Search

CN-116108552-B - Multi-part combination modeling and identification method, equipment and medium

CN116108552BCN 116108552 BCN116108552 BCN 116108552BCN-116108552-B

Abstract

The invention discloses a multi-part combination modeling and identifying method which comprises the following steps of S1, acquiring a three-dimensional model of a sample part, constructing a cloud model base of the sample part according to the three-dimensional model of the sample part, S2, acquiring image data of all parts to be identified, matching the image data with the three-dimensional model of the sample part in the cloud model base, identifying all the parts to be identified, S3, selecting a plurality of associated parts from all the parts to be identified, wherein the plurality of associated parts can be combined into one or a plurality of integral parts to obtain a three-dimensional model group corresponding to the plurality of associated parts, S4, combining the three-dimensional model groups corresponding to the plurality of associated parts to obtain a three-dimensional combination model, and the three-dimensional combination model is used for identifying and comparing the integral parts formed by the plurality of associated parts.

Inventors

  • YUAN LIANG

Assignees

  • 四川物通科技有限公司

Dates

Publication Date
20260508
Application Date
20220815

Claims (6)

  1. 1. A multi-part combination modeling identification method, characterized in that the method comprises the following steps: s1, acquiring a three-dimensional model of a sample part, and constructing a cloud model library of the sample part according to the three-dimensional model of the sample part, wherein the method for acquiring the three-dimensional model of the sample part comprises the following steps: sa, acquiring a sample part image set, and performing digital processing on the sample part image set to obtain a sample part image; Sb, performing cluster analysis according to the image atlas and pose information of the sample part to obtain a clustered image; Sc, extracting features of the sample part image to obtain feature points; Sd, matching the characteristic points of different images to obtain homonymous points of the sample part images; Se, carrying out binding constraint calculation according to the image atlas, pose information and the homonymous points of the sample part to obtain sparse point cloud; Sf, performing multi-view stereo matching according to the sparse point cloud and the clustered image to obtain a dense point cloud, and rendering the dense point cloud to obtain a three-dimensional model of a sample part; s2, collecting image data of all parts to be identified, and matching with a sample part three-dimensional model in the cloud model library to identify all the parts to be identified; s3, selecting a plurality of associated parts from all the parts to be identified, wherein the plurality of associated parts can be combined into one or more integral parts to obtain a three-dimensional model group corresponding to the plurality of associated parts; and S4, combining the three-dimensional model groups corresponding to the plurality of related parts to obtain a three-dimensional combined model, wherein the three-dimensional combined model is used for identifying and comparing the three-dimensional combined model with the whole part formed by the plurality of related parts.
  2. 2. The method for identifying the multi-part combination modeling according to claim 1, wherein the step S1 is performed, the method for acquiring the three-dimensional model comprises the steps of three-dimensionally scanning the sample part, acquiring three-dimensional image information of the sample part, and constructing the three-dimensional model by adopting a three-dimensional modeling tool or directly calling the three-dimensional model of the sample part from an existing database.
  3. 3. The multi-part combination modeling and recognition method according to claim 1, wherein the cloud model library comprises point cloud models of sample parts, feature point comparison is performed by using part image data to be recognized and the point cloud models in the cloud model library, and when the comparison similarity exceeds 95%, the part to be recognized is judged to be successfully recognized.
  4. 4. The multi-part combination modeling and identifying device is characterized by comprising a model library establishing module, a model library acquiring module and a model library acquiring module, wherein the model library establishing module is used for acquiring a three-dimensional model of a sample part and constructing a cloud model library of the sample part according to the three-dimensional model of the sample part, and the three-dimensional model of the sample part is acquired through the following steps: sa, acquiring a sample part image set, and performing digital processing on the sample part image set to obtain a sample part image; Sb, performing cluster analysis according to the image atlas and pose information of the sample part to obtain a clustered image; Sc, extracting features of the sample part image to obtain feature points; Sd, matching the characteristic points of different images to obtain homonymous points of the sample part images; Se, carrying out binding constraint calculation according to the image atlas, pose information and the homonymous points of the sample part to obtain sparse point cloud; Sf, performing multi-view stereo matching according to the sparse point cloud and the clustered image to obtain a dense point cloud, and rendering the dense point cloud to obtain a three-dimensional model of a sample part; The acquisition module is used for acquiring the image data of the part to be identified; The identification comparison module is used for collecting the image data of the part to be identified and matching the image data with the cloud model library of the sample part to identify the part to be identified; The matching module is used for selecting a plurality of associated parts, and the plurality of associated parts can be combined into an integral part to be matched with a corresponding three-dimensional model; And the combination comparison module is used for combining the matched three-dimensional models to obtain a combination model, and is used for identifying and comparing the combination model with the whole part formed by a plurality of related parts.
  5. 5. An electronic device, comprising: A memory for storing a computer program; A processor for implementing the steps of a multi-part combinatorial modeling recognition method as claimed in any one of claims 1 to 3 when executing the computer program.
  6. 6. A computer-readable storage medium, characterized in that a corresponding program of a multi-part combination modeling recognition method is stored on the computer-readable storage medium, which when executed implements a multi-part combination modeling recognition method as claimed in any one of claims 1 to 3.

Description

Multi-part combination modeling and identification method, equipment and medium Technical Field The invention relates to the technical field of visual recognition, in particular to a multi-part combined modeling recognition method, equipment and medium. Background In the process of stamping and automobile assembly of automobile panels, checking whether the stamped panels and parts meet the standard standards and whether the appearance is damaged or not is indispensable in the intelligent automobile factory panel stamping and part assembly links. The traditional target recognition algorithm is to collect two-dimensional images of parts by using a camera, and recognize and judge the types of the parts by the two-dimensional images of the parts. The other common mode is that the target recognition algorithm based on the neural network utilizes the training result of the two-dimensional image of the parts to carry out target classification and detection recognition according to the characteristics extracted from the two-dimensional image, and the two common target recognition modes are all used for recognizing the characteristics in the two-dimensional image, but the types of the automobile parts are various, the shapes are complex, the dimensions of the two-dimensional image are missing, and the overall view of the target is difficult to characterize. Therefore, when the method is used for detecting automobile parts based on the existing target recognition mode, how to improve recognition accuracy becomes a problem to be solved, and therefore, the method, the device and the medium for multi-part combination modeling recognition are provided for solving the problem. Disclosure of Invention The invention aims to provide a multi-part combination modeling and identifying method, equipment and medium, the method is based on a visual recognition algorithm and an enhanced implementation projection technology, can effectively recognize objects, and assists operators to confirm recognition results so as to reduce misjudgment. In order to solve the technical problems, the invention adopts the following scheme: a multi-part combinatorial modeling identification method, the method comprising the steps of: S1, acquiring a three-dimensional model of a sample part, and constructing a cloud model library of the sample part according to the three-dimensional model of the sample part; s2, collecting image data of all parts to be identified, and matching with a sample part three-dimensional model in the cloud model library to identify all the parts to be identified; s3, selecting a plurality of associated parts from all the parts to be identified, wherein the plurality of associated parts can be combined into one or more integral parts to obtain a three-dimensional model group corresponding to the plurality of associated parts; and S4, combining the three-dimensional model groups corresponding to the plurality of related parts to obtain a three-dimensional combined model, wherein the three-dimensional combined model is used for identifying and comparing the three-dimensional combined model with the whole part formed by the plurality of related parts. Further, when S1 is executed, the method for acquiring the three-dimensional model includes the following steps: sa, acquiring a sample part image set, and performing digital processing on the sample part image set to obtain a sample part image; Sb, performing cluster analysis according to the image atlas and pose information of the sample part to obtain a clustered image; Sc, extracting features of the sample part image to obtain feature points; Sd, matching the characteristic points of different images to obtain homonymous points of the sample part images; Se, carrying out binding constraint calculation according to the image atlas, pose information and the homonymous points of the sample part to obtain sparse point cloud; and Sf, performing multi-view stereo matching according to the sparse point cloud and the clustered image to obtain a dense point cloud, and rendering the dense point cloud to obtain a three-dimensional model of the sample part. Further, when the step S1 is executed, the method for acquiring the three-dimensional model comprises the following steps of carrying out three-dimensional scanning on the sample part, acquiring three-dimensional image information of the sample part, and constructing the three-dimensional model by adopting a three-dimensional modeling tool or directly calling the three-dimensional model of the standard part from the existing database. Further, feature point comparison is carried out by utilizing the image data of the part to be identified and the point cloud model in the three-dimensional model comparison library, and when the comparison similarity exceeds 95%, the part to be identified is judged to be successfully identified. The invention provides a multi-part combination modeling and identifying device, which comprises: The model libr