Search

EP-3818473-B1 - SYSTEM AND METHOD FOR AUTOMATED VISUAL INSPECTION

EP3818473B1EP 3818473 B1EP3818473 B1EP 3818473B1EP-3818473-B1

Inventors

  • HYATT, Yonatan

Dates

Publication Date
20260513
Application Date
20190702

Claims (9)

  1. A method for automated visual inspection of an item on an inspection line, the method comprising: using a processor (102) to: detect (202) an item on an inspection line (105) in an image of the inspection line (202'); using a processor (102) to produce (204) from the image (202') a reduced image (116; 204') comprising image data of the item or of part of the item and in which image data outside the item or the part of the item is eradicated such that the image data outside of the item or the part of the item is unreadable; storing the reduced image (116; 204') in a storage device such that the reduced image (116; 204') is accessible for further use in an inspection process; and applying a machine learning process to the reduced image (116; 204') to detect a defect on the item in the inspection process.
  2. The method of claim 1 comprising after detecting the item in the image of the inspection line (202'), creating an outline of the item in the image (202') and eradicating the image data outside the outline.
  3. The method of claim 2 comprising detecting the item in the image (202') using computer vision techniques and creating the outline based on the detection of the item.
  4. The method of claim 2 wherein the outline of the item includes the item and a predetermined periphery of the item.
  5. The method of claim 2 wherein the outline of the item includes only a portion of the item.
  6. The method of claim 1 comprising displaying the reduced image (116; 204').
  7. The method of claim 1 comprising using the stored reduced image (116; 204') in a post-inspection process.
  8. The method of claim 1 wherein the inspection process comprises a set up stage prior to an inspection stage, the method comprising: receiving a setup image of the inspection line in the setup stage; and collecting, in the setup stage, representing information of an item from the setup image, wherein the detected item is of a same-type as the item from the setup image and is detected based on the representing information; and wherein the method further comprises inputting the reduced image to the inspection process to detect a defect on the same-type item at the inspection stage.
  9. The method of claim 1 wherein the reduced image (116; 204') comprises information indicating a status of detection of the item in the image of the inspection line (202') or a status of detection of a defect on the item for use in the inspection process.

Description

FIELD The present invention relates to automated visual inspection processes, for example, inspection of items during a production process. BACKGROUND Inspection during processes can be instrumental in ensuring the quality of the process. For example, inspection during production processes at manufacturing plants helps control the quality of products by identifying defects and then acting upon this identification, for example, by fixing the defect or discarding the defective part. During production, the process of defect detection is essential for quality assurance (QA) gating and sorting on production lines, and is consequently useful in improving productivity, improving production processes and working procedures, reducing defect rates, and reducing re-work and waste. Systems that use imaging devices and image processors, generally referred to as machine vision systems, are used in inspection tasks, where a camera is set up to scan the process or item under inspection. Images obtained by the camera are then analyzed, for example, to detect defects in items. Such machine vision solutions necessitate positioning and arranging of cameras and typically require a set up stage by an operator and/or set up crew. Images obtained by the cameras during the set up stage are recorded and viewed (e.g., via a closed-loop video system) for feedback regarding the positioning and arrangement of the cameras, ambient illumination, and other parameters important for setting up a machine vision system. The images are also recorded and viewed to enable monitoring by the operator or another reviewer to correct or improve the set up process, if necessary. These images, which are obtained during the set up stage, almost always include views of the crew involved in setting up and other surroundings and background which are not relevant for the set up or for the following inspection process. Consequently, images obtained during the set up stage and during the following inspection stages, typically include personal and/or confidential information, such as one or more crew people or parts of bodies of crew people, machinery or parts of the machinery used at the manufacturing plant and other background depicting the environment of the inspection line and plant. The use of these images, which are obtained during visual inspection processes, may be extremely restricted, due to security policies and/or privacy regulations. Since use of images obtained during the set up stage and during the following inspection stages, is essential in visual inspection processes, restrictions on the use of these images may harm and even bring the inspection process to a halt, causing irreparable damage. US 9,330,339 B2 discloses systems and methods for detecting container seals. In one implementation, a processing device receives one or more images and processes the one or more images to identify one or more areas of interest within at least one of the one or more images, the one or more areas of interest including one or more areas of the one or more images within which a container seal is relatively likely to be present. The processing device compares the one or more identified areas of interest with one or more templates, each of the one or more templates being associated with one or more respective seal types. The processing device identifies, based on the comparison, one or more matches between the one or more identified areas and the one or more templates, and provides, based on the identification of the one or more matches, an identification of the container seal. EP 2 801 815 A1 discloses an image inspecting device with an inspection area setting means which takes an image of a sample to be inspected, and uses the resulting sample image to set in advance the inspection area to be used in the inspection process. The inspection area setting means seeks the optimal solution for the inspection area by evaluating both the pixel separation of the colors or brightness between the inside and outside of the inspection area and the edge overlap between the contours and edges of the inspection area based on color or brightness information from each pixel in the sample image and edge information included in the sample image. Also, the inspection area setting means adjusts the weighting when evaluating the pixel separation and edge overlap using balance parameters inputted by the user. JP 2005-208890 A discloses an inspection method in which, after performing defect inspection, history information in which the inspection result is recorded and secret level information indicating whether the inspection result is to be deleted as security information or not are acquired. The existence of defect detection is checked, and when no defect is detected, inspecting image data are deleted and the fact of deletion is recorded in the history information. When a defect is detected, whether the inspected result is to be deleted as security processing corresponding to secret