CN116912747B - Data processing system based on video identification load foreign matter - Google Patents
Data processing system based on video identification load foreign matter Download PDFInfo
- Publication number
- CN116912747B CN116912747B CN202310982053.7A CN202310982053A CN116912747B CN 116912747 B CN116912747 B CN 116912747B CN 202310982053 A CN202310982053 A CN 202310982053A CN 116912747 B CN116912747 B CN 116912747B
- Authority
- CN
- China
- Prior art keywords
- image
- vehicle
- load
- video
- foreign matter
- 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
- 238000003062 neural network model Methods 0.000 claims abstract description 37
- 238000004590 computer program Methods 0.000 claims abstract description 10
- 238000000034 method Methods 0.000 claims description 12
- 230000006870 function Effects 0.000 claims description 4
- 230000002159 abnormal effect Effects 0.000 claims description 3
- 239000002699 waste material Substances 0.000 description 6
- 230000009286 beneficial effect Effects 0.000 description 5
- 238000001514 detection method Methods 0.000 description 2
- 230000002349 favourable effect Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- 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
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Multimedia (AREA)
- Computational Linguistics (AREA)
- Databases & Information Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a data processing system based on video recognition of load foreign matter, comprising: an initial list of vehicle IDs, a processor and a memory storing a computer program which, when executed by the processor, performs the steps of: acquiring a first image list; acquiring a first priority list; confirming an image positive sample and an image negative sample; obtaining a target neural network model; the third image list corresponding to the target vehicle is input into the target neural network model to acquire the target load foreign matter identification corresponding to the target vehicle, and further whether the load foreign matter capable of influencing vehicle transportation exists in the target vehicle is determined.
Description
Technical Field
The invention relates to the technical field of video identification, in particular to a data processing system based on video identification of load foreign matters.
Background
The method for detecting the occurrence of the load foreign matter is mainly to weigh the weight of the vehicle in an empty state and in a state of pulling the object, and to artificially check whether the load foreign matter exists in the vehicle when the weighed weight of the vehicle is not in a specified weight range.
However, the above method also has the following technical problems:
the manual detection method has uncontrollable property and needs a large amount of manual participation, the work content is tedious and repeated, and when the work load is large, the condition that the manual detection is likely to have misjudgment exists, so that the accuracy of determining whether the load foreign matters exist in the vehicle is low and the resource waste is easy to cause.
Disclosure of Invention
Aiming at the technical problems, the invention adopts the following technical scheme:
a data processing system for identifying load-carrying foreign matter based on video, comprising: initial vehicle ID list a= { a 1 ，……，A i ，……，A m A processor and a memory storing a computer program, wherein A i For the i-th initial vehicle ID, i= … … m, m is the initial vehicle ID number, when the computer program is executed by the processor, the following steps are implemented:
s100, according to A, a first image list B= { B corresponding to A is obtained 1 ，……，B i ，……，B m }，B i ＝{B i1 ，……，B ij ，……，B in }，B ij ＝{B 1 ij ，B 2 ij }，B 1 ij Is A i Corresponding j-th first image, B 2 ij Is B 1 ij And corresponding second images, j= … … n, wherein n is the number of first images, the first images are images when the bearing state of the initial vehicle corresponding to the initial vehicle ID is in no-load, and the second images are images when the bearing state of the initial vehicle corresponding to the initial vehicle ID is in full-load.
S200, according to B, acquiring a first priority list C= { C corresponding to A 1 ，……，C i ，……，C m }，C i Is A i A corresponding first priority.
S300, when C i ≥C 0 At the time, all B 1 ij And B 2 ij As positive samples of the image, otherwise, all B's are taken 1 ij And B 2 ij As an imageNegative sample, wherein C 0 Is a preset first priority threshold.
S400, training the neural network model according to the image positive sample and the image negative sample to obtain a target neural network model, wherein the output result of the target neural network model is a load foreign matter identifier, the load foreign matter identifier is an identifier for representing whether load foreign matters exist in the vehicle, and the load foreign matters are abnormal objects existing in the vehicle when the load state of the vehicle is full.
S500, acquiring a third image list corresponding to the target vehicle, wherein the third image list comprises a plurality of third images and fourth images corresponding to the third images, the third images are images when the bearing state of the target vehicle is in no-load, and the fourth images are images when the bearing state of the target vehicle is in full-load.
S600, inputting the third image list into a target neural network model to obtain a target load foreign matter identifier corresponding to the target vehicle.
And S700, when the target load foreign matter mark is marked as '1', acquiring the thickness H of the load foreign matter in the target vehicle according to the photoelectric sensor.
S800 when H is greater than or equal to H 0 When the load foreign matter capable of influencing the vehicle transportation exists in the target vehicle, otherwise, the load foreign matter capable of influencing the vehicle transportation does not exist in the target vehicle, H 0 Is a preset thickness.
The invention has at least the following beneficial effects:
the embodiment of the invention provides a data processing system based on video recognition of load foreign matters, which comprises the following steps: an initial list of vehicle IDs, a processor and a memory storing a computer program which, when executed by the processor, performs the steps of: acquiring a first image list corresponding to the initial vehicle ID list according to the initial vehicle ID list; acquiring a first priority list corresponding to the initial vehicle ID list according to the first image list; confirming an image positive sample and an image negative sample according to the first priority list; training the neural network model according to the image positive sample and the image negative sample to obtain a target neural network model; acquiring a third image list corresponding to the target vehicle; inputting the third image list into a target neural network model to obtain a target load foreign matter identifier corresponding to the target vehicle; when the target load foreign matter mark is the mark '1', acquiring the thickness of the load foreign matter in the target vehicle according to the photoelectric sensor; judging the thickness of the load foreign matter, and determining whether the load foreign matter capable of influencing the vehicle transportation exists in the target vehicle, wherein the invention can acquire the target neural network model, input a third image list corresponding to the target vehicle into the target neural network model, judge whether the load foreign matter exists in the target vehicle, further judge whether the load foreign matter is the load foreign matter capable of influencing the vehicle transportation, thereby being beneficial to improving the accuracy of determining whether the load foreign matter exists in the vehicle and avoiding resource waste.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flowchart of a data processing system executing a computer program based on video recognition of a load foreign matter according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to fall within the scope of the invention.
The embodiment of the invention provides a data processing system based on video recognition of load foreign matters, which comprises the following steps: initial vehicle ID list a= { a 1 ，……，A i ，……，A m Treatment ofA memory storing a computer program, wherein A i For the i-th initial vehicle ID, i= … … m, m is the initial vehicle ID number, when the computer program is executed by the processor, the following steps are implemented, as shown in fig. 1:
s100, according to A, a first image list B= { B corresponding to A is obtained 1 ，……，B i ，……，B m }，B i ＝{B i1 ，……，B ij ，……，B in }，B ij ＝{B 1 ij ，B 2 ij }，B 1 ij Is A i Corresponding j-th first image, B 2 ij Is B 1 ij And corresponding second images, j= … … n, wherein n is the number of first images, the first images are images when the bearing state of the initial vehicle corresponding to the initial vehicle ID is in no-load, and the second images are images when the bearing state of the initial vehicle corresponding to the initial vehicle ID is in full-load.
Specifically, the initial vehicle ID is a unique identity of the initial vehicle, where the initial vehicle is a vehicle preset by a person skilled in the art according to actual requirements, and will not be described herein.
Further, each initial vehicle ID corresponds to one initial vehicle information.
Further, the initial vehicle information includes at least a vehicle type corresponding to the initial vehicle ID, a vehicle size corresponding to the initial vehicle ID, and a vehicle weight corresponding to the initial vehicle ID.
Specifically, the step S100 includes the steps of obtaining B 1 ij And B 2 ij ：
S101, obtaining A i Corresponding first empty video set D i ＝{D i1 ，……，D ij ，……，D in }，D ij Is A i And the corresponding j-th first idle video is a video acquired by the video acquisition device when the bearing state of the initial vehicle is in an idle state.
Specifically, the acquisition angle of each first idle video is different.
Specifically, the video capture device is stored in the system.
Further, the video capture device may adjust the angle and focal length.
Further, there are 1 or several video acquisition devices in the system.
S103 according to D ij Obtaining D ij Corresponding first empty image list E ij ＝{E 1 ij ，……，E r ij ，……，E s ij }，E r ij For D ij The corresponding r first idle image, r= … … s, s is the number of the first idle images, and the first idle image is any frame image in the first idle video.
S105, obtain D i Corresponding first full video set F i ＝{F i1 ，……，F ij ，……，F in }，F ij For D ij The first full-load video set corresponds to the video, and the video is acquired by the video acquisition device and has the same acquisition angle as the first no-load video when the bearing state of the initial vehicle is in full load.
S107 according to F ij Obtaining F ij Corresponding first full image list G ij ＝{G 1 ij ，……，G k ij ，……，G t ij }，G k ij Is F ij The corresponding kth first full-load image, k= … … t, t is the number of the first full-load images, and the first full-load image is any frame image in the first full-load video.
Specifically, the size and the number of pixels of the first no-load image are the same as the size and the number of pixels of the first full-load image.
S109 according to E r ij And G k ij Acquisition of B 1 ij And B 2 ij 。
Specifically, the step S109 includes the steps of:
s1091, obtain E r ij And G k ij Between (a) and (b)Second priority L rk ij 。
Specifically, the step S1091 includes the following steps:
s1, obtaining E r ij Corresponding first vehicle profile coordinate list U r ij ＝{U r1 ij ，……，U rg ij ，……，U rh ij }，U rg ij For E r ij Corresponding g first vehicle contour coordinates, g= … … h, h being E r ij The corresponding number of the first vehicle contour coordinates is that the first vehicle contour coordinates are coordinates of the contour of the image presented by the initial vehicle when the center of the first empty image is taken as the origin, the transverse direction is the abscissa, and the longitudinal direction is the ordinate in the first empty image, wherein, a person skilled in the art knows that any method for obtaining the coordinates of the contour of the object in the coordinate system in the prior art belongs to the protection scope of the present invention, and is not repeated herein.
S3, obtaining G k ij Corresponding second vehicle contour coordinate list V k ij ＝{V k1 ij ，……，V kg ij ，……，V kh ij }，V kg ij For E r ij The corresponding g second vehicle contour coordinate, where those skilled in the art know that the manner of obtaining the second vehicle contour coordinate is the same as the manner of obtaining the first vehicle contour coordinate, and will not be described herein again.
S5, according to U rg ij And V kg ij Obtaining L rk ij Wherein L is rk ij Meets the following conditions:
L rk ij =count/h, count is U rg ij And V is equal to kg ij Identical U rg ij Is a number of (3).
The method comprises the steps of obtaining a first vehicle contour coordinate list and a second vehicle coordinate contour list, obtaining a second priority according to the first vehicle contour coordinate and the second vehicle contour coordinate, judging the second priority and the third priority, determining a first image and a second image in a first image list, obtaining the first priority corresponding to an initial vehicle ID according to the first image list, further determining an image positive sample and an image negative sample, training a neural network model according to the image positive sample and the image negative sample, obtaining a target neural network model, inputting a third image list corresponding to the target vehicle into the target neural network model, judging whether a load foreign matter exists in the target vehicle, further judging whether the load foreign matter is the load foreign matter which can influence vehicle transportation, and being beneficial to improving the accuracy of determining whether the load foreign matter exists in the vehicle and avoiding resource waste.
S1093, obtain E r ij And G k ij Third priority P between rk ij 。
Specifically, the step S1093 includes the following steps:
s10, obtaining E r ij Corresponding first pixel point quantity XS r1 ij The number of the first pixels is the number of pixels occupied by the image presented by the initial vehicle in the first idle image, wherein a person skilled in the art knows that any method for obtaining the number of pixels occupied by the image presented by the object in the whole image in the prior art belongs to the protection scope of the present invention, and is not described herein.
S30, obtaining G k ij Corresponding second pixel point quantity XS k2 ij The number of the second pixels is the number of pixels occupied by the image presented by the initial vehicle in the first full image, and those skilled in the art know that the manner of obtaining the number of the second pixels is the same as the manner of obtaining the number of the first pixels, which is not described herein again.
S50, according to XS r1 ij And XS k2 ij Obtaining P rk ij Wherein P is rk ij Meets the following conditions:
P rk ij ＝min(XS r1 ij ，XS k2 ij )/max(XS r1 ij ，XS k2 ij ) Min () is a minimum valued function, and max () is a maximum valued function.
The method comprises the steps of obtaining the first pixel number and the second pixel number, obtaining the third priority according to the first pixel number and the second pixel number, judging the second priority and the third priority, determining the first image and the second image in the first image list, obtaining the first priority corresponding to the initial vehicle ID according to the first image list, further determining the image positive sample and the image negative sample, training the neural network model according to the image positive sample and the image negative sample, obtaining the target neural network model, inputting the third image list corresponding to the target vehicle into the target neural network model, judging whether the load foreign matter exists in the target vehicle, further judging whether the load foreign matter is the load foreign matter which can influence the transportation of the vehicle, and being beneficial to improving the accuracy of determining whether the load foreign matter exists in the vehicle and avoiding the resource waste.
S1095, when L rk ij ≥L 0 And P is rk ij ≥P 0 When E is to r ij As B 1 ij ，G k ij As B 2 ij Wherein, the method comprises the steps of, wherein, L0 to preset the second priority threshold, P 0 A third priority threshold is preset.
Specifically, L 0 The value range of (5) is [0.8,1 ]]The person skilled in the art knows that the person skilled in the art sets the value of the preset second priority threshold according to the actual requirement, which is not described herein.
Specifically, P 0 The value range of (5) is [0.8,1 ]]The person skilled in the art knows that the person skilled in the art sets the value of the preset third priority threshold according to the actual requirement, and will not be described herein.
S1097, when L rk ij ＜L 0 Time-order JD ij ＝JD ij +J 0 To obtain updated F ij And performs S107 step, JD ij For collecting F ij Video acquisition device of (a)Acquisition angle, J 0 The preset angle is known to those skilled in the art, and the value of the preset angle is set by those skilled in the art according to the actual requirement, which is not described herein.
S1099 when P rk ij ＜P 0 When making JU ij ＝JU ij +U 0 To obtain updated F ij And performs step S107, JU ij For collecting F ij Is the focal length of the video acquisition device, U 0 The preset focal length difference is known to those skilled in the art, and the preset focal length difference is set by those skilled in the art according to actual requirements, which will not be described herein.
Above-mentioned, obtain first empty load image and first full load image through first empty load video collection and first full load video collection, according to first empty load image and first full load image, confirm first image list, according to first image list, obtain the first priority that initial vehicle ID corresponds, and then confirm image positive sample and image negative sample, train neural network model according to image positive sample and image negative sample, obtain target neural network model, input the third image list that target vehicle corresponds into target neural network model, judge whether there is the load foreign matter in the target vehicle, further judge, whether the load foreign matter is the load foreign matter that can influence the vehicle transportation, be favorable to improving the accuracy of confirming whether there is the load foreign matter in the vehicle and avoid the wasting of resources.
S200, according to B, acquiring a first priority list C= { C corresponding to A 1 ，……，C i ，……，C m }，C i Is A i A corresponding first priority.
Specifically, the step S200 includes the steps of:
s201, obtain B 1 ij Corresponding first sub-image list TX 1 ij ＝{TX 11 ij ，……，TX 1e ij ，……，TX 1f ij }，TX 1e ij Is B 1 ij A corresponding e-th first sub-image, e= … … fF is the number of first sub-images, which are part of the first images.
S203, obtain B 2 ij Corresponding second sub-image list TX 2 ij ＝{TX 21 ij ，……，TX 2e ij ，……，TX 2f ij }，TX 2e ij Is B 2 ij And the corresponding e second sub-image is a part of images in the second image.
Specifically, the first sub-image and the second sub-image are the same size.
Further, the position of the e first sub-image in the first image is the same as the position of the e second sub-image in the second image.
S205, according to TX 1e ij And TX (transmit x) 2e ij Obtaining C i Wherein C i Meets the following conditions:
C i ＝(Σ n j＝1 (Σ f e＝1 XS e ij )/f)/n，XS e ij for TX 1e ij And TX (transmit x) 2e ij The image similarity between the two images, wherein, the person skilled in the art knows that any method for obtaining the image similarity between the two images in the prior art belongs to the protection scope of the present invention, and is not described herein.
According to the first sub-image list corresponding to the first image and the second sub-image list corresponding to the second image, the image similarity between the first sub-image and the second sub-image corresponding to the first sub-image is obtained, the first priority corresponding to the initial vehicle ID is further obtained, the image positive sample and the image negative sample are further determined, the neural network model is trained according to the image positive sample and the image negative sample, the target neural network model is obtained, the third image list corresponding to the target vehicle is input into the target neural network model, whether the load foreign matter exists in the target vehicle is judged, whether the load foreign matter is the load foreign matter which can influence the transportation of the vehicle is further judged, and the accuracy of determining whether the load foreign matter exists in the vehicle is improved, and the resource waste is avoided.
S300, when C i ≥C 0 At the time, all B 1 ij And B 2 ij As positive samples of the image, otherwise, all B's are taken 1 ij And B 2 ij As a negative sample of the image, wherein C 0 For a preset first priority threshold, C is known to those skilled in the art 0 The values of (2) are set by those skilled in the art according to actual requirements, and are not described herein.
S400, training the neural network model according to the image positive sample and the image negative sample to obtain a target neural network model, wherein the output result of the target neural network model is a load foreign matter identifier, the load foreign matter identifier is an identifier for representing whether load foreign matters exist in the vehicle, and the load foreign matters are abnormal objects existing in the vehicle when the load state of the vehicle is full.
Specifically, the load foreign matter is identified as an identification "1" or an identification "0".
Further, the designation "1" is characterized by the presence of a load carrying foreign object.
Further, the designation "0" is characterized by the absence of a payload foreign body.
S500, acquiring a third image list corresponding to the target vehicle, wherein the third image list comprises a plurality of third images and fourth images corresponding to the third images, the third images are images when the bearing state of the target vehicle is in no-load, and the fourth images are images when the bearing state of the target vehicle is in full-load.
Specifically, the step S500 includes the steps of:
s501, acquiring a third video set corresponding to the target vehicle volume, wherein the third video set comprises a plurality of third videos, and the third videos are videos acquired by a video acquisition device when the target vehicle is empty and weighted.
S503, loading the target vehicle according to a preset loading rule, wherein the preset loading rule is known to a person skilled in the art, and is preset according to actual requirements, and is not described herein.
S505, a fourth video set corresponding to the target vehicle is obtained, the fourth video set comprises a plurality of fourth videos, and the fourth videos are videos acquired by the video acquisition device when the target vehicle is loaded and weighted fully.
Specifically, the third video number is the same as the fourth video number.
Specifically, the third video corresponds to the fourth video one by one, and the acquisition angle of the video acquisition device for acquiring the third video is the same as the acquisition angle of the video acquisition device for acquiring the fourth video corresponding to the third video, which can be understood as: the acquisition angle of the video acquisition device for acquiring the first third video is the same as the acquisition angle of the video acquisition device for acquiring the first fourth video, the acquisition angle of the video acquisition device for acquiring the second third video is the same as the acquisition angle of the video acquisition device for acquiring the second fourth video, and … …, the acquisition angle of the video acquisition device for acquiring the last third video is the same as the acquisition angle of the video acquisition device for acquiring the last fourth video.
S507, obtaining a third image list according to the third video set and the fourth video set, wherein a person skilled in the art knows that a manner of obtaining the third image list is the same as a manner of obtaining the first image list in the step S100, and details are not repeated here.
Above-mentioned, obtain the third video collection and the fourth video collection that target vehicle corresponds, obtain the third image list according to third video collection and fourth video collection, input the third image list into the target neural network model, judge whether there is the load foreign matter in the target vehicle, further judge whether the load foreign matter is the load foreign matter that can influence the vehicle transportation, be favorable to improving the accuracy of confirming whether there is the load foreign matter in the vehicle and avoid the wasting of resources.
S600, inputting the third image list into a target neural network model to obtain a target load foreign matter identifier corresponding to the target vehicle.
S700, when the target load foreign matter mark is the mark '1', the thickness H of the load foreign matter in the target vehicle is obtained according to the photoelectric sensor, wherein a person skilled in the art knows that any method for obtaining the thickness of the object according to the photoelectric sensor in the prior art belongs to the protection scope of the invention, and the description is omitted herein.
S800 when H is greater than or equal to H 0 When the load foreign matter capable of influencing the vehicle transportation exists in the target vehicle, otherwise, the load foreign matter capable of influencing the vehicle transportation does not exist in the target vehicle, H 0 The preset thickness is known to those skilled in the art, and the value of the preset thickness is set by those skilled in the art according to actual requirements, and will not be described herein.
The embodiment of the invention provides a data processing system based on video recognition of load foreign matters, which comprises the following steps: an initial list of vehicle IDs, a processor and a memory storing a computer program which, when executed by the processor, performs the steps of: acquiring a first image list corresponding to the initial vehicle ID list according to the initial vehicle ID list; acquiring a first priority list corresponding to the initial vehicle ID list according to the first image list; confirming an image positive sample and an image negative sample according to the first priority list; training the neural network model according to the image positive sample and the image negative sample to obtain a target neural network model; acquiring a third image list corresponding to the target vehicle; inputting the third image list into a target neural network model to obtain a target load foreign matter identifier corresponding to the target vehicle; when the target load foreign matter mark is the mark '1', acquiring the thickness of the load foreign matter in the target vehicle according to the photoelectric sensor; judging the thickness of the load foreign matter, and determining whether the load foreign matter capable of influencing the vehicle transportation exists in the target vehicle, wherein the invention can acquire the target neural network model, input a third image list corresponding to the target vehicle into the target neural network model, judge whether the load foreign matter exists in the target vehicle, further judge whether the load foreign matter is the load foreign matter capable of influencing the vehicle transportation, thereby being beneficial to improving the accuracy of determining whether the load foreign matter exists in the vehicle and avoiding resource waste.
While certain specific embodiments of the invention have been described in detail by way of example, it will be appreciated by those skilled in the art that the above examples are for illustration only and are not intended to limit the scope of the invention. Those skilled in the art will also appreciate that many modifications may be made to the embodiments without departing from the scope and spirit of the invention. The scope of the invention is defined by the appended claims.
Claims (4)
1. A data processing system for identifying load bearing foreign objects based on video, the system comprising: initial vehicle ID list a= { a 1 ，……，A i ，……，A m A processor and a memory storing a computer program, wherein A i For the i-th initial vehicle ID, i= … … m, m is the initial vehicle ID number, when the computer program is executed by the processor, the following steps are implemented:
s100, according to A, a first image list B= { B corresponding to A is obtained 1 ，……，B i ，……，B m }，B i ＝{B i1 ，……，B ij ，……，B in }，B ij ＝{B 1 ij ，B 2 ij }，B 1 ij Is A i Corresponding j-th first image, B 2 ij Is B 1 ij The corresponding second images, j= … … n, n are the first image number, the first image is an image when the bearing state of the initial vehicle corresponding to the initial vehicle ID is in no load, and the second image is an image when the bearing state of the initial vehicle corresponding to the initial vehicle ID is in full load; the step S100 includes the steps of obtaining B 1 ij And B 2 ij ：
S101, obtaining A i Corresponding first empty video set D i ＝{D i1 ，……，D ij ，……，D in }，D ij Is A i The corresponding j-th first no-load video is a video acquired by the video acquisition device when the bearing state of the initial vehicle is in no-load;
s103 according to D ij Obtaining D ij Corresponding first no-loadImage list E ij ＝{E 1 ij ，……，E r ij ，……，E s ij }，E r ij For D ij The corresponding r first idle image, r= … … s, s is the number of the first idle images, and the first idle image is any frame image in the first idle video;
s105, obtain D i Corresponding first full video set F i ＝{F i1 ，……，F ij ，……，F in }，F ij For D ij The first full-load video set is a video which is acquired by the video acquisition device and has the same acquisition angle as the first no-load video when the bearing state of the initial vehicle is in full load;
s107 according to F ij Obtaining F ij Corresponding first full image list G ij ＝{G 1 ij ，……，G k ij ，……，G t ij }，G k ij Is F ij The corresponding kth first full-load image, k= … … t, t is the number of the first full-load images, and the first full-load image is any frame image in the first full-load video;
s109 according to E r ij And G k ij Acquisition of B 1 ij And B 2 ij The method comprises the steps of carrying out a first treatment on the surface of the The step S109 includes the steps of:
s1091, obtain E r ij And G k ij Second priority L between rk ij The method comprises the steps of carrying out a first treatment on the surface of the The step S1091 includes the following steps:
s1, obtaining E r ij Corresponding first vehicle profile coordinate list U r ij ＝{U r1 ij ，……，U rg ij ，……，U rh ij }，U rg ij For E r ij Corresponding g first vehicle contour coordinates, g= … … h, h being E r ij A corresponding number of first vehicle contour coordinates, the first vehicle contour coordinates being in the first empty image, in a firstThe center of an empty image is the origin, the transverse direction is the abscissa, and the longitudinal direction is the ordinate of the contour of the image presented by the initial vehicle;
s3, obtaining G k ij Corresponding second vehicle contour coordinate list V k ij ＝{V k1 ij ，……，V kg ij ，……，V kh ij }，V kg ij For E r ij A corresponding g-th second vehicle profile coordinate;
s5, according to U rg ij And V kg ij Obtaining L rk ij Wherein L is rk ij Meets the following conditions:
L rk ij =count/h, count is U rg ij And V is equal to kg ij Identical U rg ij Is the number of (3);
s1093, obtain E r ij And G k ij Third priority P between rk ij The method comprises the steps of carrying out a first treatment on the surface of the The step S1093 includes the following steps:
s10, obtaining E r ij Corresponding first pixel point quantity XS r1 ij The first pixel point number is the number of pixel points occupied by an image presented by the initial vehicle in the first empty image;
s30, obtaining G k ij Corresponding second pixel point quantity XS k2 ij The second pixel number is the number of pixels occupied by the image presented by the initial vehicle in the first full image;
s50, according to XS r1 ij And XS k2 ij Obtaining P rk ij Wherein P is rk ij Meets the following conditions:
P rk ij ＝min(XS r1 ij ，XS k2 ij )/max(XS r1 ij ，XS k2 ij ) Min () is a minimum value valued function, and max () is a maximum value valued function;
S1095、when L rk ij ≥L 0 And P is rk ij ≥P 0 When E is to r ij As B 1 ij ，G k ij As B 2 ij Wherein L is 0 To preset the second priority threshold, P 0 Presetting a third priority threshold;
s1097, when L rk ij ＜L 0 Time-order JD ij ＝JD ij +J 0 To obtain updated F ij And performs S107 step, JD ij For collecting F ij Acquisition angle, J of video acquisition device 0 Is a preset angle;
s1099 when P rk ij ＜P 0 When making JU ij ＝JU ij +U 0 To obtain updated F ij And performs step S107, JU ij For collecting F ij Is the focal length of the video acquisition device, U 0 A preset focal length difference value;
s200, according to B, acquiring a first priority list C= { C corresponding to A 1 ，……，C i ，……，C m }，C i Is A i A corresponding first priority; the step S200 includes the steps of:
s201, obtain B 1 ij Corresponding first sub-image list TX 1 ij ＝{TX 11 ij ，……，TX 1e ij ，……，TX 1f ij }，TX 1e ij Is B 1 ij The corresponding e first sub-image, e= … … f, f is the number of the first sub-images, and the first sub-image is a part of the first images;
s203, obtain B 2 ij Corresponding second sub-image list TX 2 ij ＝{TX 21 ij ，……，TX 2e ij ，……，TX 2f ij }，TX 2e ij Is B 2 ij A corresponding e second sub-image, the second sub-image being a portion of the second image;
s205, according to TX 1e ij And TX (transmit x) 2e ij Obtaining C i Wherein C i Meets the following conditions:
C i ＝(Σ n j＝1 (Σ f e＝1 XS e ij )/f)/n，XS e ij for TX 1e ij And TX (transmit x) 2e ij Image similarity between them;
s300, when C i ≥C 0 At the time, all B 1 ij And B 2 ij As positive samples of the image, otherwise, all B's are taken 1 ij And B 2 ij As a negative sample of the image, wherein C 0 Is a preset first priority threshold;
s400, training a neural network model according to an image positive sample and an image negative sample to obtain a target neural network model, wherein a result output by the target neural network model is a load foreign matter identifier, the load foreign matter identifier is an identifier for representing whether load foreign matters exist in a vehicle, and the load foreign matters are abnormal objects existing in the vehicle when the load state of the vehicle is full;
s500, acquiring a third image list corresponding to the target vehicle, wherein the third image list comprises a plurality of third images and fourth images corresponding to the third images, the third images are images when the bearing state of the target vehicle is in no-load, and the fourth images are images when the bearing state of the target vehicle is in full-load;
s600, inputting the third image list into a target neural network model to obtain a target load foreign matter identifier corresponding to the target vehicle;
s700, when the target load foreign matter mark is marked as '1', acquiring the thickness H of the load foreign matter in the target vehicle according to the photoelectric sensor;
s800 when H is greater than or equal to H 0 When the load foreign matter capable of influencing the vehicle transportation exists in the target vehicle, otherwise, the load foreign matter capable of influencing the vehicle transportation does not exist in the target vehicle, H 0 Is a preset thickness.
2. The data processing system based on video recognition of heavy load foreign matter of claim 1, wherein the acquisition angle of each first empty video is different.
3. The video-recognition-based data processing system of claim 1, wherein the size and the number of pixels of the first empty image are the same as the size and the number of pixels of the first full image.
4. The data processing system based on video recognition of a heavy load foreign matter according to claim 1, comprising the steps of, in step S500:
s501, acquiring a third video set corresponding to the target vehicle volume, wherein the third video set comprises a plurality of third videos, and the third videos are videos acquired by a video acquisition device when the target vehicle is empty and weighted;
s503, loading the target vehicle according to a preset loading rule;
s505, acquiring a fourth video set corresponding to the target vehicle, wherein the fourth video set comprises a plurality of fourth videos, and the fourth videos are videos acquired by a video acquisition device when the target vehicle is loaded and weighted fully;
s507, acquiring a third image list according to the third video set and the fourth video set.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310982053.7A CN116912747B (en) | 2023-08-04 | 2023-08-04 | Data processing system based on video identification load foreign matter |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310982053.7A CN116912747B (en) | 2023-08-04 | 2023-08-04 | Data processing system based on video identification load foreign matter |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116912747A CN116912747A (en) | 2023-10-20 |
| CN116912747B true CN116912747B (en) | 2024-04-05 |
Family
ID=88360151
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310982053.7A Active CN116912747B (en) | 2023-08-04 | 2023-08-04 | Data processing system based on video identification load foreign matter |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116912747B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018104068A1 (en) * | 2016-12-05 | 2018-06-14 | Koninklijke Philips N.V. | Foreign object detection in a wireless power transfer system |
| CN108960107A (en) * | 2018-06-25 | 2018-12-07 | 安徽百诚慧通科技有限公司 | A kind of overcrowding recognition methods of small mini van and device |
| CN110633690A (en) * | 2019-09-24 | 2019-12-31 | 北京邮电大学 | Vehicle feature identification method and system based on bridge monitoring |
| CN111274843A (en) * | 2018-11-16 | 2020-06-12 | 上海交通大学 | Truck overload monitoring method and system based on monitoring video |
| CN112308003A (en) * | 2020-11-06 | 2021-02-02 | 中冶赛迪重庆信息技术有限公司 | Method, system, equipment and medium for identifying loading state of scrap steel truck |
| WO2021088381A1 (en) * | 2019-11-06 | 2021-05-14 | 北京交通大学 | Power distribution and vehicle self-learning-based truck overload identification method |
| CN113033284A (en) * | 2020-12-22 | 2021-06-25 | 迪比（重庆）智能科技研究院有限公司 | Vehicle real-time overload detection method based on convolutional neural network |
| CN114758275A (en) * | 2022-04-13 | 2022-07-15 | 浪潮通信信息系统有限公司 | Vehicle video detection method based on GPU |
| CN115817506A (en) * | 2022-10-31 | 2023-03-21 | 东风商用车有限公司 | Vehicle load state identification method and device, electronic equipment and storage medium |
| CN116026441A (en) * | 2023-02-16 | 2023-04-28 | 北京中交慧联信息科技有限公司 | Method, device, equipment and storage medium for detecting abnormal load capacity of vehicle |
-
2023
- 2023-08-04 CN CN202310982053.7A patent/CN116912747B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018104068A1 (en) * | 2016-12-05 | 2018-06-14 | Koninklijke Philips N.V. | Foreign object detection in a wireless power transfer system |
| CN108960107A (en) * | 2018-06-25 | 2018-12-07 | 安徽百诚慧通科技有限公司 | A kind of overcrowding recognition methods of small mini van and device |
| CN111274843A (en) * | 2018-11-16 | 2020-06-12 | 上海交通大学 | Truck overload monitoring method and system based on monitoring video |
| CN110633690A (en) * | 2019-09-24 | 2019-12-31 | 北京邮电大学 | Vehicle feature identification method and system based on bridge monitoring |
| WO2021088381A1 (en) * | 2019-11-06 | 2021-05-14 | 北京交通大学 | Power distribution and vehicle self-learning-based truck overload identification method |
| CN112308003A (en) * | 2020-11-06 | 2021-02-02 | 中冶赛迪重庆信息技术有限公司 | Method, system, equipment and medium for identifying loading state of scrap steel truck |
| CN113033284A (en) * | 2020-12-22 | 2021-06-25 | 迪比（重庆）智能科技研究院有限公司 | Vehicle real-time overload detection method based on convolutional neural network |
| CN114758275A (en) * | 2022-04-13 | 2022-07-15 | 浪潮通信信息系统有限公司 | Vehicle video detection method based on GPU |
| CN115817506A (en) * | 2022-10-31 | 2023-03-21 | 东风商用车有限公司 | Vehicle load state identification method and device, electronic equipment and storage medium |
| CN116026441A (en) * | 2023-02-16 | 2023-04-28 | 北京中交慧联信息科技有限公司 | Method, device, equipment and storage medium for detecting abnormal load capacity of vehicle |
Non-Patent Citations (2)
| Title |
|---|
| 基于计算机视觉的运砂船超载状态检测;王炎龙 等;《现代计算机(专业版) 》;20141115;全文 * |
| 模糊神经网络的自动变速汽车换挡规律分析;刘振军;胡建军;李光辉;秦大同;;重庆大学学报;20090815(08);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116912747A (en) | 2023-10-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP3843036B1 (en) | Sample labeling method and device, and damage category identification method and device | |
| CN111639629B (en) | Pig weight measurement method and device based on image processing and storage medium | |
| CN116912747B (en) | Data processing system based on video identification load foreign matter | |
| CN115995056A (en) | Automatic bridge disease identification method based on deep learning | |
| CN110427845B (en) | Method, device and equipment for determining pixel center of article and readable storage medium | |
| CN115660647A (en) | Maintenance method for building outer wall | |
| CN115463844A (en) | Intelligent cargo sorting method and system based on dual recognition | |
| CN117161589B (en) | Intelligent detection method and system for marking deviation of laser engraving machine | |
| CN110688977B (en) | Industrial image identification method and device, server and storage medium | |
| CN111127494B (en) | Highway truck weight limiting identification method based on image processing | |
| CN112784494A (en) | Training method of false positive recognition model, target recognition method and device | |
| CN114821658B (en) | Face recognition method, operation control device, electronic equipment and storage medium | |
| CN116612098A (en) | Insulator RTV spraying quality evaluation method and device based on image processing | |
| CN111160339A (en) | License plate correction method, image processing equipment and device with storage function | |
| CN114821513A (en) | Image processing method and device based on multilayer network and electronic equipment | |
| CN116091389A (en) | Image detection method based on classification model, electronic equipment and medium | |
| CN113705672A (en) | Threshold value selection method, system and device for image target detection and storage medium | |
| CN115147852A (en) | Ancient book identification method, ancient book identification device, ancient book storage medium and ancient book storage equipment | |
| CN114387451A (en) | Training method, device and medium for abnormal image detection model | |
| CN112529928A (en) | Part assembly detection method, computer device and storage medium | |
| CN111222473A (en) | Analysis and recognition method for clustering faces in video | |
| CN111062298A (en) | Power distribution network power equipment target identification method and system | |
| CN112801214B (en) | Mouse quantity prediction method based on interaction of mouse recognition terminal and cloud computing platform | |
| CN112016842B (en) | Method and device for automatically distributing distribution tasks based on Bayesian algorithm | |
| CN117392594A (en) | Video intelligent monitoring system and method for coal conveying trestle |
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 |