being re-inserted into the inspection system, the
shape of the cap and its orientation are verifed.
Since the operator must remove the part to insert
this cap, it is also necessary to again perform 2D
matrix barcode reading to ensure that the part
being inspected is the same as the one previously inspected for PCB component placement. In
this way, any operator error is eliminated.
To perform these tasks, the electrical housing is frst placed into a custom-built nest after
which it is positioned into the machine vision
system. In the design of the system, two cameras are placed at on and off-axis locations to
view the part (Figure 2). The on-axis camera,
an IVS-USB4600c color USB3 camera with
a 4608x3288 CMOS image sensor, is used to
image the PCB.
To ensure that the components on the PCB
are imaged accurately and repeatably and that
any blurred edges that may be caused by diffuse
refections are eliminated, the camera is coupled
to a bi-telecentric lens via a C-mount thread to
provide a 160 x 120 mm feld of view (FOV) of
the PCB. A custom 90º diffuse axial light from
DCM Sistemes (Valencia, Spain; www.dcmsis-
temes.com) mounted in front of the lens provides
the diffuse light illumination needed to improve
image consistency while avoiding any refections
and shadowing effects that may otherwise occur.
Data Matrix code
Before the on-axis camera is used to check
whether components are correctly in place,
however, the system must frst validate wheth-
er the 2D Data Matrix code on both the PCB
and the side of the unit’s housing are correct. To
do so, a second camera, an IVS-USB640c USB
color camera, is positioned at 90° to the part. To
illuminate the 2D Data Matrix code on the side
of the housing, an MSP 1.25in white LED spot
light from Spectrum Illumination (Montague,
MI, USA; www.spectrumillumination.com) is
mounted next to the camera (Figure 3).
To read both the Data Matrix code on the
PCB and the unit’s housing, images captured
by both cameras are transferred to an indus-
trial PC over a USB3 interface using the IVS
soft ware. After images are captured, both ECC
200 Data Matrix codes are read by the IVS
software. After reading, barcode data is trans-
ferred over an industrial Ethernet interface to
the automotive manufacturer’s factory infor-
mation system that is then used to initiate the
fnal testing procedure.
Since the electrical components have
already been tested electrically and the solder
joints whether on the PCB have been deter-
mined acceptable, the fnal inspection pro-
cess requires a check of whether components
have been correctly placed and not damaged.
Thus, at the next stage of the inspection pro-
cess, components on the PCB must be validat-
ed as being correctly placed and that the PCB
is properly ftted into its aluminium housing.
To do so, requires frst defning over 170 regions
of interest (ROIs) within the feld of view of the
image using a known good part. These ROIs
are then used to perform pattern matching on
any new part being imaged.
The vision system’s human machine interface (HMI) developed by IVS provides an
analysis of the inspected PCB showing failed
areas highlighted in red. This provides feedback to the operator of missing and/or damaged components.
Figure 2: In the design of the system, two cameras are used to inspect the part. The frst,
through a telecentric lens and on-axis illumination system inspects for correct component placement. The second, mounted at 90° inspects the 2D barcode on the side of the assembly.
Figure 3: To inspect the part, an off-axis illumination system is used (top) and images through
the cameras bi-telecentric lens. Reading the barcode on the side of the part is accomplished using an on-axis bar light (left) and a second camera mounted to the right.