Search

EP-4361554-B1 - AUTOMOTIVE WIRING INSPECTION METHOD

EP4361554B1EP 4361554 B1EP4361554 B1EP 4361554B1EP-4361554-B1

Inventors

  • MARTÍNEZ ZAMBRANA, JUAN ANTONIO

Dates

Publication Date
20260506
Application Date
20221024

Claims (14)

  1. An automobile wiring inspection method which uses of a worktable (1) on which one or more automobile cables (3) are arranged such that the cables (3) are stretched, the method comprising the steps of: - acquiring (101), with a camera (4), an image of the cables (3) arranged on the worktable (1), - processing (104) the image, transforming it into a binary image, in which the cables (3) are lines representing their geometry and arrangement, - creating (106) reference nodes on the cables (3) of the processed image, - calculating (107) a distance in pixels between the reference nodes, - converting the distance in pixels between the reference nodes to a distance in units of measurement using an artificial intelligence model previously trained by means of a known image geometry for configuring many calibration points, generating with all these points a set of conversion factors to be applied to different areas of the image, and transforming the distance measured in pixels to a distance in length units using the conversion factors; and - repeating the calculation (107) and conversion steps until a set of distances of interest between the reference nodes is obtained.
  2. The method according to claim 1, comprising a previous step of fixing the cables (3) to the worktable (1) by means of fastening harnesses (2), such that the cables (3) are stretched and tensioned.
  3. The method according to claim 1, wherein the step of processing (104) the image comprises the sub-steps of: - converting (204) the image to grayscale, - blurring (205) the image, maintaining the focus only on the cables (3), - binarizing (206) the image, obtaining an image with black and white pixels, and - thinning (208) the cables (3) in the image, such that they are reduced to a geometry with a set of centered curves.
  4. The method according to claim 3, wherein a Gaussian blur is performed in the blurring step (205).
  5. The method according to claim 3, comprising an additional step after the binarizing step (206) of filling in (207) the pixels left empty inside the cables (3) of the image.
  6. The method according to claim 1, wherein acquiring (101) the images is performed from the top, at a point opposite the worktable (1).
  7. The method according to claim 1, wherein acquiring (101) the images is performed laterally with respect to the worktable (1) and the cables (3).
  8. The method according to claim 1, which may comprise an additional step prior to the step of obtaining the images (102) of illuminating the worktable (1) to improve the quality of the images.
  9. The method according to claim 1, comprising an additional step of identifying shapes, comprising a sub-step of applying template matching on the acquired image of the cables (3) in order to identify an element of interest.
  10. The method according to claim 1, comprising an additional step of identifying the color of the cable (3), comprising the sub-steps of: - locating each of the pixels of the image in an HSV color space, - determining the color of each of the pixels from its location in the HSV model, taking into account a predefined minimum and maximum acceptance range, - determining the color of the cable (3), - from the color of each of the pixels of the cables (3), determining defects having a color different from that of the cable (3).
  11. The method according to claim 3, comprising an additional step of identifying protuberances in the cable (3), comprising the following sub-steps: - determining, in each pixel of the curves centered on the original shape of the cable (3) created in the thinning step, the orientation of the pixel and of its neighboring pixels, - determining when a protuberance starts, as well as its size based on the comparison of the orientations of the pixels, - if a protuberance is detected, determining its orientation and position with respect to the closest node.
  12. The method according to claim 1, wherein, when the cables (3) are connected to connectors, it comprises an additional step of detecting the connectors, comprising the following sub-steps: - segmenting the image to determine an area in which the connector is located, - inspecting the area to determine the limits of the connector, - performing a histogram analysis to determine an area of attachment of the cable (3) with the connector, and - positioning a node in the area of attachment of the cable (3) with the connector, in order to be able to subsequently determine the length of the cable (3).
  13. A computer program configured when run on a computer used in combination with a camera for executing the steps of the method according to any of claims 1, 2, 4 to 7 and 9 to 11.
  14. A storage unit configured for storing the computer program according to claim 13.

Description

OBJECT OF THE INVENTION The object of the present invention relates to an automobile wiring inspection method, particularly applicable in the technological field of automatic measurements of dimensions, specifically in the sector of length measurement and component detection by shape and color, within automobile wiring. In the method, measurement nodes are defined in images taken by means of cameras positioned on the cables and a length relationship is established between measurements in the International System and pixels, in order to be able to obtain the complete or partial length of a cable or a cable assembly. In addition, the different components connected to the cables can be identified based on recognition of their shape and color. BACKGROUND OF THE INVENTION The cable harnesses or cable assembly of an automobile refer to a component of the electrical circuit formed by grouped copper cables and wires, with subsequent plastic compression of an insulator or mechanical element. Today, with the increase in the supply of electric vehicles in which the electrical wiring is becoming very important, it is necessary to check the cables that will be incorporated in the vehicle and their dimensional characteristics, which are very strict. Cable harnesses in automobiles have become a frequent cause of failures in electric vehicles, one of these types of failures being factory defects in terms of lengths or dimensions, which may cause problems relating to information loss between elements or simply to them not corresponding to the expected dimensions, and therefore cannot be installed, when being introduced into the car. To ensure the safety and reliability of the cable harness and quality approval, the cable harness must be inspected and examined strictly and quickly in order to find and eliminate unqualified products with problems such as poor contact, poor insulation, and assembly errors in time. Electrical faults are performed on electrical test benches but the checking of dimensions and certain wiring elements is, however, currently performed by means of a manual visual inspection process, causing several problems, such as increased production time, loss of efficiency because it is performed by an operator who may be subjected to stress or fatigue, and therefore leading to inspection errors. Kamble Supriya S. and Kulkami Ashwini A. "Automatic Optical Inspection System for wiring harness using Computer Vision", 2021 IEEE International Conference on Electronics, Computing and Communication Technologies (CONECCT), IEEE, 9 July 2021 (pp. 1-5) discloses a method for determining the length of a cable disposed on a worktable based on an image captured by a camera. DESCRIPTION OF THE INVENTION The present invention as defined in the independent claim 1 seeks to solve the problems arising from wiring inspection methods known in the state of the art by means of automatically measuring the length of wiring in automobiles. In a first step of the method, a single cable or an assembly of cables attached to one another are placed on a worktable, stretching each cable as much as possible in order to be able to capture images as accurately as possible, i.e., with the curvatures typical of the flexible geometry of the cable being absent as much as possible. The cables may incorporate elements such as clips and connectors. The cables can be arranged directly on the table without fastening or can be fixed to the table by means of fastening harnesses such that they remain taut. In addition, in either case, references can be placed on the table to guide an operator when placing the cables. This provides great flexibility because the fact that some branches are not fastened to a tension device but rather rest directly on the table makes the placement thereof much faster. An artificial intelligence algorithm will then look for the locations where the connectors are not fastened to the elements. The worktable can also be illuminated to prevent shades and to ensure that the images are of the highest quality possible. Image acquisition is performed with one or more cameras, such as 20-megapixel Ethernet cameras, for example, which take images from top views (on a face opposite the worktable) and/or side views (with respect to the worktable) of the wiring assembly. The top or side images are taken depending on the geometry of the cables, such that images which provide the clearest image of the cables are taken. Once the images have been taken, the processing thereof, which consists of transforming the color image into a binary image, is carried out, in which the cables are reduced to lines that represent their geometry and arrangement in a simplified manner. A preferred embodiment of the different steps to be followed for processing the acquired image is described below. First, the images are converted to grayscale. A blurring step, preferably Gaussian blur, is then performed, eliminating elements that are not essential in the acquire