wherein ZZi is represented as a positive value, u2 is represented as a weight coefficient corresponding to the sample storage ordering data, u3 is represented as a weight coefficient corresponding to the positive sorting data, e1 is represented as an integrated conversion factor assigned to the sample storage ordering data and the positive sorting data, and u3 > u 2.
The invention has the beneficial effects that:
(1) the method comprises the steps of integrating data of a previous collected sample, performing correlation calculation on the integrated related data, calculating the related data corresponding to the related data, and performing data processing and calculation on a new collected target according to the related data, so that a new sample collecting method is determined, the data collecting accuracy is improved, the data analysis time is saved, and the working efficiency is improved;
(2) and (3) obtaining corresponding judging conditions through integration processing of new sample data, performing primary data classification judgment on related data according to the judging conditions, performing secondary division judgment according to the classification judgment result of the primary data, performing secondary division judgment according to the result of the secondary division judgment, thus realizing accurate classification of the sample data, and performing sequencing storage according to the classified data.
Drawings
The invention is further described below with reference to the accompanying drawings.
FIG. 1 is a system block diagram of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, the present invention is a sample collection and classification monitoring system for rail flaw detection, including a sample collection unit, a sample recording unit, a sample position unit, a sample management monitoring unit, a sample classification unit and a sample storage and display unit;
the sample set unit is used for acquiring sample base information related to the steel rail corresponding to the road section and transmitting the sample base information to the sample position unit;
the sample recording unit is used for storing related information of past sample collection and marking the related information of the past sample collection as sample recording information, wherein the sample recording information comprises sample recording data, sample loss data, sample good data and a sample loss image, the sample loss image refers to an image after the sample is damaged, the sample recording data refers to all steel rails of a certain road section collected in a record, the sample loss data refers to damaged steel rails corresponding to all steel rails in the certain road section in the record, and the sample good data refers to that the steel rails are not damaged;
the sample processing unit is used for acquiring the steel recording data, the sample loss data and the sample good data from the sample recording unit, and carrying out sample processing operation on the steel recording data, the sample loss data and the sample good data and the sample base information, wherein the specific operation process of the sample processing operation is as follows:
selecting a plurality of marking steel data, sequencing the section steel rails corresponding to the plurality of marking steel data in sequence, and marking the section steel rails corresponding to the plurality of marking steel data in sequence as sample base sequencing;
marking the sample base sequence with a digital mark Aa, wherein a is a positive integer, marking the sample base sequence according to sample loss data and sample good data, marking the sample loss data and the sample good data with Y-and Y + respectively, and marking the sample base sequence with the digital mark Aa according to the Y-and Y + so as to obtain a mark AaY-or AaY +;
selecting a plurality of sample loss data AaY-corresponding to the steel recording data, listing AaY-one by one to obtain a plurality of AaY-arrays, selecting a numerical value in each AaY-array, counting the total number of each group AaY-, summing the total numbers of a plurality of groups AaY-, calculating the sum of the total number of AaY-, dividing the sum of the AaY-total number by the number of the groups AaY-total number to obtain the average value of the total numbers of a plurality of groups AaY-, and calibrating the average value as the sample loss average value;
the total number of AaY-and the number of corresponding Aa are subjected to proportion calculation by a plurality of groups, the damage proportion is calculated, the average value of a plurality of damage proportions is calculated, the damage average value is calculated, and the damage average value is marked as SJ;
selecting the number of words with the same a data in a plurality of groups AaY-arrays, calibrating the number of words as a heavy sample loss value, selecting the minimum value of the a data in a plurality of groups AaY-arrays, calibrating the minimum value as a minimum sample loss value, selecting the sorting difference value between the key sample loss value and the minimum sample loss value in a plurality of groups, carrying out mean value calculation on the sorting difference value, calculating an interval mean value, and bringing the minimum sample loss value, the key sample loss value and the interval mean value into a sample loss relation calculation formula: heavy sample loss value = minimum sample loss value + mean interval value u1 e, where u1 is expressed as a fold rule value and e is expressed as a sample loss influence regulator;
extracting sample base information, marking the steel rail of each road section in the sample base information as steel sampling data, selecting the steel sampling data corresponding to all the road sections, and sequentially arranging the steel sampling data corresponding to all the road sections, wherein the specific steps are as follows: selecting one of the road section starting points, sequentially arranging and marking the road section starting points towards a set end point direction according to the road section starting points, and marking the corresponding steel acquisition data as Gi, wherein i =1,2,3, i.
The method comprises the following steps of identifying the total amount i of the corresponding steel rails of all road sections in a steel rail mark Gi, selecting the steel rails with i =1,2 and 3, marking the steel rails with i =1,2 and 3 as test samples, carrying out flaw detection on the test samples through a steel rail flaw detector, marking the result images detected by the steel rail flaw detector as test images, and matching the test images with the damage sample images, wherein the method specifically comprises the following steps:
when the data which are the same as the test image are matched from the damaged sample image, the damaged sample image is calibrated to be a damaged image to generate a damaged image signal, and when the data which are the same as the test image cannot be matched from the damaged sample image, the damaged sample image is calibrated to be a lossless image to generate a lossless signal;
extracting a damaged map signal and a lossless signal, identifying the damaged map signal and the lossless signal, judging damaged values when the corresponding i =1,2 and 3 are identified when the damaged map signal is identified, selecting the minimum value of the damaged values, calibrating the minimum value as a first detection steel rail, bringing the value of i corresponding to the first detection steel rail into a sample damage relation calculation formula, selecting the interval mean value as 2, calculating the next detection steel rail, and performing analogism calculation in sequence, thereby calculating the sequencing data of the detection steel rail;
when a lossless signal is identified, selecting a numerical value corresponding to the minimum sample loss value, carrying out numerical value replacement according to a sample loss relation calculation formula, converting the minimum sample loss value into a heavy sample loss value, wherein the calculation formula is as follows: the minimum sample loss value = a first detected steel rail + a spacing average value u1 × e, wherein the spacing average value is 1, the first detected steel rail is calculated, the minimum sample loss value is calibrated as a second detected steel rail, a sample sequence numerical value with the spacing average value of 2 is calculated according to a sample loss relation calculation formula, the sample sequence numerical value is calibrated as a third detected steel rail, and the like are repeated to calculate two detected steel rail sequencing data, wherein the two detected steel rail sequencing data and the one detected steel rail sequencing data are different in sequence numerical value of the first detected steel rail, and the second numerical value of the two detected steel rail sequencing data is the same as the first numerical value of the one detected steel rail sequencing data;
transmitting the damaged mean value and the damaged sample image to a sample grading unit through a sample monitoring unit;
transmitting the first detected steel rail sequencing data or the second detected steel rail sequencing data to a sample set unit;
the sample collecting unit collects data according to the first detected steel rail sequencing data or the second detected steel rail sequencing data, so that samples are collected on the steel rails of the corresponding road sections, the collected samples are marked as sample detailed information, the sample detailed information comprises sample image data, sample storage data, sample name data and sample division data, the sample image data refers to sample detection images, the sample storage data refers to the sizes of the sample detection images, the sample name data refers to the corresponding sample names of the sample detections, the sample division data refers to the time consumed by the sample detections, namely the playback time of the sample detections, and the sample image data, the sample storage data, the sample name data and the sample division data are transmitted to the sample division unit;
the sample grading unit is used for carrying out sample grading processing on the damaged mean value and the damaged sample image together with the sample image data, the sample storage data, the sample name data and the sample grading data, and the specific processing process of the sample grading processing is as follows:
selecting sample name data, extracting corresponding sample image data according to the sample name data, and matching the sample image data with a damaged sample image, wherein the sample name data specifically comprises the following steps: when the matching result of the sample image data and the damaged sample image is consistent, the sample image data is judged to be damaged data, a real damaged signal is generated, and when the matching result of the sample image data and the damaged sample image is inconsistent, the sample image data is judged to be lossless data, and a real non-signal is generated;
extracting real loss signals and real no signals, identifying the real loss signals and the real no signals, calibrating the real no signals into positive track sample data when the real no signals are identified, and calibrating the real no signals into loss track sample data when the real loss signals are identified;
extracting real loss signals and real no signals, and counting the number of the real loss signals and the number of the real no signals, so as to bring the number corresponding to the real loss signals and the real no signals into a calculation formula: the real loss rate = the number of real loss signals/(the number of real loss signals + the number of real no signals), the real loss rate is calculated, and the real loss rate is marked as Sv;
comparing the actual damage rate Sv with the damage mean value SJ, and specifically:
when Sv p1 is larger than or equal to SJ, judging that the damage rate is high, and generating a heavy loss signal;
when Sv p1 < SJ, judging that the damage rate is low, and generating a low-loss signal, wherein p1 is a deviation adjustment factor of the real loss rate, and the heavy-loss signal and the low-loss signal are marked as rate-loss signals;
extracting corresponding sample name data and sample division data according to the normal orbit sample data, and carrying out sample division judgment on the sample division data and the rate loss signal, wherein the sample division judgment specifically comprises the following steps: according to the positive orbit sample data, sorting the sample sorting data corresponding to the sample name data from large to small so as to obtain positive sorting data, identifying the rate loss signal, and when the identification result is the heavy loss signal, storing and sorting the positive sorting data and a plurality of corresponding sample storage data, specifically:
sorting a plurality of sample memory data from large to small to obtain sample memory sorting data, assigning the sorting in the sample memory sorting data, assigning the sorting data of the first sample memory to F1, assigning the sorting data of the second sample memory to F1-k1, assigning the sorting data of the third sample memory to F1-2k1, and assigning the sorting data of the g to F1- (g-1) k 1;
assigning ranks in the positive rank ordering data, assigning the rank of the first positive rank ordering data to F2, assigning the rank of the second positive rank ordering data to F2-k2, assigning the rank of the third positive rank ordering data to F2-2k2, and assigning the rank of the g-th positive rank ordering data to F2- (g-1) k 2;
extracting the numerical values of the sample name data corresponding to each positive orbit sample data in the positive sorting data and the sample storage sorting data, and bringing the numerical values into a positive value calculation formula:
wherein ZZi is represented as a positive value, u2 is represented as a weight coefficient corresponding to the sample storage ordering data, u3 is represented as a weight coefficient corresponding to the positive sorting data, e1 is represented as an integrated conversion factor assigned to the sample storage ordering data and the positive sorting data, and u3 is greater than u 2;
sorting according to the numerical value of the positive value factor corresponding to the sample name data to obtain positive value sorting data, and sorting and storing the sample name data and the corresponding sample image data according to the positive value sorting data;
when the identification result is a low-loss signal, setting a sample preset value M1, matching the sample preset value M1 with the forward sorting data, matching out a value corresponding to M1 in the forward sorting data, marking the value as a judgment value, marking the forward sorting data which is arranged in front of the judgment value in the forward sorting data as suspected data (including the matched value), marking the forward sorting data which is arranged behind the judgment value in the forward sorting data as normal data, and processing the suspected data and the normal data according to a processing mode of storage sorting processing, so as to obtain suspected storage sorting data corresponding to the suspected data and normal storage sorting data corresponding to the normal data;
extracting corresponding sample name data and sample division data according to the loss track sample data, and carrying out sample division judgment on the sample division data and the rate loss signal, wherein the sample division judgment specifically comprises the following steps: sorting sample sub data corresponding to the sample name data from large to small according to the loss track sample data to obtain loss track sorting data, identifying the rate loss signal, and analyzing the positive loss sorting data and the loss value sorting data according to the positive value sorting data when the identification result is the heavy loss signal;
when the identification result is a low-loss signal, setting a sample preset value M2, matching the sample preset value M2 with the normal-loss sorting data, matching a value corresponding to M2 in the normal-loss sorting data, marking the value as a judgment value, marking the normal-loss sorting data which is arranged in front of the judgment value in the normal-loss sorting data as loss normal data (including the matched value), marking the normal-loss sorting data which is arranged behind the judgment value in the normal-loss sorting data as loss abnormal data, and processing the loss normal data and the loss abnormal data according to a processing mode of storage sorting processing, so that loss normal storage sorting data corresponding to the loss normal data and loss abnormal storage sorting data corresponding to the loss abnormal data are obtained;
extracting positive value sorting data or suspected storage sorting data and positive similar storage sorting data and loss value sorting data or loss value sorting data and loss difference storage sorting data, wherein the positive value sorting data, the suspected storage sorting data and the positive similar storage sorting data only can be one, namely one condition of the positive value sorting data occurs, or the suspected storage sorting data and the positive similar storage sorting data occur, and the loss value sorting data, the loss value sorting data and the loss difference storage sorting data only can be one sort, namely one condition of the loss value sorting data occurs, or one condition of the positive storage sorting data and the loss difference storage sorting data occurs;
the occurrence condition of positive-value sorted data is marked as D1, the occurrence condition of suspected storage sorted data and positive-similar storage sorted data is marked as D2, the occurrence condition of damaged-value sorted data is marked as D3, the occurrence condition of damaged positive-stored sorted data and damaged different storage sorted data is marked as D4, the occurrence conditions are divided into four kinds, D1 and D3 occur simultaneously, D1 and D4 occur simultaneously, D2 and D3 occur simultaneously, or D2 and D4 occur simultaneously;
when D1 and D3 occur simultaneously, two classification modes occur, the corresponding sample name data and the sample image data are selected for sequencing storage according to the corresponding score sequencing;
when D2 and D3 appear simultaneously, three classification modes appear, corresponding sample name data and sample image data are selected for sequencing storage according to corresponding score sequencing;
when D1 and D4 occur simultaneously, three classification modes occur, corresponding sample name data and sample image data are selected for sequencing storage according to corresponding score sequencing;
when D2 and D4 occur simultaneously, four classification modes occur, corresponding sample name data and sample image data are selected for sorting and storage according to corresponding score sorting;
sorting the sample name data and the sample image data corresponding to the final corresponding sorting, calibrating the sample name data and the sample image data into sample storage sorting, and transmitting the sample storage sorting to a sample storage display unit;
the sample storage and display unit is used for storing and displaying sample storage and sequencing, and is specifically a tablet computer.
When the device works, the sample set unit is used for collecting sample base information related to a steel rail corresponding to a road section, and the sample base information is transmitted to the sample position unit; the sample marking unit stores related information of past sample collection, marks the related information of the past sample collection as sample marking information, the sample position unit acquires marking steel data, sample damage data and sample good data from the sample marking unit, carrying out sample position operation on the recorded steel data, the sample loss data, the sample good data and the sample base information to obtain a damage mean value, a sample loss image, a first detected steel rail sequencing data or a second detected steel rail sequencing data, transmitting the first detected steel rail sequencing data or the second detected steel rail sequencing data to a sample collection unit, transmitting the damage mean value and the sample loss image to a sample grade unit through a sample management monitoring unit, carrying out data acquisition by the sample collection unit according to the first detected steel rail sequencing data or the second detected steel rail sequencing data, thus, the steel rail of the corresponding road section is sampled, the sampled sample is marked as sample detail information, and the sample detail information is transmitted to a sample grading unit; the sample grading unit performs sample grading processing on the damaged mean value and the damaged sample image together with the sample image data, the sample storage data, the sample name data and the sample grading data to obtain sample storage sequencing, and transmits the sample storage sequencing to the sample storage and display unit; and the sample storage and display unit stores and displays the sample storage and sorting.
The foregoing is merely exemplary and illustrative of the present invention and various modifications, additions and substitutions may be made by those skilled in the art to the specific embodiments described without departing from the scope of the invention as defined in the following claims.
Claims (1)
1. A sample collection and classification supervision system for steel rail flaw detection is characterized by comprising a sample collection unit, a sample recording unit, a sample position unit, a sample management monitoring unit, a sample grading unit and a sample storage and display unit;
the sample set unit is used for acquiring sample base information related to the steel rail corresponding to the road section and transmitting the sample base information to the sample position unit;
the sample recording unit is used for storing sample recording information related to previous sample collection, and the sample recording information comprises steel recording data, sample loss data, sample good data and a sample loss image;
the sample processing unit is used for carrying out sample processing operation on the recorded steel data, the sample loss data, the sample good data and the sample base information to obtain a damaged average value, a sample loss image, a detected steel rail sequencing data or two detected steel rail sequencing data, transmitting the detected steel rail sequencing data or the two detected steel rail sequencing data to the sample set unit, transmitting the damaged average value and the sample loss image to the sample management monitoring unit, wherein the specific operation process of the sample processing operation is as follows:
selecting a plurality of marking steel data, sequencing the road section steel rails corresponding to the plurality of marking steel data in sequence and marking the road section steel rails as sample base sequencing, carrying out sample base processing according to the sample base sequencing, processing to obtain a damage mean value SJ and a multiple rule value u1, and carrying out the specific processing process of the sample base processing according to the sample base sequencing:
marking the sample base sequence with a digital mark Aa, wherein a is a positive integer, marking the sample base sequence according to sample loss data and sample good data, marking the sample loss data and the sample good data with Y-and Y + respectively, and marking the sample base sequence with the digital mark Aa according to the Y-and Y + so as to obtain a mark AaY-or AaY +;
selecting a plurality of sample loss data AaY-corresponding to the steel recording data, listing AaY-one by one to obtain a plurality of AaY-arrays, selecting a numerical value in each AaY-array, counting the total number of each group AaY-, summing the total numbers of a plurality of groups AaY-, calculating the sum of the total number of AaY-, dividing the sum of the AaY-total number by the number of the groups AaY-total number to obtain the average value of the total numbers of a plurality of groups AaY-, and calibrating the average value as the sample loss average value;
the total number of AaY-and the number of corresponding Aa are subjected to proportion calculation by a plurality of groups, the damage proportion is calculated, the average value of a plurality of damage proportions is calculated, the damage average value is calculated, and the damage average value is marked as SJ;
selecting the same number of words of a data appearing in a plurality of groups of AaY-arrays, calibrating the same number of words as a heavy sample loss value, selecting the minimum value of the a data appearing in a plurality of groups of AaY-arrays, calibrating the minimum value of the a data as a minimum sample loss value, selecting a sequencing difference value of a key sample loss value and the minimum sample loss value in the plurality of groups, carrying out mean value calculation on the sequencing difference value, calculating an interval mean value, and bringing the minimum sample loss value, the key sample loss value and the interval mean value into a sample loss relation calculation formula: heavy sample loss value = minimum sample loss value + mean interval value u1 e, where u1 is expressed as a fold rule value and e is expressed as a sample loss influence regulator;
extracting sample base information, marking the steel rail of each road section in the sample base information as steel sampling data, selecting the steel sampling data corresponding to all the road sections, and sequentially arranging the steel sampling data corresponding to all the road sections, wherein the specific steps are as follows: selecting one of the road section starting points, sequentially arranging and marking the road section starting points towards a set end point direction according to the road section starting points, and marking the corresponding steel acquisition data as Gi, wherein i =1,2,3, i.
The method comprises the following steps of identifying the total amount i of steel rails corresponding to all road sections in a steel rail mark Gi, selecting steel rails with i =1,2 and 3, marking the steel rails with i =1,2 and 3 as test samples, carrying out flaw detection on the test samples through a steel rail flaw detector, marking a result image detected by the steel rail flaw detector as a test image, matching the test image with a damage sample image to obtain first detection steel rail sequencing data and second detection steel rail sequencing data, and carrying out matching on the test image and the damage sample image in the specific process that:
when the data which are the same as the test image are matched from the damaged sample image, the damaged sample image is calibrated to be a damaged image to generate a damaged image signal, and when the data which are the same as the test image cannot be matched from the damaged sample image, the damaged sample image is calibrated to be a lossless image to generate a lossless signal;
extracting a damage map signal and a lossless signal, identifying the damage map signal and the lossless signal, judging damage values when the damage map signal is identified, selecting the minimum value of the damage values when the corresponding i =1,2 and 3, calibrating the minimum value as a first detection steel rail, bringing the value of i corresponding to the first detection steel rail into a sample damage relation calculation formula, selecting the interval mean value as 2, calculating the next detection steel rail, and performing analogism calculation in sequence to calculate the sequencing data of the detection steel rail;
when a lossless signal is identified, selecting a numerical value corresponding to the minimum sample loss value, carrying out numerical value replacement according to a sample loss relation calculation formula, converting the minimum sample loss value into a heavy sample loss value, wherein the calculation formula is as follows: the minimum sample loss value = the first detected steel rail + the interval mean value u1 e, the interval mean value is 1, the first detected steel rail is calculated, the minimum sample loss value is calibrated to be the second detected steel rail, the sample sequence numerical value with the interval mean value of 2 is calculated according to the sample loss relation calculation formula and is calibrated to be the third detected steel rail, and the sequencing data of the second detected steel rail are calculated by analogy;
the sample collection unit carries out data acquisition according to the first detection steel rail sequencing data or the second detection steel rail sequencing data, so as to carry out sample acquisition on the steel rails of the corresponding road section, mark the acquired samples as sample detailed information, and transmit the sample detailed information to the sample management monitoring unit, wherein the sample detailed information comprises sample image data, sample storage data, sample name data and sample division data;
the sample management monitoring unit is used for carrying out sample grade processing on the damaged mean value and the damaged sample image together with the sample image data, the sample storage data, the sample name data and the sample grade data to obtain sample storage sequencing, and transmitting the sample storage sequencing to the sample storage and display unit, wherein the specific processing process of the sample grade processing is as follows:
selecting sample name data, extracting corresponding sample image data according to the sample name data, matching the sample image data with a loss sample image, and generating a real loss signal and a real no-signal according to the consistency or inconsistency of the matching results;
identifying real loss signals and real no signals, calibrating sample image data into positive track sample data when the real no signals are identified, and calibrating the sample image data into loss track sample data when the real loss signals are identified;
calculating the real loss rate according to the occurrence number of the real loss signals and the real no-signals, comparing the real loss rate with the damage mean value SJ to obtain heavy loss signals and low loss signals, and marking the heavy loss signals and the low loss signals as rate loss signals;
extracting corresponding sample name data and sample division data according to the normal orbit sample data, and carrying out sample division judgment on the sample division data and the rate loss signal, wherein the sample division judgment specifically comprises the following steps: sorting sample data corresponding to sample name data from large to small according to the positive orbit sample data to obtain positive sorting data, identifying a rate loss signal, storing and sorting the positive sorting data and a plurality of corresponding sample storage data to obtain positive sorting data when an identification result is a heavy loss signal, sorting and storing the sample name data and the corresponding sample image data according to the positive sorting data, wherein the specific processing process of the storage and sorting processing is as follows:
sorting a plurality of sample memory data from large to small to obtain sample memory sorting data, assigning the sorting in the sample memory sorting data, assigning the sorting data of the first sample memory to F1, assigning the sorting data of the second sample memory to F1-k1, assigning the sorting data of the third sample memory to F1-2k1, and assigning the sorting data of the g to F1- (g-1) k 1;
assigning ranks in the positive rank ordering data, assigning the rank of the first positive rank ordering data to F2, assigning the rank of the second positive rank ordering data to F2-k2, assigning the rank of the third positive rank ordering data to F2-2k2, and assigning the rank of the g-th positive rank ordering data to F2- (g-1) k 2;
extracting the numerical values of the sample name data corresponding to each positive orbit sample data in the positive sorting data and the sample storage sorting data, and bringing the numerical values into a positive value calculation formula:
calculating a positive value ZZi, wherein ZZi represents a positive value, u2 represents a weight coefficient corresponding to the sample sorting data, u3 represents a weight coefficient corresponding to the positive sorting data, e1 represents an integrated conversion factor assigned to the sample sorting data and the positive sorting data, and u3 > u 2;
sorting according to the numerical value of the positive value factor corresponding to the sample name data so as to obtain positive value sorting data;
when the identification result is a low-loss signal, setting a sample preset value M1, matching the sample preset value M1 with the forward sorting data, matching a value corresponding to M1 in the forward sorting data, marking the value as a judgment value, marking the forward sorting data which is arranged in front of the judgment value in the forward sorting data as suspected data, marking the forward sorting data which is arranged behind the judgment value in the forward sorting data as positive data, and processing the suspected data and the positive data according to a processing mode of storage sorting processing, so as to obtain suspected storage sorting data corresponding to the suspected data and positive storage sorting data corresponding to the positive data;
extracting corresponding sample name data and sample division data according to the loss track sample data, and carrying out sample division judgment on the sample division data and the rate loss signal, wherein the sample division judgment specifically comprises the following steps: sorting sample sub data corresponding to the sample name data from large to small according to the loss track sample data to obtain loss track sorting data, identifying the rate loss signal, and analyzing the positive loss sorting data and the loss value sorting data according to the positive value sorting data when the identification result is the heavy loss signal;
when the identification result is a low-loss signal, setting a sample preset value M2, matching the sample preset value M2 with the positive loss sorting data, matching a value corresponding to M2 in the positive loss sorting data, marking the value as a judgment value, marking the positive sorting data which is arranged in front of the judgment value in the positive sorting data as loss positive data, marking the positive sorting data which is arranged behind the judgment value in the positive sorting data as loss differential data, and processing the loss positive data and the loss differential data according to a processing mode of storage sorting processing, so that loss positive storage sorting data corresponding to the loss positive data and loss differential data corresponding to the loss differential data are obtained;
extracting positive value sorting data or suspected storage sorting data and positive similar storage sorting data and loss value sorting data or loss value sorting data and loss difference storage sorting data, wherein the positive value sorting data, the suspected storage sorting data and the positive similar storage sorting data only can be one, namely one condition of the positive value sorting data occurs, or the suspected storage sorting data and the positive similar storage sorting data occur, and the loss value sorting data, the loss value sorting data and the loss difference storage sorting data only can be one sort, namely one condition of the loss value sorting data occurs, or one condition of the positive storage sorting data and the loss difference storage sorting data occurs;
the occurrence condition of positive-value sorted data is marked as D1, the occurrence condition of suspected storage sorted data and positive-similarity storage sorted data is marked as D2, the occurrence condition of loss-value sorted data is marked as D3, the occurrence condition of loss-positive-value sorted data and loss-difference storage sorted data is marked as D4, and the occurrence conditions are divided into four kinds, namely D1 and D3 occur simultaneously, or D1 and D4 occur simultaneously, or D2 and D3 occur simultaneously, or D2 and D4 occur simultaneously;
when D1 and D3 occur simultaneously, two classification modes occur, the corresponding sample name data and the sample image data are selected for sequencing storage according to the corresponding score sequencing;
when D2 and D3 appear simultaneously, three classification modes appear, corresponding sample name data and sample image data are selected for sequencing storage according to corresponding score sequencing;
when D1 and D4 appear simultaneously, three classification modes appear, corresponding sample name data and sample image data are selected for sequencing storage according to corresponding score sequencing;
when D2 and D4 occur simultaneously, four classification modes occur, corresponding sample name data and sample image data are selected for sorting and storage according to corresponding score sorting;
sorting the sample name data and the sample image data corresponding to the final corresponding sorting, calibrating the sample name data and the sample image data into sample storage sorting, and transmitting the sample storage sorting to a sample storage and display unit;
the sample loss image refers to an image of a damaged sample, the steel recording data refers to all steel rails collected in a certain section in a record, the sample loss data refers to damaged steel rails corresponding to all steel rails in a certain section in the record, and the sample good data refers to no damage to the steel rails;
the sample image data refers to a sample detection image, the sample storage data refers to the size of the sample detection image, the sample name data refers to a corresponding sample name of the sample detection, and the sample division data refers to the time consumed by the sample detection, namely the playback time during the sample detection.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111161213.9A CN113791034B (en) | 2021-09-30 | 2021-09-30 | Sample collection and classification supervision system for steel rail flaw detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111161213.9A CN113791034B (en) | 2021-09-30 | 2021-09-30 | Sample collection and classification supervision system for steel rail flaw detection |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113791034A CN113791034A (en) | 2021-12-14 |
| CN113791034B true CN113791034B (en) | 2022-09-06 |
Family
ID=78877647
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111161213.9A Active CN113791034B (en) | 2021-09-30 | 2021-09-30 | Sample collection and classification supervision system for steel rail flaw detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113791034B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104634872A (en) * | 2015-01-10 | 2015-05-20 | 哈尔滨工业大学（威海） | Online high-speed railway steel rail damage monitoring method |
| WO2018040117A1 (en) * | 2016-08-30 | 2018-03-08 | 广东汕头超声电子股份有限公司 | Method and system for ultrasonic imaging detection of welding seam of dual-array probe-based steel-rail |
| WO2019201177A1 (en) * | 2018-04-17 | 2019-10-24 | 江苏必得科技股份有限公司 | Train component crack damage monitoring method and system |
| CN110378869A (en) * | 2019-06-05 | 2019-10-25 | 北京交通大学 | A kind of rail fastening method for detecting abnormality of sample automatic marking |
| CN111855810A (en) * | 2020-07-20 | 2020-10-30 | 济南大学 | Rail foot damage identification method and system based on recurrent neural network |
| CN111896625A (en) * | 2020-08-17 | 2020-11-06 | 中南大学 | Real-time monitoring method and monitoring system for rail damage |
| CN113030123A (en) * | 2021-05-27 | 2021-06-25 | 南昌华梦达航空科技发展有限公司 | AOI detection feedback system based on Internet of things |
| CN113254572A (en) * | 2021-07-06 | 2021-08-13 | 深圳市知酷信息技术有限公司 | Electronic document classification supervision system based on cloud platform |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106650780B (en) * | 2016-10-18 | 2021-02-12 | 腾讯科技（深圳）有限公司 | Data processing method and device, classifier training method and system |
-
2021
- 2021-09-30 CN CN202111161213.9A patent/CN113791034B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104634872A (en) * | 2015-01-10 | 2015-05-20 | 哈尔滨工业大学（威海） | Online high-speed railway steel rail damage monitoring method |
| WO2018040117A1 (en) * | 2016-08-30 | 2018-03-08 | 广东汕头超声电子股份有限公司 | Method and system for ultrasonic imaging detection of welding seam of dual-array probe-based steel-rail |
| WO2019201177A1 (en) * | 2018-04-17 | 2019-10-24 | 江苏必得科技股份有限公司 | Train component crack damage monitoring method and system |
| CN110378869A (en) * | 2019-06-05 | 2019-10-25 | 北京交通大学 | A kind of rail fastening method for detecting abnormality of sample automatic marking |
| CN111855810A (en) * | 2020-07-20 | 2020-10-30 | 济南大学 | Rail foot damage identification method and system based on recurrent neural network |
| CN111896625A (en) * | 2020-08-17 | 2020-11-06 | 中南大学 | Real-time monitoring method and monitoring system for rail damage |
| CN113030123A (en) * | 2021-05-27 | 2021-06-25 | 南昌华梦达航空科技发展有限公司 | AOI detection feedback system based on Internet of things |
| CN113254572A (en) * | 2021-07-06 | 2021-08-13 | 深圳市知酷信息技术有限公司 | Electronic document classification supervision system based on cloud platform |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113791034A (en) | 2021-12-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103150900B (en) | Traffic jam event automatic detecting method based on videos | |
| CN102436624B (en) | A kind of based on the power marketing live traffic treatment scheme of handheld terminal and the method for specification | |
| CN107958031B (en) | Resident travel OD distribution extraction method based on fusion data | |
| CN105046959B (en) | Urban Travel Time extracting method based on Dual-window shiding matching mechanism | |
| CN103631681A (en) | Method for online restoring abnormal data of wind power plant | |
| CN112153574B (en) | Method and system for checking accuracy of roadside device clock based on floating vehicle | |
| Abou Chacra et al. | Fully automated road defect detection using street view images | |
| CN102663252A (en) | Combined type pavement usability performance evaluation method for underground road | |
| CN108270636A (en) | Link-quality-evaluating method and device | |
| CN110567662B (en) | Short-term bridge monitoring and evaluating method based on engineering simulation | |
| CN116308958A (en) | Carbon emission online detection and early warning system and method based on mobile terminal | |
| CN113657747B (en) | Intelligent assessment system for enterprise safety production standardization level | |
| CN113640380A (en) | Multi-stage classification method and system for rail damage detection | |
| CN116466241A (en) | Thermal runaway positioning method for single battery | |
| CN113791034B (en) | Sample collection and classification supervision system for steel rail flaw detection | |
| CN112989660B (en) | Method for predicting corrosion of pipeline under subway stray current based on partial least square method | |
| CN115376315A (en) | Road network emission accounting-oriented multi-level bayonet quality control method | |
| CN114167837B (en) | Intelligent fault diagnosis method and system for railway signal system | |
| CN114461476B (en) | Memory bank fault detection method, device and system | |
| CN112732773B (en) | Method and system for checking uniqueness of relay protection defect data | |
| CN111831635B (en) | Method and device for accurately extracting meteorological data | |
| CN115081916A (en) | DNA digital management system | |
| CN113516407B (en) | Snow disaster distribution identification method and system for regions along high-speed rail | |
| CN117315943B (en) | Monitoring analysis and early warning method and system for overrun transportation violations | |
| CN110852516A (en) | Data quality judging method based on big data information entropy traffic flow detection equipment |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20231024 Address after: Floor 3, Building 3, Guochuang Science and Technology Park, No. 10 Tianzhi Road, High tech Zone, Hefei City, Anhui Province, 230000 Patentee after: Hefei Detai Nondestructive Testing Technology Co.,Ltd. Address before: 230051 b1-1, xiezhihui Industrial Park, Zhongguancun, Baohe District, Hefei City, Anhui Province Patentee before: Hefei detai rail transit Data Co.,Ltd. |