CN114626470A - Aircraft skin key feature detection method based on multi-type geometric feature operator - Google Patents
Aircraft skin key feature detection method based on multi-type geometric feature operator Download PDFInfo
- Publication number
- CN114626470A CN114626470A CN202210268285.1A CN202210268285A CN114626470A CN 114626470 A CN114626470 A CN 114626470A CN 202210268285 A CN202210268285 A CN 202210268285A CN 114626470 A CN114626470 A CN 114626470A
- Authority
- CN
- China
- Prior art keywords
- point cloud
- rivet
- operator
- features
- attention
- 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
- 238000001514 detection method Methods 0.000 title abstract description 9
- 238000000034 method Methods 0.000 claims abstract description 54
- 210000001503 joint Anatomy 0.000 claims abstract description 47
- 230000008447 perception Effects 0.000 claims abstract description 24
- 238000000605 extraction Methods 0.000 claims abstract description 15
- 238000004458 analytical method Methods 0.000 claims abstract description 12
- 238000005070 sampling Methods 0.000 claims description 24
- 230000004927 fusion Effects 0.000 claims description 18
- 230000008569 process Effects 0.000 claims description 14
- 230000006870 function Effects 0.000 claims description 13
- 239000011159 matrix material Substances 0.000 claims description 13
- 239000013598 vector Substances 0.000 claims description 13
- 230000004913 activation Effects 0.000 claims description 9
- 230000002708 enhancing effect Effects 0.000 claims description 9
- 230000000717 retained effect Effects 0.000 claims description 6
- 239000000284 extract Substances 0.000 claims description 4
- 230000007246 mechanism Effects 0.000 claims description 4
- 238000011176 pooling Methods 0.000 claims description 4
- 238000013527 convolutional neural network Methods 0.000 claims description 3
- 238000005520 cutting process Methods 0.000 claims description 3
- 238000009826 distribution Methods 0.000 claims description 3
- 238000012847 principal component analysis method Methods 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 2
- 230000000712 assembly Effects 0.000 description 1
- 238000000429 assembly Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- 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
- G06N3/045—Combinations of 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
- 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)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an aircraft skin key feature detection method based on a multi-type geometric feature operator, which comprises the following steps: acquiring three-dimensional point cloud data of a large aircraft skin; constructing an operator atlas containing multi-type 2D geometric features for each point in the three-dimensional point cloud data; inputting the constructed multi-type 2D geometric feature operator sub-image set into a trained double-attention and multi-scale perception point cloud classification network, learning the features of rivets and butt joints in a plurality of operators, and classifying three-dimensional point cloud data; extracting rivet contour features in a rivet area and butt seam contour features in a butt seam area by adopting a multilevel model fitting algorithm; and performing rivet levelness analysis and butt seam step analysis based on the extracted rivet profile characteristics and butt seam profile characteristics. The method for detecting the butt joint of the rivet on the surface of the airplane realizes the accurate and automatic extraction of the small-scale key characteristic rivet and the small butt joint on the surface of the airplane skin, and improves the assembly quality of the airplane.
Description
Technical Field
The invention relates to the field of three-dimensional point cloud model detection, in particular to an aircraft skin key feature detection method based on multi-type geometric feature operators.
Background
The assembly of large aircraft is a process of combining millions of aircraft parts and connecting pieces according to design and technical requirements, and gradually connecting the parts into assemblies, parts, large parts and complete machines. Because the large-scale airplane has large number of parts, various configurations and complex assembly levels and mutual constraint relations, various errors of the components are mutually superposed and transmitted in the step-by-step assembly process, so that joint surfaces cannot be tightly combined, and a butt joint with the length of tens of meters is generated; in addition, due to the special nature of the materials from which large aircraft are made, it is determined that the skin cannot be welded to the fuselage, but rather that millions of rivets are used to secure the skin to the fuselage. The aeronautical manufacturing industry has strict standard requirements on two key characteristics of butt joint and rivets, and if the two indexes of 'butt joint step difference' and 'rivet flatness' are not controlled in a design range, the structural strength of a large airplane can be reduced, and the potential safety hazard is formed. Therefore, analysis of critical skin features for large aircraft is required.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides the aircraft skin key feature detection method based on the multi-type geometric feature operator, which realizes the accurate automatic extraction of small-scale key feature rivets and small butt seams on the surface of the aircraft skin and improves the aircraft assembly quality.
In order to achieve the purpose, the invention adopts the following technical scheme: a method for detecting aircraft skin key features based on multi-type geometric feature operators comprises the following steps:
(1) acquiring three-dimensional point cloud data of a large aircraft skin;
(2) constructing an operator atlas set containing multi-type 2D geometric features for each point in the three-dimensional point cloud data based on multiple geometric attributes of the three-dimensional point cloud data;
(3) inputting the multi-type 2D geometric feature operator image set constructed in the step (2) into a trained double-attention and multi-scale perception point cloud classification network, learning the features of rivets and butt joints in a plurality of operators, and classifying three-dimensional point cloud data into a rivet region, a butt joint region and a non-feature region;
(4) extracting rivet contour features in a rivet area and butt seam contour features in a butt seam area by adopting a multilevel model fitting algorithm;
(5) and performing rivet levelness analysis and butt seam step analysis based on the extracted rivet profile characteristics and butt seam profile characteristics.
Further, step (1) comprises the following sub-steps:
(1.1) repeatedly acquiring three-dimensional point cloud data of the skin of the airplane by using a LeicaATS960 absolute tracker, wherein the repeatedly acquired three-dimensional point cloud data cover a key feature area to be detected on the surface of the skin of the airplane;
and (1.2) splicing the three-dimensional point cloud data acquired for multiple times into complete large aircraft skin three-dimensional point cloud data by using an iterative closest point algorithm.
Further, the step (2) comprises the following sub-steps:
(2.1) for each point p in the three-dimensional point cloud data, calculating the distance from all points in the three-dimensional point cloud data to the point, taking the nearest k points as a neighborhood point set, and calculating the normal vectors of the point p and all points in the neighborhood point set by using a principal component analysis method; establishing a local coordinate system by taking the point p as an origin and taking the normal vector of the point p as a Z axis, projecting the point p and a neighborhood point set thereof to an XY-plane of the local coordinate system, constructing a minimum square bounding box with sides parallel to the X axis and the Y axis, and cutting the bounding box into n multiplied by n plane grids;
(2.2) calculating the projection height, normal vector, Gaussian curvature, average curvature and geometric information of the dense density coding neighborhood points of the neighborhood points in each grid to form a multi-type 2D geometric feature operator graph, which comprises the following steps: height map operator, normal vector map operator, curvature map operator and dense map operator.
Further, the double-attention and multi-scale perception point cloud classification network in the step (3) is formed by sequentially connecting a multi-type geometric feature operator set feature extraction module, a double-attention feature enhancement module and a key feature point classification module; the multi-type geometric feature operator set feature extraction module extracts features of the multi-type 2D geometric feature operator graphs by using a convolutional neural network sharing weight; the double attention feature enhancement module consists of a self-attention module and a channel attention SK network module, the self-attention module uses a self-attention mechanism, obtains the weight of each data in the 2D geometric feature operator features through a convolution layer and a softmax activation function for the features of each type of 2D geometric feature operator, performs Hadamard product operation on the weight and the corresponding 2D geometric feature operator features to obtain the self-attention enhancement features of a 2D geometric feature operator graph, inputs the self-attention enhancement features into the channel attention SK network module, learns the contribution degree of the self-attention enhancement features of different types of geometric feature operator graphs to final classification, and outputs multi-operator fusion features; the key feature point classification module is formed by sequentially connecting a multi-scale sensing network module MSP, a convolution layer and a softmax activation function, and the multi-scale sensing network module MSP performs convolution, pooling and addition operations of different scales on multi-operator fusion features to obtain multi-scale fusion features; and then, carrying out convolution operation and softmax activation functions on the multi-scale fusion features to obtain the probability of each point of the three-dimensional point cloud data on a rivet region, a butt joint region and a non-feature region.
Further, the training process of the double attention and multi-scale perception point cloud classification network in the step (3) specifically comprises the following steps:
(a) inputting a multi-type 2D geometric feature operator set into a multi-type geometric feature operator set feature extraction module in a double-attention and multi-scale perception point cloud classification network to extract operator features;
(b) inputting the extracted operator characteristics into a double-attention characteristic enhancing module in a double-attention and multi-scale perception point cloud classification network, obtaining a weight matrix with the same dimension as the operator characteristics through convolution and softmax operation by using a self-attention learning module for each operator characteristic, carrying out Hadamard product operation on the weight matrix and the operator characteristics to obtain self-attention enhancing characteristics, splicing the self-attention enhancing characteristics of all operators to obtain splicing characteristics, inputting the splicing characteristics into an SK channel attention module, carrying out convolution and softmax operation on channel dimensions to obtain a weight matrix on the channel dimensions, and carrying out Hadamard product operation on the weight matrix and the splicing characteristics to obtain multi-operator fusion characteristics;
(c) inputting the fusion characteristics of the multiple operators into a key characteristic point classification module in a double attention and multi-scale perception point cloud classification network, predicting the probability of each three-dimensional point cloud data on a rivet area, a butt joint area and a non-characteristic area, and taking the area with the maximum probability value as the area where the three-dimensional point cloud data is located;
(d) and (c) repeating the steps (a) - (c) until the cross entropy loss function is converged, and finishing the training of the double attention and multi-scale perception point cloud classification network.
Further, the specific process of extracting the rivet profile features in the rivet area in the step (4) is as follows:
(4.1.1) randomly sampling point cloud subsets from all three-dimensional point cloud data of a rivet area, wherein the number of the sampled point cloud subsets is greater than that of rivets, fitting each point cloud subset into a circular model hypothesis by using a random consistency sampling method, and obtaining parameters of each circular model hypothesis to form a circular model hypothesis set; the assumed parameters of the circle model comprise a circle center and a radius;
(4.1.2) for each circular model hypothesis, retaining the circular model hypothesis when the radius thereof has an error from the actual rivet radius less than a threshold value;
(4.1.3) according to the spatial characteristics of rivet discrete distribution, comparing the distance between the centers of circles of each circle model hypothesis and other circle model hypotheses in the reserved circle model hypotheses, and clustering the circle model hypotheses with the distance being less than 2 times of the rivet radius into a rivet model hypothesis set;
(4.1.4) merging the three-dimensional point cloud data of the rivet areas of all the circle model hypotheses in each rivet model hypothesis set into a point cloud, deleting the point cloud from the rivet areas, fitting the point cloud into a circle by using a random consistency sampling method, and taking the fitted circle as a rivet contour feature;
(4.1.5) if the rivet contour features obtained in the step (4.4) are less than the total number of the rivets to be detected, repeating the steps (4.1.1) - (4.1.4) on the three-dimensional point cloud data which are not sampled in the rivet area.
Further, the process of extracting the butt seam contour features in the butt seam region in the step (4) specifically includes:
(4.2.1) randomly sampling point cloud subsets from all three-dimensional point cloud data of the butt joint area, wherein the number of the sampled point cloud subsets is greater than that of the butt joint area, and fitting each point cloud subset into a curve model hypothesis by using a random consistency sampling method to form a curve model hypothesis set;
(4.2.2) for each curve model hypothesis, when the distance between a point in the point cloud of the butt seam area and the fitted curve is smaller than a distance threshold value, and the number of points with the distance smaller than the distance threshold value is larger than a set point threshold value, keeping the curve model hypothesis;
(4.2.3) clustering a curve model hypothesis into a curve model hypothesis set when the distance from the retained curve model hypothesis to the closest point of other curve model hypotheses is smaller than a threshold value and the curve normal included angle from the retained curve model hypothesis to the closest point of other curve model hypotheses is smaller than a set threshold value;
(4.2.4) merging the three-dimensional point cloud data of the butt joint area of all the curve model hypotheses in each curve model hypothesis set into a point cloud, deleting the point cloud from the butt joint area, fitting the point cloud into two parallel curves by using a random consistency sampling method, and taking the two fitted parallel curves as a butt joint contour feature;
and (4.2.5) if the seam aligning profile features obtained in the step (4.2.4) are less than the total number of seams to be detected, repeating the steps (4.2.1) - (4.2.4) on the three-dimensional point cloud data which are not sampled in the rivet area.
Further, the rivet levelness analysis process in the step (5) specifically comprises the following steps: for any rivet contour feature, acquiring a rivet local neighborhood point cloud according to the geometric parameters of the rivet contour feature, fitting a plane by using a random consistency sampling method, calculating the directed distance between each point in the rivet contour and the plane, and judging that the rivet flushness is qualified when the maximum directed distance is smaller than a set flushness threshold value.
Further, the step (5) of analyzing the seam difference specifically includes: for any butt seam contour feature, according to the geometric parameters of the butt seam contour feature, calculating the directed distance D from each point of two curves of the butt seam contour to the other curve, obtaining the local neighborhood point cloud of the point, fitting the point into two approximately parallel planes by using a random consistency sampling method, calculating the directed distance of the two planes as the butt seam step difference F of the point, and then calculating the butt seam gap of the point
Calculating the butt seam step difference and the butt seam gap of all points of the butt seam contour characteristics, and taking the average value of the butt seam step differences of all points
Average value of gap between butt joint
When average butt joint step difference
And average butt gap
And when the seam is smaller than the respectively set threshold value, judging that the pair of seams is qualified.
Compared with the prior art, the invention has the following beneficial effects: the method for detecting the key features of the aircraft skin uses multiple kinds of geometric feature information to construct a geometric feature operator set as input, uses a double attention module to learn the contribution degree of final classification of each operator feature, deeply fuses multiple kinds of feature information, accurately extracts rivets and butt seams with unobvious features on the large aircraft skin, uses a multi-level model fitting method to improve the stability and accuracy of rivet contour and butt seam contour extraction, realizes accurate automatic extraction of small-scale key feature rivets and tiny butt seams on the surface of the aircraft skin, and improves the aircraft assembly quality.
Drawings
FIG. 1 is a flow chart of the method for detecting key features of an aircraft skin based on a multi-type geometric feature operator according to the present invention;
FIG. 2 is a schematic diagram illustrating the construction of a multi-type geometric feature operator according to the present invention;
FIG. 3 is a schematic structural diagram of a double-attention and multi-scale perception point cloud classification network according to the present invention;
FIG. 4 is a schematic illustration of a rivet region and a butt seam region on an aircraft skin;
FIG. 5 is a schematic representation of the results of detection of key features of an aircraft skin by the present invention.
Detailed Description
The technical solution of the present invention is further described in detail below with reference to the accompanying drawings.
Fig. 1 is a flow chart of the method for detecting aircraft skin key features based on multi-type geometric feature operators, and the method for detecting aircraft skin key features comprises the following steps:
(1) acquiring three-dimensional point cloud data of a large aircraft skin; the method specifically comprises the following substeps:
(1.1) as the aircraft skin has larger dimension, all three-dimensional point cloud data cannot be acquired at one time during scanning, and complete data needs to be acquired by adopting a mode of acquiring for multiple times and splicing, specifically, a Leica ATS960 absolute tracker is used for acquiring the aircraft skin three-dimensional point cloud data for multiple times, the three-dimensional point cloud data acquired for multiple times cover a key feature region to be detected on the surface of the aircraft skin, and the overlapping rate of the three-dimensional point cloud data acquired for two adjacent times is more than 30%;
and (1.2) splicing the three-dimensional point cloud data acquired for multiple times into complete large aircraft skin three-dimensional point cloud data by using an iterative closest point algorithm.
(2) Based on multiple geometric attributes of the three-dimensional point cloud data, an operator map set containing multiple types of 2D geometric features is constructed for each point in the three-dimensional point cloud data so as to obtain multiple types of information and improve the identification precision of a rivet area and a butt joint area; as shown in fig. 2, the method specifically includes the following sub-steps:
(2.1) for each point p in the three-dimensional point cloud data, calculating the distance from all points in the three-dimensional point cloud data to the point, taking the nearest k points as a neighborhood point set, wherein the value of k is 32, and calculating the normal vectors of the point p and all points in the neighborhood point set by using a principal component analysis method; the point p is used as an origin, the normal vector of the point p is used as a Z axis, a local coordinate system is established, the influence of the position and the direction of a point cloud in a three-dimensional space is avoided, and the complexity of a learning task is reduced; projecting the point p and the neighborhood point set thereof to an XY-plane of a local coordinate system, constructing a minimum square bounding box with sides parallel to an X axis and a Y axis, cutting the bounding box into n multiplied by n planar grids, and converting the unstructured three-dimensional neighborhood point set into a structured 2D grid through the operation, so that the neighborhood point set can be processed by a deep learning module in the image field;
(2.2) calculating the projection height, normal vector, Gaussian curvature, density and other geometric information of each point in the neighborhood point set, dividing the neighborhood point set into different grids according to the projection position of each point in the neighborhood point set on the plane network, taking the average height, average normal vector, average Gaussian curvature and average density of all points in each grid as multi-type geometric information of the grid, forming a matrix by each type of geometric information of the whole plane grid, and taking all the matrices as multi-type 2D geometric feature calculation graphs, wherein the method comprises the following steps: height map operator, normal vector operator, curvature map operator and dense map operator. By extracting the geometric information of various types of information, the difference characteristics of the rivet area, the butt joint area and the non-characteristic area can be better sensed and distinguished, and the identification precision is improved.
(3) Inputting the multi-type 2D geometric feature operator image set constructed in the step (2) into a trained double-attention and multi-scale perception point cloud classification network, learning the features of rivets and butt joints in a plurality of operators, and classifying the three-dimensional point cloud data into a rivet region, a butt joint region and a non-feature region.
Referring to fig. 3, the double-attention and multi-scale perception point cloud classification network of the invention is formed by sequentially connecting a multi-type geometric feature operator set feature extraction module, a double-attention feature enhancement module and a key feature point classification module; the multi-type geometric feature operator set feature extraction module extracts features of the multi-type 2D geometric feature operator graphs by using a convolutional neural network sharing weight, the multi-type geometric feature operator set feature extraction module comprises 4 convolutional layers, and a maximum pooling layer is arranged behind each 2 convolutional layers so as to reduce the dimensionality of input data and reduce the computation complexity; the double attention feature enhancement module consists of a self attention module and a channel attention SK network module, the self attention module uses a self attention mechanism, obtains the weight of each data in 2D geometric feature operator features through a convolutional layer and a softmax activation function for the features of each type of 2D geometric feature operator, performs Hadamard product operation on the data and the corresponding 2D geometric feature operator features to obtain the self attention enhancement features of a 2D geometric feature operator graph, inputs the self attention enhancement features into the channel attention SK network module, learns the contribution degree of the self attention enhancement features of different types of geometric feature operator graphs to final classification, and outputs multi-operator fusion features, the double attention mechanism can enable the network to pay more attention to help feature region learning important information, and restrain some irrelevant detailed information; the key feature point classification module is formed by sequentially connecting a multi-scale sensing network module MSP, a convolution layer and a softmax activation function, and the multi-scale sensing network module MSP performs convolution, pooling and addition operations of different scales on the multi-operator fusion feature to obtain the multi-scale fusion feature; and then, carrying out convolution operation and softmax activation functions on the multi-scale fusion features to obtain the probability of each point of the three-dimensional point cloud data on a rivet region, a butt joint region and a non-feature region. According to the invention, the double-attention and multi-scale perception point cloud classification network can accurately detect rivets and butt joints with unobvious features in the aircraft skin point cloud data.
The training process of the double-attention and multi-scale perception point cloud classification network specifically comprises the following steps:
(a) inputting a multi-type 2D geometric feature operator set into a multi-type geometric feature operator set feature extraction module in a double-attention and multi-scale perception point cloud classification network to extract operator features;
(b) inputting the extracted operator characteristics into a double attention characteristic enhancing module in a double attention and multi-scale perception point cloud classification network, obtaining a weight matrix with the same dimension size as the operator characteristics through convolution and softmax operation by using a self-attention learning module for each operator characteristic, carrying out Hadamard product operation on the weight matrix and the operator characteristics to obtain self-attention enhancing characteristics, splicing the self-attention enhancing characteristics of all operators to obtain splicing characteristics, inputting the splicing characteristics into an SK channel attention module, carrying out convolution and softmax operation on channel dimensions to obtain a weight matrix on channel dimensions, and carrying out Hadamard product operation on the weight matrix and the splicing characteristics to obtain multi-operator fusion characteristics;
(c) inputting the fusion characteristics of the multiple operators into a key characteristic point classification module in a double attention and multi-scale perception point cloud classification network, predicting the probability of each three-dimensional point cloud data on a rivet area, a butt joint area and a non-characteristic area, and taking the area with the maximum probability value as the area where the three-dimensional point cloud data is located;
(d) and (c) repeating the steps (a) - (c) until the cross entropy loss function is converged, and finishing the training of the double attention and multi-scale perception point cloud classification network.
The cross entropy loss function L in the present invention is:
wherein k represents the total number of classification regions; i represents a classification region index; y isiIndicating the accuracy of the classification region, when correct, y i1 is ═ 1; otherwise, yi＝0；piRepresenting the probability of predicting the ith category.
(4) The method adopts a multilevel model fitting algorithm to extract the rivet profile characteristics in the rivet area and the butt seam profile characteristics in the butt seam area, and can obtain a more stable profile characteristic extraction result; specifically, the specific process of extracting the rivet profile features in the rivet area is as follows:
(4.1.1) randomly sampling point cloud subsets from all three-dimensional point cloud data of a rivet area, wherein the number of the sampled point cloud subsets is greater than that of rivets, fitting each point cloud subset into a circular model hypothesis by using a random consistency sampling method, and obtaining parameters of each circular model hypothesis to form a circular model hypothesis set; the assumed parameters of the circular model comprise a circle center and a radius;
(4.1.2) for each circular model hypothesis, when the error between the radius and the actual rivet radius is smaller than a threshold value, retaining the circular model hypothesis, wherein the threshold value can be 0.3 times the actual rivet radius;
(4.1.3) according to the spatial characteristic of discrete rivet distribution, the rivet distance is far larger than the rivet radius, the distances between the circle centers of each circle model hypothesis and other circle model hypotheses in the reserved circle model hypotheses are compared, and the circle model hypotheses with the distances being smaller than 2 times of the rivet radius are clustered into a rivet model hypothesis set;
(4.1.4) combining the three-dimensional point cloud data of the rivet areas of all the circular model hypotheses of each rivet model hypothesis set into a point cloud, deleting the point cloud from the rivet areas, fitting the point cloud into a circle by using a random consistency sampling method, and taking the fitted circle as a rivet outline characteristic;
(4.1.5) if the rivet contour features obtained in the step (4.1.4) are less than the total number of the rivets to be detected, repeating the steps (4.1.1) - (4.1.4) on the three-dimensional point cloud data which are not sampled in the rivet area.
The process of extracting the butt seam contour features in the butt seam area specifically comprises the following steps:
(4.2.1) randomly sampling point cloud subsets from all three-dimensional point cloud data of the butt joint area, wherein the number of the sampled point cloud subsets is larger than that of the butt joint area, and fitting each point cloud subset into a curve model hypothesis by using a random consistency sampling method to form a curve model hypothesis set;
(4.2.2) for each curve model hypothesis, when the distance between a point in the point cloud of the butt seam area and the fitted curve is smaller than a distance threshold value, and the number of points with the distance smaller than the distance threshold value is larger than a set point threshold value, keeping the curve model hypothesis; in the invention, the distance threshold value is 1 cm; the point number threshold is determined by the length L of the curve model hypothesis and the average closest point distance d, the point number threshold is generally 0.5L/d, after all the points with the distance less than the distance threshold are obtained, the distance between two points with the farthest distance in the points is taken as the length L of the curve model hypothesis, and the average closest point distance d is the average value of the distance from each point to the closest point in the whole point cloud data;
(4.2.3) clustering a curve model hypothesis into a curve model hypothesis set when the distance from the retained curve model hypothesis to the closest point of other curve model hypotheses is smaller than a threshold value and the curve normal included angle from the retained curve model hypothesis to the closest point of other curve model hypotheses is smaller than a set threshold value; the threshold value of the normal phase angle is set to be 10 degrees;
(4.2.4) merging the three-dimensional point cloud data of the butt joint area of all the curve model hypotheses in each curve model hypothesis set into a point cloud, deleting the point cloud from the butt joint area, fitting the point cloud into two parallel curves by using a random consistency sampling method, and taking the two fitted parallel curves as a butt joint contour feature;
(4.2.5) if the butt seam contour features obtained in the step (4.2.4) are less than the total number of butt seams to be detected, repeating the steps (4.2.1) - (4.2.4) on the three-dimensional point cloud data which are not sampled in the rivet area.
(5) Carrying out rivet levelness analysis and butt joint step analysis based on the extracted rivet profile characteristics and butt joint profile characteristics, specifically, the process of rivet levelness analysis specifically is: for any rivet contour feature, acquiring a rivet local neighborhood point cloud according to the geometric parameters of the rivet contour feature, fitting a plane by using a random consistency sampling method, calculating the directed distance between each point in the rivet contour and the plane, judging that the rivet flushness is qualified when the maximum directed distance is smaller than a set flushness threshold value, and determining that the specific flushness threshold value is determined by a detection qualified standard, wherein the flushness threshold value is set to be 1 mm in one technical scheme of the invention. The process of analyzing the gap step specifically comprises the following steps: for any butt seam contour feature, calculating the directed distance D from each point of two curves of the butt seam contour to the other curve according to the geometric parameters of the butt seam contour feature to obtainTaking local neighborhood point cloud of the point, fitting the point cloud into two approximately parallel planes by using a random consistency sampling method, calculating the directed distance of the two planes as the butt joint step difference F of the point, and then calculating the butt joint gap of the point
Calculating the butt seam step difference and the butt seam gap of all points of the butt seam contour characteristics, and taking the average value of the butt seam step differences of all points
Average value of gap between butt joint
Mean step of butt joint
And average butt gap
And when the seam is smaller than the respectively set threshold value, judging that the pair of seams is qualified. The specific threshold value of the step difference of the butt joint and the threshold value of the gap of the butt joint are determined by the qualified detection standard, and in one technical scheme of the invention, the threshold value of the step difference of the butt joint is 1 mm and the threshold value of the gap of the butt joint is 3 mm.
FIG. 4 is a schematic illustration of a rivet region and a butt seam region on an aircraft skin; as shown in FIG. 5, the method for detecting the key features of the aircraft skin based on the multi-type geometric feature operator is used for detecting the rivet and butt joint features, and through comparison between FIG. 4 and FIG. 5, the method for detecting the key features of the aircraft skin can detect all the rivet and butt joint features on the aircraft skin.
The above is only a preferred embodiment of the present invention, and the scope of the present invention is not limited to the above embodiment, and any technical solutions that fall under the spirit of the present invention fall within the scope of the present invention. It should be noted that modifications and embellishments within the scope of the invention may be made by those skilled in the art without departing from the principle of the invention.
Claims (9)
1. A method for detecting aircraft skin key features based on multi-type geometric feature operators is characterized by comprising the following steps:
(1) acquiring three-dimensional point cloud data of a large aircraft skin;
(2) constructing an operator atlas set containing multi-type 2D geometric features for each point in the three-dimensional point cloud data based on multiple geometric attributes of the three-dimensional point cloud data;
(3) inputting the multi-type 2D geometric feature operator image set constructed in the step (2) into a trained double-attention and multi-scale perception point cloud classification network, learning the features of rivets and butt joints in a plurality of operators, and classifying three-dimensional point cloud data into a rivet region, a butt joint region and a non-feature region;
(4) extracting rivet contour features in a rivet area and butt seam contour features in a butt seam area by adopting a multilevel model fitting algorithm;
(5) and performing rivet levelness analysis and butt seam step analysis based on the extracted rivet profile characteristics and butt seam profile characteristics.
2. The method for detecting the key features of the aircraft skin based on the multi-type geometric feature operator as claimed in claim 1, wherein the step (1) comprises the following sub-steps:
(1.1) repeatedly acquiring three-dimensional point cloud data of the skin of the airplane by using a Leica ATS960 absolute tracker, wherein the repeatedly acquired three-dimensional point cloud data cover a key feature region to be detected on the surface of the skin of the airplane;
and (1.2) splicing the three-dimensional point cloud data acquired for multiple times into complete large aircraft skin three-dimensional point cloud data by using an iterative closest point algorithm.
3. The method for detecting the key features of the aircraft skin based on the multi-type geometric feature operator as claimed in claim 1, wherein the step (2) comprises the following sub-steps:
(2.1) for each point p in the three-dimensional point cloud data, calculating the distance from all points in the three-dimensional point cloud data to the point, taking the nearest k points as a neighborhood point set, and calculating the normal vectors of the point p and all points in the neighborhood point set by using a principal component analysis method; establishing a local coordinate system by taking the point p as an original point and taking a normal vector of the point p as a Z axis, projecting the point p and a neighborhood point set thereof to an XY-plane of the local coordinate system, constructing a minimum square bounding box with sides parallel to an X axis and a Y axis, and cutting the bounding box into planar grids of n multiplied by n;
(2.2) calculating the projection height, normal vector, Gaussian curvature, average curvature and geometric information of the dense density coding neighborhood points of the neighborhood points in each grid to form a multi-type 2D geometric feature operator graph, which comprises the following steps: height map operator, normal vector map operator, curvature map operator and dense map operator.
4. The method for detecting the key features of the aircraft skin based on the multi-type geometric feature operators according to claim 1, wherein the double-attention and multi-scale perception point cloud classification network in the step (3) is formed by sequentially connecting a multi-type geometric feature operator set feature extraction module, a double-attention feature enhancement module and a key feature point classification module; the multi-type geometric feature operator set feature extraction module extracts features of the multi-type 2D geometric feature operator graphs by using a convolutional neural network sharing weight; the double attention feature enhancement module consists of a self-attention module and a channel attention SK network module, the self-attention module uses a self-attention mechanism, obtains the weight of each data in the 2D geometric feature operator features through a convolution layer and a softmax activation function for the features of each type of 2D geometric feature operator, performs Hadamard product operation on the weight and the corresponding 2D geometric feature operator features to obtain the self-attention enhancement features of a 2D geometric feature operator graph, inputs the self-attention enhancement features into the channel attention SK network module, learns the contribution degree of the self-attention enhancement features of different types of geometric feature operator graphs to final classification, and outputs multi-operator fusion features; the key feature point classification module is formed by sequentially connecting a multi-scale sensing network module MSP, a convolution layer and a softmax activation function, and the multi-scale sensing network module MSP performs convolution, pooling and addition operations of different scales on multi-operator fusion features to obtain multi-scale fusion features; and then, carrying out convolution operation and softmax activation functions on the multi-scale fusion features to obtain the probability of each point of the three-dimensional point cloud data on a rivet region, a butt joint region and a non-feature region.
5. The method for detecting key features of aircraft skin based on multi-type geometric feature operators as claimed in claim 1, wherein the training process of the double attention and multi-scale perception point cloud classification network in the step (3) is specifically as follows:
(a) inputting a multi-type 2D geometric feature operator set into a multi-type geometric feature operator set feature extraction module in a double-attention and multi-scale perception point cloud classification network to extract operator features;
(b) inputting the extracted operator characteristics into a double attention characteristic enhancing module in a double attention and multi-scale perception point cloud classification network, obtaining a weight matrix with the same dimension size as the operator characteristics through convolution and softmax operation by using a self-attention learning module for each operator characteristic, carrying out Hadamard product operation on the weight matrix and the operator characteristics to obtain self-attention enhancing characteristics, splicing the self-attention enhancing characteristics of all operators to obtain splicing characteristics, inputting the splicing characteristics into an SK channel attention module, carrying out convolution and softmax operation on channel dimensions to obtain a weight matrix on channel dimensions, and carrying out Hadamard product operation on the weight matrix and the splicing characteristics to obtain multi-operator fusion characteristics;
(c) inputting the fusion characteristics of the multiple operators into a key characteristic point classification module in a double attention and multi-scale perception point cloud classification network, predicting the probability of each three-dimensional point cloud data on a rivet area, a butt joint area and a non-characteristic area, and taking the area with the maximum probability value as the area where the three-dimensional point cloud data is located;
(d) and (c) repeating the steps (a) - (c) until the cross entropy loss function is converged, and finishing the training of the double attention and multi-scale perception point cloud classification network.
6. The method for detecting the key features of the aircraft skin based on the multi-type geometric feature operator according to claim 1, wherein the specific process for extracting the rivet contour features in the rivet region in the step (4) is as follows:
(4.1.1) randomly sampling point cloud subsets from all three-dimensional point cloud data of a rivet area, wherein the number of the sampled point cloud subsets is greater than that of rivets, fitting each point cloud subset into a circular model hypothesis by using a random consistency sampling method, and obtaining parameters of each circular model hypothesis to form a circular model hypothesis set; the assumed parameters of the circle model comprise a circle center and a radius;
(4.1.2) for each circular model hypothesis, retaining the circular model hypothesis when the radius thereof has an error from the actual rivet radius less than a threshold value;
(4.1.3) according to the spatial characteristics of rivet discrete distribution, comparing the distance between the centers of circles of each circle model hypothesis and other circle model hypotheses in the reserved circle model hypotheses, and clustering the circle model hypotheses with the distance being less than 2 times of the rivet radius into a rivet model hypothesis set;
(4.1.4) merging the three-dimensional point cloud data of the rivet areas of all the circle model hypotheses in each rivet model hypothesis set into a point cloud, deleting the point cloud from the rivet areas, fitting the point cloud into a circle by using a random consistency sampling method, and taking the fitted circle as a rivet contour feature;
(4.1.5) if the rivet contour features obtained in the step (4.1.4) are less than the total number of the rivets to be detected, repeating the steps (4.1.1) - (4.1.4) on the three-dimensional point cloud data which are not sampled in the rivet area.
7. The method for detecting the key features of the aircraft skin based on the multi-type geometric feature operator according to claim 1, wherein the step (4) of extracting the butt seam contour features in the butt seam region specifically comprises the following steps:
(4.2.1) randomly sampling point cloud subsets from all three-dimensional point cloud data of the butt joint area, wherein the number of the sampled point cloud subsets is greater than that of the butt joint area, and fitting each point cloud subset into a curve model hypothesis by using a random consistency sampling method to form a curve model hypothesis set;
(4.2.2) for each curve model hypothesis, when the distance between a point in the point cloud of the butt seam area and the fitted curve is smaller than a distance threshold value, and the number of points with the distance smaller than the distance threshold value is larger than a set point threshold value, keeping the curve model hypothesis;
(4.2.3) clustering a curve model hypothesis into a curve model hypothesis set when the distance from the retained curve model hypothesis to the closest point of other curve model hypotheses is less than a threshold value and the curve normal included angle from the retained curve model hypothesis to the closest point of other curve model hypotheses is less than a set threshold value;
(4.2.4) merging the three-dimensional point cloud data of the butt joint area of all the curve model hypotheses in each curve model hypothesis set into a point cloud, deleting the point cloud from the butt joint area, fitting the point cloud into two parallel curves by using a random consistency sampling method, and taking the two fitted parallel curves as a butt joint contour feature;
and (4.2.5) if the seam aligning profile features obtained in the step (4.2.4) are less than the total number of seams to be detected, repeating the steps (4.2.1) - (4.2.4) on the three-dimensional point cloud data which are not sampled in the rivet area.
8. The method for detecting the key features of the aircraft skin based on the multi-type geometric feature operator according to claim 1, wherein the rivet levelness analysis in the step (5) specifically comprises the following steps: and for any rivet contour feature, acquiring a rivet local neighborhood point cloud according to the geometric parameters of the rivet contour feature, fitting a plane by using a random consistency sampling method, calculating the directed distance between each point in the rivet contour and the plane, and judging that the rivet levelness is qualified when the maximum directed distance is smaller than a set levelness threshold value.
9. The method for detecting the key features of the skin of an aircraft based on multi-type geometric feature operators as claimed in claim 1Characterized in that the process of analyzing the gap jump in the step (5) specifically comprises the following steps: for any butt seam contour feature, according to the geometric parameters of the butt seam contour feature, calculating the directed distance D from each point of two curves of the butt seam contour to the other curve, obtaining the local neighborhood point cloud of the point, fitting the point into two approximately parallel planes by using a random consistency sampling method, calculating the directed distance of the two planes as the butt seam step difference F of the point, and then calculating the butt seam gap of the point
Calculating the butt seam step difference and the butt seam gap of all points of the butt seam contour characteristics, and taking the average value of the butt seam step differences of all points
Average value of gap between butt joint
Mean step of butt joint
And average butt gap
And when the seam is smaller than the respectively set threshold value, judging that the pair of seams is qualified.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210268285.1A CN114626470B (en) | 2022-03-18 | 2022-03-18 | Aircraft skin key feature detection method based on multi-type geometric feature operator |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210268285.1A CN114626470B (en) | 2022-03-18 | 2022-03-18 | Aircraft skin key feature detection method based on multi-type geometric feature operator |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114626470A true CN114626470A (en) | 2022-06-14 |
| CN114626470B CN114626470B (en) | 2024-02-02 |
Family
ID=81901862
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210268285.1A Active CN114626470B (en) | 2022-03-18 | 2022-03-18 | Aircraft skin key feature detection method based on multi-type geometric feature operator |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114626470B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115965788A (en) * | 2023-01-12 | 2023-04-14 | 黑龙江工程学院 | Point cloud semantic segmentation method based on multi-view image structural feature attention convolution |
| CN115982864A (en) * | 2023-03-21 | 2023-04-18 | 南京航空航天大学 | Reconstruction method for large composite material component assembly coordination boundary characteristics |
| CN116778260A (en) * | 2023-08-17 | 2023-09-19 | 南京航空航天大学 | Aviation rivet flushness detection method, device and system based on AdaBoost ensemble learning |
| CN117494534A (en) * | 2024-01-03 | 2024-02-02 | 南京航空航天大学 | Aircraft skin repairing method based on point cloud and finite element analysis |
| CN117788472A (en) * | 2024-02-27 | 2024-03-29 | 南京航空航天大学 | Method for judging corrosion degree of rivet on surface of aircraft skin based on DBSCAN algorithm |
| CN118049938A (en) * | 2024-04-15 | 2024-05-17 | 西安第六镜网络科技有限公司 | Rail measuring method, device, electronic equipment, readable storage medium and system |
Citations (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108151660A (en) * | 2017-12-29 | 2018-06-12 | 西北工业大学 | The measurement equipment of a kind of aircraft components butt-joint clearance and scale, method and system |
| CN108408080A (en) * | 2018-03-29 | 2018-08-17 | 南昌航空大学 | A kind of aircraft wing body Butt Assembling device, method and system |
| CN109598306A (en) * | 2018-12-06 | 2019-04-09 | 西安电子科技大学 | Hyperspectral image classification method based on SRCM and convolutional neural networks |
| CN109596059A (en) * | 2019-01-07 | 2019-04-09 | 南京航空航天大学 | A kind of aircraft skin gap based on parallel lines structure light and scale measurement method |
| CN109886273A (en) * | 2019-02-26 | 2019-06-14 | 四川大学华西医院 | A kind of CMR classification of image segmentation system |
| CN110276445A (en) * | 2019-06-19 | 2019-09-24 | 长安大学 | Domestic communication label category method based on Inception convolution module |
| CN111028220A (en) * | 2019-12-11 | 2020-04-17 | 南京航空航天大学 | Automatic detection method for levelness of point cloud rivet |
| CN111028221A (en) * | 2019-12-11 | 2020-04-17 | 南京航空航天大学 | Airplane skin butt-joint measurement method based on linear feature detection |
| CN111080629A (en) * | 2019-12-20 | 2020-04-28 | 河北工业大学 | Method for detecting image splicing tampering |
| CN111524154A (en) * | 2020-04-21 | 2020-08-11 | 南京航空航天大学 | Image-based tunnel segment automatic segmentation method |
| CN111524129A (en) * | 2020-04-29 | 2020-08-11 | 南京航空航天大学 | Aircraft skin butt joint gap calculation method based on end face extraction |
| CN111582126A (en) * | 2020-04-30 | 2020-08-25 | 浙江工商大学 | Pedestrian re-identification method based on multi-scale pedestrian contour segmentation fusion |
| KR20200122897A (en) * | 2019-04-19 | 2020-10-28 | 서울대학교산학협력단 | System and method for monitoring the forest gaps using airborne lidar datasets |
| CN111862054A (en) * | 2020-07-23 | 2020-10-30 | 南京航空航天大学 | Rivet contour point cloud extraction method |
| CN111967480A (en) * | 2020-09-07 | 2020-11-20 | 上海海事大学 | Multi-scale self-attention target detection method based on weight sharing |
| CN112053361A (en) * | 2020-10-15 | 2020-12-08 | 南京航空航天大学 | Aircraft skin butt joint detection method based on large-scale point cloud |
| CN112053426A (en) * | 2020-10-15 | 2020-12-08 | 南京航空航天大学 | Deep learning-based large-scale three-dimensional rivet point cloud extraction method |
| CN112446591A (en) * | 2020-11-06 | 2021-03-05 | 太原科技大学 | Evaluation system for student comprehensive capacity evaluation and zero sample evaluation method |
| CN113205055A (en) * | 2021-05-11 | 2021-08-03 | 北京知见生命科技有限公司 | Fungus microscopic image classification method and system based on multi-scale attention mechanism |
| CN113378791A (en) * | 2021-07-09 | 2021-09-10 | 合肥工业大学 | Cervical cell classification method based on double-attention mechanism and multi-scale feature fusion |
| CN113935977A (en) * | 2021-10-22 | 2022-01-14 | 河北工业大学 | Solar cell panel defect generation method based on generation countermeasure network |
-
2022
- 2022-03-18 CN CN202210268285.1A patent/CN114626470B/en active Active
Patent Citations (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108151660A (en) * | 2017-12-29 | 2018-06-12 | 西北工业大学 | The measurement equipment of a kind of aircraft components butt-joint clearance and scale, method and system |
| CN108408080A (en) * | 2018-03-29 | 2018-08-17 | 南昌航空大学 | A kind of aircraft wing body Butt Assembling device, method and system |
| CN109598306A (en) * | 2018-12-06 | 2019-04-09 | 西安电子科技大学 | Hyperspectral image classification method based on SRCM and convolutional neural networks |
| CN109596059A (en) * | 2019-01-07 | 2019-04-09 | 南京航空航天大学 | A kind of aircraft skin gap based on parallel lines structure light and scale measurement method |
| CN109886273A (en) * | 2019-02-26 | 2019-06-14 | 四川大学华西医院 | A kind of CMR classification of image segmentation system |
| KR20200122897A (en) * | 2019-04-19 | 2020-10-28 | 서울대학교산학협력단 | System and method for monitoring the forest gaps using airborne lidar datasets |
| CN110276445A (en) * | 2019-06-19 | 2019-09-24 | 长安大学 | Domestic communication label category method based on Inception convolution module |
| CN111028220A (en) * | 2019-12-11 | 2020-04-17 | 南京航空航天大学 | Automatic detection method for levelness of point cloud rivet |
| CN111028221A (en) * | 2019-12-11 | 2020-04-17 | 南京航空航天大学 | Airplane skin butt-joint measurement method based on linear feature detection |
| CN111080629A (en) * | 2019-12-20 | 2020-04-28 | 河北工业大学 | Method for detecting image splicing tampering |
| CN111524154A (en) * | 2020-04-21 | 2020-08-11 | 南京航空航天大学 | Image-based tunnel segment automatic segmentation method |
| CN111524129A (en) * | 2020-04-29 | 2020-08-11 | 南京航空航天大学 | Aircraft skin butt joint gap calculation method based on end face extraction |
| CN111582126A (en) * | 2020-04-30 | 2020-08-25 | 浙江工商大学 | Pedestrian re-identification method based on multi-scale pedestrian contour segmentation fusion |
| CN111862054A (en) * | 2020-07-23 | 2020-10-30 | 南京航空航天大学 | Rivet contour point cloud extraction method |
| CN111967480A (en) * | 2020-09-07 | 2020-11-20 | 上海海事大学 | Multi-scale self-attention target detection method based on weight sharing |
| CN112053361A (en) * | 2020-10-15 | 2020-12-08 | 南京航空航天大学 | Aircraft skin butt joint detection method based on large-scale point cloud |
| CN112053426A (en) * | 2020-10-15 | 2020-12-08 | 南京航空航天大学 | Deep learning-based large-scale three-dimensional rivet point cloud extraction method |
| CN112446591A (en) * | 2020-11-06 | 2021-03-05 | 太原科技大学 | Evaluation system for student comprehensive capacity evaluation and zero sample evaluation method |
| CN113205055A (en) * | 2021-05-11 | 2021-08-03 | 北京知见生命科技有限公司 | Fungus microscopic image classification method and system based on multi-scale attention mechanism |
| CN113378791A (en) * | 2021-07-09 | 2021-09-10 | 合肥工业大学 | Cervical cell classification method based on double-attention mechanism and multi-scale feature fusion |
| CN113935977A (en) * | 2021-10-22 | 2022-01-14 | 河北工业大学 | Solar cell panel defect generation method based on generation countermeasure network |
Non-Patent Citations (4)
| Title |
|---|
| QIAN XIE ET AL: "RRCNet: Rivet Region Classification Network for Rivet Flush Measurement Based on 3-D Point Cloud", 《 IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT》, pages 5 - 6 * |
| 丁祖娇: "面向结构光对缝测量的光条细化方法", 《中国优秀硕士学位论文全文数据库信息科技辑》, vol. 2018, no. 3, pages 1 - 5 * |
| 张波等: "基于SVM的三维对缝点云间隙阶差提取方法", 《航空制造技术》, vol. 63, no. 7, pages 47 - 54 * |
| 王德重等: "基于双目多线结构光的铆钉齐平度测量方法", 《航空制造技术》, vol. 64, pages 57 - 65 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115965788A (en) * | 2023-01-12 | 2023-04-14 | 黑龙江工程学院 | Point cloud semantic segmentation method based on multi-view image structural feature attention convolution |
| CN115982864A (en) * | 2023-03-21 | 2023-04-18 | 南京航空航天大学 | Reconstruction method for large composite material component assembly coordination boundary characteristics |
| CN116778260A (en) * | 2023-08-17 | 2023-09-19 | 南京航空航天大学 | Aviation rivet flushness detection method, device and system based on AdaBoost ensemble learning |
| CN116778260B (en) * | 2023-08-17 | 2023-11-17 | 南京航空航天大学 | Aviation rivet flushness detection method, device and system based on AdaBoost ensemble learning |
| CN117494534A (en) * | 2024-01-03 | 2024-02-02 | 南京航空航天大学 | Aircraft skin repairing method based on point cloud and finite element analysis |
| CN117494534B (en) * | 2024-01-03 | 2024-03-22 | 南京航空航天大学 | Aircraft skin repairing method based on point cloud and finite element analysis |
| CN117788472A (en) * | 2024-02-27 | 2024-03-29 | 南京航空航天大学 | Method for judging corrosion degree of rivet on surface of aircraft skin based on DBSCAN algorithm |
| CN117788472B (en) * | 2024-02-27 | 2024-05-14 | 南京航空航天大学 | Method for judging corrosion degree of rivet on surface of aircraft skin based on DBSCAN algorithm |
| CN118049938A (en) * | 2024-04-15 | 2024-05-17 | 西安第六镜网络科技有限公司 | Rail measuring method, device, electronic equipment, readable storage medium and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114626470B (en) | 2024-02-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114626470B (en) | Aircraft skin key feature detection method based on multi-type geometric feature operator | |
| Movaghati et al. | Road extraction from satellite images using particle filtering and extended Kalman filtering | |
| Chen et al. | Principal axes descriptor for automated construction-equipment classification from point clouds | |
| CN109740588B (en) | X-ray picture contraband positioning method based on weak supervision and deep response redistribution | |
| Landry et al. | CELLO-3D: Estimating the Covariance of ICP in the Real World | |
| CN111080627A (en) | 2D +3D large airplane appearance defect detection and analysis method based on deep learning | |
| CN112529015A (en) | Three-dimensional point cloud processing method, device and equipment based on geometric unwrapping | |
| CN105957076A (en) | Clustering based point cloud segmentation method and system | |
| CN113343740B (en) | Table detection method, device, equipment and storage medium | |
| CN109284779A (en) | Object detection method based on deep full convolution network | |
| Xu et al. | Power line extraction from mobile LiDAR point clouds | |
| CN105809113A (en) | Three-dimensional human face identification method and data processing apparatus using the same | |
| CN111597943B (en) | Table structure identification method based on graph neural network | |
| Dai et al. | Aircraft seam feature extraction from 3D raw point cloud via hierarchical multi-structure fitting | |
| CN117274774A (en) | Yolov 7-based X-ray security inspection image dangerous goods detection algorithm | |
| Dey et al. | A robust performance evaluation metric for extracted building boundaries from remote sensing data | |
| CN117928385A (en) | Engineering construction intelligent measurement method based on remote unmanned aerial vehicle and sensor | |
| CN113378694A (en) | Method and device for generating target detection and positioning system and target detection and positioning | |
| Wu et al. | A Systematic Point Cloud Edge Detection Framework for Automatic Aircraft Skin Milling | |
| CN115761502A (en) | SAR image change detection method based on hybrid convolution | |
| CN107220987A (en) | A kind of building roof Fast Edge Detection method based on principal component analysis | |
| Yang et al. | Recurrence network-based 3D geometry representation learning for quality control in additive manufacturing of metamaterials | |
| CN113674227B (en) | Layer spacing detection method for ion thruster grid assembly | |
| CN113673534B (en) | RGB-D image fruit detection method based on FASTER RCNN | |
| Kaleci et al. | Plane segmentation of point cloud data using split and merge based method |
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 |