CN116652396A - Safety early warning method and system for laser inner carving machine - Google Patents
Safety early warning method and system for laser inner carving machine Download PDFInfo
- Publication number
- CN116652396A CN116652396A CN202310959497.9A CN202310959497A CN116652396A CN 116652396 A CN116652396 A CN 116652396A CN 202310959497 A CN202310959497 A CN 202310959497A CN 116652396 A CN116652396 A CN 116652396A
- Authority
- CN
- China
- Prior art keywords
- target object
- key point
- sequence
- item
- index
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 34
- 230000001932 seasonal effect Effects 0.000 claims description 62
- 230000005856 abnormality Effects 0.000 claims description 48
- 238000013459 approach Methods 0.000 claims description 40
- 238000010147 laser engraving Methods 0.000 claims description 29
- 230000000737 periodic effect Effects 0.000 claims description 21
- 230000008859 change Effects 0.000 claims description 17
- 238000004364 calculation method Methods 0.000 claims description 14
- 238000004590 computer program Methods 0.000 claims description 6
- 230000006870 function Effects 0.000 claims description 5
- 238000010606 normalization Methods 0.000 claims description 4
- 206010000117 Abnormal behaviour Diseases 0.000 abstract description 15
- 210000001508 eye Anatomy 0.000 description 66
- 230000006399 behavior Effects 0.000 description 14
- 230000002159 abnormal effect Effects 0.000 description 9
- 230000008569 process Effects 0.000 description 6
- 238000001514 detection method Methods 0.000 description 5
- 230000014759 maintenance of location Effects 0.000 description 4
- 230000009467 reduction Effects 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 3
- 210000003423 ankle Anatomy 0.000 description 2
- 238000000354 decomposition reaction Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 210000002683 foot Anatomy 0.000 description 2
- 210000003128 head Anatomy 0.000 description 2
- 210000000707 wrist Anatomy 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000001217 buttock Anatomy 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 210000005069 ears Anatomy 0.000 description 1
- 210000001513 elbow Anatomy 0.000 description 1
- 230000004438 eyesight Effects 0.000 description 1
- 210000003127 knee Anatomy 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 210000001331 nose Anatomy 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 239000000523 sample Substances 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 210000002832 shoulder Anatomy 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 231100000331 toxic Toxicity 0.000 description 1
- 230000002588 toxic effect Effects 0.000 description 1
- 239000012780 transparent material Substances 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B23—MACHINE TOOLS; METAL-WORKING NOT OTHERWISE PROVIDED FOR
- B23K—SOLDERING OR UNSOLDERING; WELDING; CLADDING OR PLATING BY SOLDERING OR WELDING; CUTTING BY APPLYING HEAT LOCALLY, e.g. FLAME CUTTING; WORKING BY LASER BEAM
- B23K26/00—Working by laser beam, e.g. welding, cutting or boring
- B23K26/36—Removing material
- B23K26/362—Laser etching
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B23—MACHINE TOOLS; METAL-WORKING NOT OTHERWISE PROVIDED FOR
- B23K—SOLDERING OR UNSOLDERING; WELDING; CLADDING OR PLATING BY SOLDERING OR WELDING; CUTTING BY APPLYING HEAT LOCALLY, e.g. FLAME CUTTING; WORKING BY LASER BEAM
- B23K26/00—Working by laser beam, e.g. welding, cutting or boring
- B23K26/70—Auxiliary operations or equipment
- B23K26/702—Auxiliary equipment
-
- 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
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Optics & Photonics (AREA)
- Plasma & Fusion (AREA)
- Mechanical Engineering (AREA)
- Laser Beam Processing (AREA)
- Lasers (AREA)
Abstract
The invention relates to the technical field of laser internal engraving machines, in particular to a safety early warning method and a system for a laser internal engraving machine. According to the invention, the related time sequence of each target object around the laser inner carving machine is obtained, so that abnormal behaviors of observers of the laser inner carving machine can be reliably detected and early-warned, and the reliability of the safety early warning of the laser inner carving machine is improved.
Description
Technical Field
The invention relates to the technical field of laser internal engraving machines, in particular to a safety early warning method and system for a laser internal engraving machine.
Background
The laser internal engraving machine is the most advanced and popular novel processing equipment for engraving transparent materials internationally at present, and is widely applied to various industries. Although any toxic or harmful gas is not generated in the operation process of the laser internal engraving machine, if the body parts of surrounding staff or other staff are in the dangerous range, the laser generated by the laser internal engraving machine can hurt the skin, and if the laser is watched by eyes, the laser can hurt the eyes, so that the timely detection and early warning of dangerous behaviors around the laser internal engraving machine are important.
In the prior art, a ranging beacon and a ranging base station are used for ranging, namely, an early warning distance is set, and when the distance between the ranging beacon and the ranging base station is smaller than the early warning distance, an audible and visual alarm is triggered to give an alarm. When the ranging method is applied to detection and early warning of surrounding dangerous behaviors of the laser engraving machine, the following defects exist: firstly, alarming is carried out only when the surrounding staff or other staff has parts to reach the dangerous range, and the human body is injured at the moment, so that early warning has hysteresis and no predictability; secondly, even if the surrounding staff or other staff does not reach the dangerous range, eyes can be burnt when the staff or other staff looks at the laser for a long time, and the possible human eye burning situation cannot be early warned. Therefore, when the existing ranging method is applied to detection and early warning of surrounding dangerous behaviors of the laser inner carving machine, the problem that the detection and early warning reliability is poor exists, the effective early warning cannot be carried out, and the problem of safety accidents possibly generated in the working process of the laser inner carving machine cannot be completely avoided.
Disclosure of Invention
The invention aims to provide a safety early warning method and a system for a laser inner carving machine, which are used for solving the problem of poor safety early warning reliability of the existing laser inner carving machine.
In order to solve the technical problems, the invention provides a safety pre-warning method for a laser engraving machine, which comprises the following steps:
acquiring video images around the laser engraving machine, and acquiring a distance sequence of each key point of each target object and a position sequence of the eye key point of each target object according to the video images;
respectively decomposing the distance sequence of each key point of each target object, and determining the trend item sequence, the season item sequence and the residual item sequence of each key point of each target object;
determining each close key point in each key point of each target object and an accuracy index, a close index value and a close speed value of the trend item sequence of each close key point according to the trend item sequence of each key point of each target object;
according to the seasonal item sequences of the key points of each target object, determining the periodic index of the seasonal item sequences of the key points of each target object, and according to the residual item sequences of the key points of each target object, determining the pure randomness index of the residual item sequences of the key points of each target object;
Determining a key degree factor corresponding to each near key point of each target object according to the accuracy index, the periodicity index and the pure randomness index corresponding to each near key point of each target object;
determining a stay index of the eye key points of each target object according to the position sequence of the eye key points of each target object;
and determining the abnormality degree of each target object according to the key degree factor, the approach index value, the approach speed value and the stay index of the eye key point corresponding to each approach key point of each target object, and further determining whether to perform early warning.
Further, determining an accuracy index, a closeness index value, and a closeness rate value of each close to a key point and a trend item sequence of each close to the key point in each key point of each target object includes:
clustering the trend item sequences of the key points of each target object to obtain trend item clusters, determining trend item clusters close to the internal engraving machine in the trend item clusters, and determining the key points corresponding to the trend item sequences in the trend item clusters close to the internal engraving machine as the close key points of each target object;
Calculating a distance value of a trend item sequence close to a key point in each trend item cluster of the approaching internal engraving machine of each target object and a trend item sequence of a cluster center when the trend item cluster of the approaching internal engraving machine is acquired, and taking a result of normalizing the opposite number of the distance value as an accuracy index of the trend item sequence corresponding to the approaching key point, thereby obtaining an accuracy index of each trend item sequence close to the key point of each target object;
determining the difference value between the next trend item and the previous trend item in the adjacent two trend item in each trend item sequence close to the key point of each target object, and determining the total number of all the difference values smaller than 0 as a close index value of the trend item sequence close to the key point, thereby obtaining the close index value of each trend item sequence close to the key point of each target object;
and determining the ratio of the absolute value of the sum of all differences which are smaller than 0 and correspond to each trend item sequence which is close to the key point of each target object to the total time which corresponds to all differences which are smaller than 0, and determining the ratio as the close speed value of the trend item sequence which corresponds to the key point, thereby obtaining the close speed value of each trend item sequence which is close to the key point of each target object.
Further, determining a trend item cluster close to the engraving machine in each trend item cluster comprises the following steps:
and determining the total number of all differences which are smaller than 0 and correspond to each trend item sequence in each trend item cluster, calculating the total sum of the total numbers which correspond to all trend item sequences in each trend item cluster, and determining the trend item cluster corresponding to the largest sum as the trend item cluster close to the engraving machine.
Further, determining a periodicity index of the sequence of seasonal items near the keypoint for each target object includes:
clustering the seasonal item sequences of the key points of each target object to obtain seasonal item clusters;
dividing any one seasonal item sequence in each seasonal item cluster into a plurality of subsequences, calculating the absolute value of a pearson correlation coefficient between any two subsequences, setting the periodicity index of each seasonal item sequence in the seasonal item cluster to be 1 if the absolute value of the pearson correlation coefficient is larger than a set coefficient threshold value, otherwise setting the periodicity index of each seasonal item sequence in the seasonal item cluster to be 0, and thus obtaining the periodicity index of each seasonal item sequence, close to a key point, of each target object.
Further, determining a pure randomness index of a residual term sequence of each target object near the key point comprises:
and carrying out autocorrelation calculation on the residual error item sequences of each target object, which are close to the key points, so as to determine the pure randomness index of the residual error item sequences of each target object, which are close to the key points.
Further, determining a key degree factor corresponding to each near key point of each target object, wherein the corresponding calculation formula is as follows:
wherein ,for the first object of each target objectiA key degree factor corresponding to the key point, and +.>For the first object of each target objectiPure randomness index of residual sequence near key point,/for the residual sequence near key point>For the first object of each target objectiAn accuracy index of the trend sequence close to the key point,/->For the first object of each target objectiNear the key pointPeriodic index of the seasonal sequence of items, +.>To set the weight coefficient.
Further, determining a retention indicator for the eye keypoints for each target subject includes:
the eye key points of each target object comprise a left eye key point and a right eye key point, the position change tracks of the left eye key point and the right eye key point of each target object are determined according to the position sequences of the left eye key point and the right eye key point of each target object, and the stay indexes of the left eye key point and the right eye key point are determined according to the coordinates of each point in the position change tracks.
Further, determining the degree of abnormality of each target object, wherein the corresponding calculation formula is as follows:
wherein ,for the degree of abnormality of each target object +.>Stay index for right eye key point of each target object, +.>Stay index for left eye key point of each target object, +.>For the first object of each target objectiA key degree factor corresponding to the key point, and +.>For the first object of each target objectiApproach speed value of approach key, +.>For each purposeObject of targetiA near index value near the key point,Nfor the total number of near key points of each target object, +.>Is a normalization function.
Further, determining whether to perform early warning includes:
determining the maximum abnormality degree in the abnormality degrees of all the target objects, and if the maximum abnormality degree is smaller than a first abnormality degree threshold value, not performing early warning; if the maximum abnormality degree is not less than the first abnormality degree threshold value and is less than the second abnormality degree threshold value, performing first-level early warning; and if the maximum abnormality degree is not less than the second abnormality degree threshold, performing second level early warning.
In order to solve the technical problem, the invention also provides a safety early warning system for the laser internal engraving machine, which comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor, wherein the processor realizes the steps of the safety early warning method for the laser internal engraving machine when executing the computer program.
The invention has the following beneficial effects: the distance sequence of each key point of each target object around the laser internal carving machine and the position sequence of the eye key point of each target object are obtained, and the sequences are analyzed, so that abnormal behaviors of a surrounding person can be reliably detected and early-warned, and the reliability of the safety early warning of the laser internal carving machine is improved. Specifically, the distance sequence of each key point of each target object is decomposed, so that a trend item sequence, a season item sequence and a residual item sequence are obtained, and the trend behavior of the target object is detected in advance based on the three sequences. And analyzing the trend item sequence of each key point of each target object, screening out key points close to the laser engraving machine, namely key points close to the laser engraving machine, and analyzing the trend item sequence close to the key points, so as to obtain accuracy indexes, close index values and close speed values corresponding to each close key point. The accuracy index represents the accuracy condition that the approaching key point is the key point which really tends to the laser inner engraving machine, the approaching index value represents the frequency condition that the approaching key point approaches the laser inner engraving machine, and the approaching speed value represents the speed condition that the approaching key point approaches the laser inner engraving machine. And analyzing the seasonal term sequence and the residual term sequence of each key point of each target object so as to obtain a periodic index and a pure randomness index corresponding to each close key point, wherein the periodic index represents the condition that periodic actions exist near the key points, and the pure randomness index represents the accuracy index and the reliability condition of the periodic index determined by the trend term sequence and the seasonal term sequence. And determining a key degree factor corresponding to each close key point of each target object based on the accuracy index, the periodicity index and the pure randomness index corresponding to each close key point of each target object, wherein the key degree factor characterizes the importance degree of the close key point in all the close key points. On the one hand, according to the key degree factor, the approach index value and the approach speed value corresponding to the approach key point, the trend situation of the body of the target object can be measured, so that the abnormal behavior of the target object approaching the laser engraving machine is detected in advance; on the other hand, according to the position sequence of the eye key points of the target object, determining the stay index of the eye key points, wherein the stay index represents the sight stay condition of the eye key points. And finally determining the abnormality degree of each target object by combining the trend degree of the body of the target object and the stay condition of the sight line, so as to determine whether to perform early warning. The method can analyze and identify the trend behavior of the observer in advance, and can detect the sight stay condition of the observer, so that the abnormal dangerous behavior of the observer can be accurately determined and early warned in time, the safety accident problem possibly generated in the working process of the laser internal engraving machine can be avoided as much as possible, and the problem of poor safety early warning reliability of the existing laser internal engraving machine is effectively solved.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a security early warning method for a laser engraving machine according to an embodiment of the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the present invention to achieve the preset purpose, the following detailed description is given below of the specific implementation, structure, features and effects of the technical solution according to the present invention with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
Aiming at the problem of poor safety precaution reliability of the existing laser inner carving machine, the embodiment provides a safety precaution method for the laser inner carving machine, which is used for acquiring time sequences of a plurality of key points of a human body, carrying out corresponding analysis based on the time sequences, and detecting abnormal behaviors of the human body, which are close to the laser inner carving machine and stay on the laser inner carving machine for a long time, so that the early warning is completed more accurately and reliably, and the safety precaution reliability of the laser inner carving machine is effectively improved.
Specifically, a flow chart corresponding to the safety precaution method for the laser engraving machine is shown in fig. 1, and the method comprises the following steps:
step S1: and acquiring video images around the laser engraving machine, and acquiring a distance sequence of each key point of each target object and a position sequence of the eye key point of each target object according to the video images.
In order to analyze abnormal behaviors of a surrounding person of the laser inner carving machine in a teaching scene or in a working process of the laser inner carving machine, dangerous behaviors are timely early-warned, and RGB-D cameras are placed around the laser inner carving machine to collect a close video image of the surrounding person. Based on the RGB image in the collected video image, 17 key points of people around the laser engraving machine in the video image are marked by using a human body key point detection model OpenPose, wherein the key points comprise a nose, eyes, ears, shoulders, elbows, wrists, buttocks, knees and ankles which are respectively positioned on the left side and the right side, and each target object, namely each key point of a person is tracked. Based on the obtained positions of the key points of each target object in the corresponding RGB image frames in the video image, obtaining depth values of the same positions in the corresponding D image frames in the video image, wherein the depth values are also called as distances, so that the distances of the key points of each target object in the corresponding image frames in the video image are obtained.
After the positions and the distances of the key points of each target object in the corresponding image frames in the video image are obtained, the distances of the key points are arranged according to a time sequence, so that a distance sequence of the key points of each target object can be obtained, and the abnormal behaviors of the target object, which are close to the laser engraving machine, can be analyzed and early warned based on the distance sequences of the key points. Meanwhile, according to the positions of the left eye and the right eye of each target object in the corresponding image frames in the video image, the positions of each eye are arranged according to time sequence, so that a position sequence of each eye can be obtained, and the stay condition of the human eye sight on the laser inner carving machine can be detected based on the position sequences.
Step S2: and decomposing the distance sequence of each key point of each target object respectively, and determining the trend item sequence, the season item sequence and the residual item sequence of each key point of each target object.
After the distance sequence of 17 key points of each target object is obtained in the step S1, it is inconvenient to directly analyze because the distance sequence of each key point contains various information. And considering that the time sequence data can be expressed as the sum of trend items, season items and residual items by an addition model, the trend and periodicity of human behaviors can be analyzed in a finer manner by decomposing the time sequence data. Therefore, the present embodiment decomposes the distance sequence of each key point of each target object using time series decomposition (Seasonal and Trend decomposition using Loess, STL), thereby obtaining a trend term sequence, a season term sequence, and a residual term sequence of the key point. Since the specific implementation process of decomposing each distance sequence by using STL belongs to the prior art, the description thereof is omitted here.
Step S3: and determining each close key point in each key point of each target object and an accuracy index, a close index value and a close speed value of the trend item sequence of each close key point according to the trend item sequence of each key point of each target object.
Since a large amount of information in the distance sequence is contained in the trend item and the season item, the closer the residual item is to Gaussian white noise, the more the first two items are extracted fully. Therefore, in this embodiment, based on the trend item sequences of the key points of each target object, the trend item sequences are clustered, a common rule is found according to the clustering result, potential time sequence patterns in the distance sequences of the key points are identified, and the key points with similar change patterns are classified into one type. By analyzing each type of key points, determining the category close to the key points, and analyzing the sequence information corresponding to the key points in the category close to the key points, the abnormal behavior can be accurately identified and early-warned in time, and the implementation mode is as follows:
clustering the trend item sequences of the key points of each target object to obtain trend item clusters, determining trend item clusters close to the internal engraving machine in the trend item clusters, and determining the key points corresponding to the trend item sequences in the trend item clusters close to the internal engraving machine as the close key points of each target object;
Calculating a distance value of a trend item sequence close to a key point in each trend item cluster of the approaching internal engraving machine of each target object and a trend item sequence of a cluster center when the trend item cluster of the approaching internal engraving machine is acquired, and taking a result of normalizing the opposite number of the distance value as an accuracy index of the trend item sequence corresponding to the approaching key point, thereby obtaining an accuracy index of each trend item sequence close to the key point of each target object;
determining the difference value between the next trend item and the previous trend item in the adjacent two trend item in each trend item sequence close to the key point of each target object, and determining the total number of all the difference values smaller than 0 as a close index value of the trend item sequence close to the key point, thereby obtaining the close index value of each trend item sequence close to the key point of each target object;
and determining the ratio of the absolute value of the sum of all differences which are smaller than 0 and correspond to each trend item sequence which is close to the key point of each target object to the total time which corresponds to all differences which are smaller than 0, and determining the ratio as the close speed value of the trend item sequence which corresponds to the key point, thereby obtaining the close speed value of each trend item sequence which is close to the key point of each target object.
When the trend item clusters close to the internal engraving machine in each trend item cluster are determined, the total number of all differences smaller than 0 corresponding to each trend item sequence in each trend item cluster is determined, the total accumulated sum of the total numbers corresponding to all trend item sequences in each trend item cluster is calculated, and the trend item cluster corresponding to the largest accumulated sum is determined to be the trend item cluster close to the internal engraving machine.
Specifically, for trend item sequences of 17 key points of each target object, a dynamic time warping algorithm (Dynamic Time Warping, DTW) is utilized to calculate a dynamic time warping distance between any two sequences, then based on the dynamic time warping distances, the 17 trend item sequences are divided into three types by utilizing a Kmedoid clustering algorithm, so that three trend item clusters are obtained, and a cluster center of each type is obtained. In the same trend item cluster, the difference of the trend item sequences is small, and in different trend item clusters, the difference of the trend item sequences is large.
At each trendIn potential item clustering, a reduction counter is constructed, for each trend item sequence, if the latter trend item in any two adjacent trend items in the sequence is smaller than the former trend item, the reduction counter counts by one, and each trend item sequence obtains a reduction counter count Calculating the counter count of decrease of all trend item sequences in the trend item cluster>Is a sum of the sums of the numbers. And according to the sequence of accumulating the counts corresponding to the three trend item clusters from large to small, the three trend item clusters are correspondingly close to the trend item cluster of the internal engraving machine, relatively stable in distance and far from the trend item cluster of the internal engraving machine. The key points corresponding to all trend item sequences in the trend item clusters approaching the internal engraving machine are key points approaching the laser internal engraving machine of the target object, which are called as approaching key points herein, for example, when the target object gradually faces the laser internal engraving machine side probe in the surrounding view, the key points on the head of the target object are the key points approaching the laser internal engraving machine. The key points corresponding to all trend item sequences in the distance relative stable trend item clusters are key points of the target object, which are kept at a stable distance from the laser engraving machine, and are called stable key points herein, for example, when the feet of the target object are kept still all the time when the target object is in the surrounding view, the ankle key points are key points with relatively stable distances. The key points corresponding to all trend item sequences in the trend item clusters far away from the internal engraving machine are key points of the target object far away from the laser internal engraving machine, which are called far away key points herein, for example, when the arm of the target object is gradually far away from one side of the laser internal engraving machine when the target object is in the surrounding view, the wrist key points are the key points far away from the laser internal engraving machine.
For each target object, measuring the clustering error of each trend item sequence according to the DTW distance value of each trend item sequence in the trend item clusters of the approaching internal carving machine and the trend item sequence of the clustering center of the clusters, thereby determining the accuracy of dividing each trend item sequence into approaching classes, wherein the corresponding calculation formula is as follows:
wherein ,for the first object of each target objectiAn accuracy index of the trend sequence close to the key point,/->The first in the trend item clusters of the approaching engraving machine for each target objectiA sequence of trend items close to the key point, +.>Trend item sequence approaching the clustering center of the inter-engraving machine trend item clusters for each target object,/->The first in the trend item clusters of the approaching engraving machine for each target objectiA trend item sequence close to the key point +.>Dynamic time-warping distance,/for trend item sequence of clustering center of the near-engraving machine trend item cluster>Is an exponential function based on a natural constant e for normalizing dynamic time by a distance +.>Mapping to [0,1 ]]Interval.
According to the first object of each target objectiAccuracy index of trend item sequence close to key pointAs can be seen from the calculation formula of (2) iThe greater the dynamic time-warping distance between the trend item sequence near the key point and the trend item sequence of the clustering centerDescription of the first embodimentiThe trend item sequence close to the key point is more dissimilar to the trend item sequence of the clustering center, and the first trend item sequence isiThe lower the accuracy of the classification of the trend item sequence close to the key point into the proximity class, the firstiThe smaller the probability that the approaching key point is actually approaching the key point, the smaller the corresponding accuracy index.
For each trend item sequence in the trend item clusters of the approaching engraving machine of each target object, namely each trend item sequence of each approaching key point of each target object, counting a reduction counter corresponding to each trend item sequenceAnd determining a closeness index value corresponding to the trend item sequence close to the key point. When the approach index value is larger, the corresponding approach key point is more likely to approach the internal engraving machine, and the behavior abnormality is more likely to exist.
Meanwhile, for each trend item sequence close to the key point of each target object, the trend item sequence represents the distance between the key point and the laser engraving machine, and the overall trend is smaller, but the trend item sequence does not necessarily show monotonous decrease in time sequence relation, and the situation that the distance becomes closer and farther is caused. In order to measure the trend degree of each approaching key point of each target object to the laser engraving machine, extracting sequence segments with reduced distances from a trend item sequence, and calculating the ratio of the total length of the sequence segments to the total time of the sequence segments so as to obtain an approaching speed value. That is, for each trend item sequence near the key point of each target object, when the next trend item in any two adjacent trend items in the sequence is smaller than the previous trend item, calculating the difference value between the next trend item and the previous trend item, simultaneously calculating the time difference value between the next trend item and the previous trend item, calculating the absolute value of the sum of all the differences in the sequence, thereby obtaining the total approach distance, simultaneously calculating the sum of all the time differences, thereby obtaining the total time, and determining the ratio of the total approach distance to the total time as the approach speed value of the trend item sequence near the key point. When the approaching speed value is larger, the corresponding key points are indicated to continuously trend to the laser internal engraving machine at a higher average speed, the abnormal degree of the behavior is higher, the risk is higher, and stronger early warning is required.
Step S4: according to the seasonal term sequence of each key point of each target object, determining a periodic index of the seasonal term sequence of each target object close to the key point, and according to the residual term sequence of each target object close to the key point, determining a pure randomness index of the residual term sequence of each target object close to the key point.
According to the seasonal item sequence of each key point of each target object, determining the periodic index of the seasonal item sequence of each target object close to the key point, wherein the implementation steps comprise:
clustering the seasonal item sequences of the key points of each target object to obtain seasonal item clusters;
dividing any one seasonal item sequence in each seasonal item cluster into a plurality of subsequences, calculating the absolute value of a pearson correlation coefficient between any two subsequences, setting the periodicity index of each seasonal item sequence in the seasonal item cluster to be 1 if the absolute value of the pearson correlation coefficient is larger than a set coefficient threshold value, otherwise setting the periodicity index of each seasonal item sequence in the seasonal item cluster to be 0, and thus obtaining the periodicity index of each seasonal item sequence, close to a key point, of each target object.
According to the residual error item sequence of each target object close to the key point, determining the pure randomness index of the residual error item sequence of each target object close to the key point, wherein the implementation steps comprise:
and carrying out autocorrelation calculation on the residual error item sequences of each target object, which are close to the key points, so as to determine the pure randomness index of the residual error item sequences of each target object, which are close to the key points.
Specifically, according to the seasonal term sequences of 17 key points of each target object, the 17 seasonal term sequences are divided into two types by using a Kmedoid algorithm, so that two seasonal term clusters, namely periodic and aperiodic types, are obtained. In order to distinguish periodic and non-periodic classes from two seasonal item clusters, one seasonal item sequence is arbitrarily selected from the two seasonal item clusters, each selected seasonal item sequence is divided into a plurality of segments, the absolute value of a Pearson correlation coefficient between any two segments of each seasonal item sequence is calculated, when the absolute value of the Pearson correlation coefficient between the two segments is larger than a set coefficient threshold value of 0.8, the corresponding seasonal item sequence is indicated to be a periodic sequence, periodicity occurs, and the seasonal item cluster where the seasonal item sequence of the periodic sequence is located is the periodic class; when the absolute value of the Pearson correlation coefficient between any two segments is not more than the set coefficient threshold value of 0.8, the corresponding seasonal item sequence is an aperiodic sequence, no periodicity occurs, and at the moment, the seasonal item cluster where the seasonal item sequence of the aperiodic sequence is located is of an aperiodic type.
For the seasonal item sequence in the periodic class, due to the periodicity, the repeated actions of people, such as standing on the foot and watching the people, or continuously probing beside the laser engraving machine, alternately approaching and moving away from the laser engraving machine, and the like, are more likely to have abnormal behaviors after the repeated actions are generated. Therefore, after the periodic class and the aperiodic class are distinguished from the two seasonal item clusters, the periodicity index of each seasonal item sequence in the periodic class is set to 1, and the periodicity index of each seasonal item sequence in the aperiodic class is set to 0, so that the abnormal behavior can be analyzed and judged later.
Meanwhile, according to each residual term sequence close to the key point of each target object, calculating the pure randomness of each residual term sequence component close to the key point by using a time sequence lag first-order autocorrelation formula, thereby obtaining a pure randomness index of each residual term sequence, wherein the corresponding calculation formula is as follows:
wherein ,for the first object of each target objectiPure randomness index of residual sequence near key point,/for the residual sequence near key point>For the first object of each target objectiThe first of the sequence of residual items near the keypointtResidual items->For the first object of each target object iThe first of the sequence of residual items near the keypointt-1 residual item,/->For the first object of each target objectiMean value of all residual terms in residual term sequence of individual key points, +.>For the first object of each target objectiThe total number of all residual terms in the sequence of residual terms that are close to the keypoint.
The first object of each target objectiPure randomness index of residual error item sequence near key pointIn the calculation formula of (a), when no correlation exists between each residual item in the residual item sequence, the corresponding pure randomness index +.>The smaller the value of (c), the more fully the information in the trend term and season term corresponding to the residual term sequence.
Step S5: and determining a key degree factor corresponding to each near key point of each target object according to the accuracy index, the periodicity index and the pure randomness index corresponding to each near key point of each target object.
When judging whether the target object has abnormal behaviors, the performance of a plurality of key points needs to be integrated. For any target object, the distance sequence of 17 key points of the target object has corresponding trend items and season itemsAnd residual terms, sharing according to clustering results of the trend terms The trend terms are divided into proximity categories taking this into account>The influence of each trend item corresponding to the approaching key point is different, firstly, the accuracy index, the periodicity index and the pure randomness index are combined, the key degree factor corresponding to each approaching key point is calculated, and the corresponding calculation formula is as follows:
wherein ,for the first object of each target objectiA key degree factor corresponding to the key point, and +.>For the first object of each target objectiPure randomness index of residual sequence near key point,/for the residual sequence near key point>For the first object of each target objectiAn accuracy index of the trend sequence close to the key point,/->For the first object of each target objectiA periodic index of a sequence of seasonal items near the key point,/->To set the weight coefficient.
The first object of each target objectiThe key degree factors corresponding to the key pointsIn the calculation formula of (2), the pure randomness index of the residual error item sequence near the key point is +.>The smaller the trend and season terms representing the near key point, the higher the distance sequence, thus using +.>As coefficients for calculating the criticality factor. Accuracy index of trend item sequence near key point +.>Characterized in that the trend item is divided into periodic indexes of seasonal item sequences close to key points, which are close to the accuracy of a laser engraving machine >Representing whether the season term shows abnormal periodic behavior, setting a weight coefficient +.>For coordinating accuracy index->And periodicity index->Is a ratio of 0.65. When the pure randomness index->The smaller the accuracy index->And periodicity index->The greater the critical degree factor->The larger this approach key point is represented as more important in all approach key points, the greater the weight.
Step S6: and determining the stay index of the eye key point of each target object according to the position sequence of the eye key point of each target object.
Considering that when the sight line of the target object stays on the laser engraving machine, the head of the laser engraving machine basically keeps stable no matter the laser engraving machine is in static view or is in advancing view, the position change range of two eyes between frames is smaller, the track of the key points of the two eyes is stable or the fluctuation is smaller, and the damage to the target object caused by the behavior is possibly larger. Therefore, it is necessary to determine the stay indicators of the left eye key point and the right eye key point according to the position sequences of the eye key points, that is, the left eye key point and the right eye key point, of the target object, so as to measure the line of sight stay condition of the target object.
When determining the stay index of the eye key point of each target object, the eye key point of each target object comprises a left eye key point and a right eye key point, the position change track of the left eye key point and the right eye key point of each target object is determined according to the position sequence of the left eye key point and the right eye key point of each target object, and the stay index of the left eye key point and the right eye key point is determined according to the coordinates of each point in the position change track. That is, according to the position sequences of the left eye key point and the right eye key point of each target object, the positions in the position sequences are connected according to the time sequence, so that the position change track of the left eye key point and the right eye key point of each target object can be obtained, and the ordinate of the position change track represents the upper and lower positions of the eyes, so that the change condition of the ordinate can be used for representing the change condition of the positions of the eyes. And respectively calculating the variance of the ordinate of the position change track of the left-eye key point and the variance of the ordinate of the position change track of the right-eye key point, and taking the two variances as the stay indexes of the left-eye key point and the right-eye key point. Of course, other embodiments in the prior art may be used to measure the position change of the eye key point, so as to obtain a corresponding retention index, where when the position change of the eye key point is smaller, the corresponding retention index is smaller.
Step S7: and determining the abnormality degree of each target object according to the key degree factor, the approach index value, the approach speed value and the stay index of the eye key point corresponding to each approach key point of each target object, and further determining whether to perform early warning.
Based on the key degree factor, the approach index value, the approach speed value and the retention index of the eye key point corresponding to each approach key point of each target object, the abnormal condition of each target object is measured to determine the abnormal degree of each target object, and the corresponding calculation formula is as follows:
wherein ,for the degree of abnormality of each target object +.>Stay index for right eye key point of each target object, +.>Stay index for left eye key point of each target object, +.>For the first object of each target objectiA key degree factor corresponding to the key point, and +.>For the first object of each target objectiApproach speed value of approach key, +.>For the first object of each target objectiA near index value near the key point,Nfor the total number of near key points of each target object, +.>Is a normalization function.
Degree of abnormality of each target object described aboveIn the calculation formula of (a) iThe product of the approach speed value and the approach index value corresponding to the approach key point +.>Characterizing the firstiThe trend degree of the laser engraving machine near the key point is equal to the product value +.>The larger the indication of the firstiThe greater the tendency of approaching the key point, the more the first is utilizediA critical degree factor of about a critical point>The degree of trending is weighted. Taking the average value of the stay index of the right eye key point and the stay index of the left eye key point into consideration in addition to the trend degree>The reciprocal of (2) is a coefficient of trend, and the smaller the stay index is, the smaller the position change of the eye key point is, and the higher the possibility of the line of sight stay. Finally adopt->Normalizing the final value of the trend degree by a normalization function so as to obtain the abnormality degree of the target object, wherein the value range of the abnormality procedure is [0,1 ]]The larger the value of the abnormality degree is, the more possible abnormal behaviors or line-of-sight stay of the target object are indicated, and early warning is needed.
After obtaining the abnormality degree of each target object, determining whether to perform early warning according to the abnormality degree of each target object, wherein the implementation process comprises the following steps:
determining the maximum abnormality degree in the abnormality degrees of all the target objects, and if the maximum abnormality degree is smaller than a first abnormality degree threshold value, not performing early warning; if the maximum abnormality degree is not less than the first abnormality degree threshold value and is less than the second abnormality degree threshold value, performing first-level early warning; and if the maximum abnormality degree is not less than the second abnormality degree threshold, performing second level early warning.
Specifically, the first abnormality degree threshold is preset to 0.4, and the second abnormality degree threshold is preset to 0.6. After obtaining the degree of abnormality of each target object, the maximum degree of abnormality among all the degrees of abnormality is calculatedAnd (5) early warning is carried out. That is, when->0.4, indicating that slight abnormal behaviors exist around the laser internal engraving machine, and that people can slowly approach the laser internal engraving machine in a safe area, wherein no early warning is sent out; when 0.4 is less than or equal to%>0.6, indicating that a moderate abnormal behavior exists around the laser internal engraving machine, and possibly someone continuously approaches the laser internal engraving machine at a high speed, and sending out an abnormal early warning with moderate sound quantity at the moment to remind not to go on to an unsafe area; when->And when the alarm is more than or equal to 0.6, the serious abnormal behavior exists around the laser internal carving machine, and then loud abnormal early warning is sent out and an alarm is lightened at the same time, so that attention is acquired to warn and prohibit approaching actions.
Based on the same inventive concept, the embodiment also provides a safety pre-warning system for the laser internal engraving machine, which comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor, wherein the processor realizes the steps of the safety pre-warning method for the laser internal engraving machine when executing the computer program. Because the core of the system is to implement the above-mentioned safety pre-warning method for the laser engraving machine, the method is already described in detail in the above-mentioned content, and the system will not be described in detail here.
The application obtains the video image around the laser engraving machine in the past period of time, and obtains the distance sequence of each key point of each target object and the position sequence of the eye key point of each target object according to the video image. The trend item sequence, the season item sequence and the residual item sequence of each key point are obtained by decomposing the distance sequence of each key point of each target object, trend conditions of the target objects around the laser engraving machine can be analyzed in advance by analyzing the trend item sequence, the season item sequence and the residual item sequence of each key point, abnormal trend behaviors can be detected when the body part of a person does not reach a dangerous range, and the sight stay condition of the target objects can be analyzed according to the position sequence of the eye key points. Finally, by combining the trend condition and the sight line stay condition of the target object, the abnormal behaviors of the surrounding staff are analyzed and identified, so that the abnormal dangerous behaviors of the surrounding staff can be accurately determined and early-warned in time, and the problem of poor safety early-warning reliability of the existing laser internal carving machine is effectively solved.
It should be noted that: the above embodiments are only for illustrating the technical solution of the present application, and not for limiting the same; although the application has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present application, and are intended to be included in the scope of the present application.
Claims (10)
1. The safety early warning method for the laser inner carving machine is characterized by comprising the following steps of:
acquiring video images around the laser engraving machine, and acquiring a distance sequence of each key point of each target object and a position sequence of the eye key point of each target object according to the video images;
respectively decomposing the distance sequence of each key point of each target object, and determining the trend item sequence, the season item sequence and the residual item sequence of each key point of each target object;
determining each close key point in each key point of each target object and an accuracy index, a close index value and a close speed value of the trend item sequence of each close key point according to the trend item sequence of each key point of each target object;
according to the seasonal item sequences of the key points of each target object, determining the periodic index of the seasonal item sequences of the key points of each target object, and according to the residual item sequences of the key points of each target object, determining the pure randomness index of the residual item sequences of the key points of each target object;
determining a key degree factor corresponding to each near key point of each target object according to the accuracy index, the periodicity index and the pure randomness index corresponding to each near key point of each target object;
Determining a stay index of the eye key points of each target object according to the position sequence of the eye key points of each target object;
and determining the abnormality degree of each target object according to the key degree factor, the approach index value, the approach speed value and the stay index of the eye key point corresponding to each approach key point of each target object, and further determining whether to perform early warning.
2. The method of claim 1, wherein determining the accuracy index, the approach index value, and the approach speed value of each approach key point and each trend item sequence of each approach key point in each key point of each target object comprises:
clustering the trend item sequences of the key points of each target object to obtain trend item clusters, determining trend item clusters close to the internal engraving machine in the trend item clusters, and determining the key points corresponding to the trend item sequences in the trend item clusters close to the internal engraving machine as the close key points of each target object;
calculating a distance value of a trend item sequence close to a key point in each trend item cluster of the approaching internal engraving machine of each target object and a trend item sequence of a cluster center when the trend item cluster of the approaching internal engraving machine is acquired, and taking a result of normalizing the opposite number of the distance value as an accuracy index of the trend item sequence corresponding to the approaching key point, thereby obtaining an accuracy index of each trend item sequence close to the key point of each target object;
Determining the difference value between the next trend item and the previous trend item in the adjacent two trend item in each trend item sequence close to the key point of each target object, and determining the total number of all the difference values smaller than 0 as a close index value of the trend item sequence close to the key point, thereby obtaining the close index value of each trend item sequence close to the key point of each target object;
and determining the ratio of the absolute value of the sum of all differences which are smaller than 0 and correspond to each trend item sequence which is close to the key point of each target object to the total time which corresponds to all differences which are smaller than 0, and determining the ratio as the close speed value of the trend item sequence which corresponds to the key point, thereby obtaining the close speed value of each trend item sequence which is close to the key point of each target object.
3. The method of claim 2, wherein determining a trend item cluster of the near-engraving machine from among the trend item clusters comprises:
and determining the total number of all differences which are smaller than 0 and correspond to each trend item sequence in each trend item cluster, calculating the total sum of the total numbers which correspond to all trend item sequences in each trend item cluster, and determining the trend item cluster corresponding to the largest sum as the trend item cluster close to the engraving machine.
4. The method of claim 1, wherein determining the periodicity index of the seasonal sequence of items near the keypoints for each target object comprises:
clustering the seasonal item sequences of the key points of each target object to obtain seasonal item clusters;
dividing any one seasonal item sequence in each seasonal item cluster into a plurality of subsequences, calculating the absolute value of a pearson correlation coefficient between any two subsequences, setting the periodicity index of each seasonal item sequence in the seasonal item cluster to be 1 if the absolute value of the pearson correlation coefficient is larger than a set coefficient threshold value, otherwise setting the periodicity index of each seasonal item sequence in the seasonal item cluster to be 0, and thus obtaining the periodicity index of each seasonal item sequence, close to a key point, of each target object.
5. The method of claim 1, wherein determining a purely stochastic indicator of a sequence of residual terms for each target object near a keypoint comprises:
and carrying out autocorrelation calculation on the residual error item sequences of each target object, which are close to the key points, so as to determine the pure randomness index of the residual error item sequences of each target object, which are close to the key points.
6. The safety precaution method for a laser engraving machine according to claim 1, characterized in that determining a key degree factor corresponding to each near key point of each target object, wherein the corresponding calculation formula is:
wherein ,for the first object of each target objectiA key degree factor corresponding to the key point, and +.>For the first object of each target objectiPure randomness index of residual sequence near key point,/for the residual sequence near key point>For the first object of each target objectiAn accuracy index of the trend sequence close to the key point,/->For the first object of each target objectiA periodic index of a sequence of seasonal items near the key point,/->To set the weight coefficient.
7. The method of claim 1, wherein determining the stay indicator for the eye key point of each target object comprises:
the eye key points of each target object comprise a left eye key point and a right eye key point, the position change tracks of the left eye key point and the right eye key point of each target object are determined according to the position sequences of the left eye key point and the right eye key point of each target object, and the stay indexes of the left eye key point and the right eye key point are determined according to the coordinates of each point in the position change tracks.
8. The method for safety precaution of laser engraving machine according to claim 7, characterized in that determining the degree of abnormality of each target object corresponds to the formula:
wherein ,for the degree of abnormality of each target object +.>Stay index for right eye key point of each target object, +.>Stay index for left eye key point of each target object, +.>For the first object of each target objectiA key degree factor corresponding to the key point, and +.>For the first object of each target objectiApproach speed value of approach key, +.>For the first object of each target objectiA near index value near the key point,Nfor the total number of near key points of each target object, +.>Is a normalization function.
9. The method of claim 1, wherein determining whether to perform the pre-warning comprises:
determining the maximum abnormality degree in the abnormality degrees of all the target objects, and if the maximum abnormality degree is smaller than a first abnormality degree threshold value, not performing early warning; if the maximum abnormality degree is not less than the first abnormality degree threshold value and is less than the second abnormality degree threshold value, performing first-level early warning; and if the maximum abnormality degree is not less than the second abnormality degree threshold, performing second level early warning.
10. A safety precaution system for a laser engraving machine comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor, when executing the computer program, realizes the steps of the safety precaution method for a laser engraving machine according to any one of the preceding claims 1-9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310959497.9A CN116652396B (en) | 2023-08-01 | 2023-08-01 | Safety early warning method and system for laser inner carving machine |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310959497.9A CN116652396B (en) | 2023-08-01 | 2023-08-01 | Safety early warning method and system for laser inner carving machine |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116652396A true CN116652396A (en) | 2023-08-29 |
| CN116652396B CN116652396B (en) | 2023-10-10 |
Family
ID=87721051
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310959497.9A Active CN116652396B (en) | 2023-08-01 | 2023-08-01 | Safety early warning method and system for laser inner carving machine |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116652396B (en) |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106096316A (en) * | 2016-07-13 | 2016-11-09 | 复旦大学 | Serious symptom based on the analysis of real time medical data flow point and operation monitoring and pre-warning method and system |
| US20200364618A1 (en) * | 2019-05-14 | 2020-11-19 | Msd International Gmbh | Automated quality check and diagnosis for production model refresh |
| CN113033521A (en) * | 2021-05-25 | 2021-06-25 | 南京甄视智能科技有限公司 | Perimeter dynamic early warning method and system based on target analysis |
| CN113111747A (en) * | 2021-03-31 | 2021-07-13 | 新疆爱华盈通信息技术有限公司 | Abnormal limb behavior detection method, device, terminal and medium |
| CN114067271A (en) * | 2021-11-20 | 2022-02-18 | 河南鑫安利安全科技股份有限公司 | Safety risk early warning method based on AIOT and personnel trajectory analysis |
| CN115659185A (en) * | 2021-12-08 | 2023-01-31 | 云智慧（北京）科技有限公司 | Method and device for processing time sequence data of operation and maintenance system |
| CN115690902A (en) * | 2022-10-19 | 2023-02-03 | 南通鹏宝运动用品有限公司 | Abnormal posture early warning method for body building action |
| CN115761803A (en) * | 2022-11-23 | 2023-03-07 | 山东鲁软数字科技有限公司 | Peripheral pedestrian safety early warning method for electric power operation area |
| CN115937793A (en) * | 2023-03-02 | 2023-04-07 | 广东汇通信息科技股份有限公司 | Student behavior abnormity detection method based on image processing |
| CN116030449A (en) * | 2023-02-17 | 2023-04-28 | 济南邦德激光股份有限公司 | Automatic sorting method and automatic sorting system for laser cutting pieces |
| CN116385948A (en) * | 2023-06-07 | 2023-07-04 | 成都考拉悠然科技有限公司 | System and method for early warning railway side slope abnormality |
-
2023
- 2023-08-01 CN CN202310959497.9A patent/CN116652396B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106096316A (en) * | 2016-07-13 | 2016-11-09 | 复旦大学 | Serious symptom based on the analysis of real time medical data flow point and operation monitoring and pre-warning method and system |
| US20200364618A1 (en) * | 2019-05-14 | 2020-11-19 | Msd International Gmbh | Automated quality check and diagnosis for production model refresh |
| CN113111747A (en) * | 2021-03-31 | 2021-07-13 | 新疆爱华盈通信息技术有限公司 | Abnormal limb behavior detection method, device, terminal and medium |
| CN113033521A (en) * | 2021-05-25 | 2021-06-25 | 南京甄视智能科技有限公司 | Perimeter dynamic early warning method and system based on target analysis |
| CN114067271A (en) * | 2021-11-20 | 2022-02-18 | 河南鑫安利安全科技股份有限公司 | Safety risk early warning method based on AIOT and personnel trajectory analysis |
| CN115659185A (en) * | 2021-12-08 | 2023-01-31 | 云智慧（北京）科技有限公司 | Method and device for processing time sequence data of operation and maintenance system |
| CN115690902A (en) * | 2022-10-19 | 2023-02-03 | 南通鹏宝运动用品有限公司 | Abnormal posture early warning method for body building action |
| CN115761803A (en) * | 2022-11-23 | 2023-03-07 | 山东鲁软数字科技有限公司 | Peripheral pedestrian safety early warning method for electric power operation area |
| CN116030449A (en) * | 2023-02-17 | 2023-04-28 | 济南邦德激光股份有限公司 | Automatic sorting method and automatic sorting system for laser cutting pieces |
| CN115937793A (en) * | 2023-03-02 | 2023-04-07 | 广东汇通信息科技股份有限公司 | Student behavior abnormity detection method based on image processing |
| CN116385948A (en) * | 2023-06-07 | 2023-07-04 | 成都考拉悠然科技有限公司 | System and method for early warning railway side slope abnormality |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116652396B (en) | 2023-10-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107545224B (en) | The method and device of transformer station personnel Activity recognition | |
| CN104605827B (en) | The determination methods of a kind of human body strenuous exercise state and device | |
| CN112926541B (en) | Sleeping post detection method and device and related equipment | |
| CN103186902A (en) | Trip detecting method and device based on video | |
| CN110289061B (en) | Time sequence prediction method for traumatic hemorrhagic shock injury | |
| CN111739243B (en) | D-S evidence theory-based fire state detection method | |
| CN116687363B (en) | Limited space operation personnel health monitoring system based on wearing equipment | |
| CN111753724A (en) | Abnormal behavior identification method and device | |
| CN106792883A (en) | Sensor network abnormal deviation data examination method and system | |
| CN111643092A (en) | Epilepsia alarm device and epilepsia detection method | |
| CN114782897A (en) | Dangerous behavior detection method and system based on machine vision and deep learning | |
| US20110265026A1 (en) | Novelty detection with missing parameters | |
| CN112597903B (en) | Electric power personnel safety state intelligent identification method and medium based on stride measurement | |
| CN116652396B (en) | Safety early warning method and system for laser inner carving machine | |
| CN110428579A (en) | Indoor Monitoring System, method and device based on image recognition | |
| WO2020217812A1 (en) | Image processing device that recognizes state of subject and method for same | |
| CN112168188A (en) | Processing method and device for pressure detection data | |
| Boehner | A smartphone application for a portable fall detection system | |
| KR101483218B1 (en) | Activity diagnosis apparatus | |
| CN116115239A (en) | Embarrassing working gesture recognition method for construction workers based on multi-mode data fusion | |
| CN106840213B (en) | A kind of dangerous material detection method and detector | |
| TWI438726B (en) | System and method for assessment of fall risk | |
| Luca et al. | Anomaly detection using the Poisson process limit for extremes | |
| Chawan et al. | Person Identification And Tinetti Score Prediction Using Balance Parameters: A Machine Learning Approach To Determine Fall Risk | |
| CN111626193A (en) | Face recognition method, face recognition device and readable 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |