HOME  > Low Cost Automation Tutorial  > #327 Know-how on automation: Automating screw fastening - 5: Image processing and parts transport
Low Cost Automation Tutorial

#327 Know-how on automation: Automating screw fastening - 5: Image processing and parts transport

Category :
February 9, 2018

One of the powerful production technologies useful for addressing troubles in automating parts transport is image processing.Image processing technique is not only effective for addressing troubles such as wrong orientation or positioning error for transported parts on the alignment unit or conveyor, but also for as countermeasure for entry of foreign substances.This section provides an overview of image processing techniques.

(1)Automation and image processing

・In the production lines, image processing technique can be used as a tool for visual inspection for all processes including transporting, machining/assembling, and inspection.
・Image processing, adopted as a countermeasure for troubles like positioning error, helps to measure the correctness of position and orientation of screw parts before the positioning stage, and to reject inappropriately oriented parts (Fig. 1).

[Fig. 1] Alignment in Linear Feeder from NTN Corporation

・Registering an appearance pattern of part for inspection to avoid entry of foreign substances will help to reject parts with wrong pattern (foreign substances).
・The image processing technique is also applied to the field of inspection as a means of examining the appearance after fastening the screw.

(2)Basics of image processing technique

Image processing mainly uses the four methods in recognizing patterns of states, following the model of the human eye's functions:
a)Determine region by binarizing the image signal
b)Extract local characteristics of object and apply them to inspection
c)Register a template shape for inspection in advance as a pattern for orientation, circuit, etc., and apply it to positioning or inspection of scratches by checking it against the template (template matching) (Fig. 2)
d)Calculate and determine global characteristics for areas or circumference length of an object, and apply them to positioning or classification of the parts.

[Fig. 1] Variety of screw parts

Explanation

・In the process of binarization, the image signals are digitized into either "1" or "0", depending on whether they are greater or less than the threshold value for the image signal (example: brightness).

・Binarization will make shape inspections possible such as follows when appearance recognition of the object is made possible:
1)Recognition of product and inspection for foreign substances from projected shape
2)Gravity of projected shape
3)Distance from gravity to contour (maximum distance, average distance)

・Local characteristics determined by appearance recognition of object cannot be applied to examinations of printed circuit boards or inspection for scratches.
・In the template matching technique, characteristics of shape required for recognizing the image of object are registered as templates for pattern recognition, which are used for precise positioning and orientation control for complex shapes, or determination of circuit pattern defects.
・Since shapes of objects vary a great deal (three dimensional object with convexes and concaves), an appropriate technique must be selected according to the following fluctuating parameters:
1)Orientation of image processing camera (point of sight, largeness of image area, etc.)
2)Illuminating condition of image processing area (luminous intensity, orientation, change rate in luminosity, etc.)
3)Optimization of object state (avoidance of parts overlap, etc.)

MISUMI USA eCatalog