BOBCAT Hardware User’s Manual
Imperx, Inc.
Rev. 2.0.2
6421 Congress Ave.
11/20/2012
Boca Raton, FL 33487
+1 (561) 989-0006
144 of 265
3.1 OVERVIEW
The camera has built-in several basic image processing functions. More functions will be
added later. Please contact Imperx for more information.
3.2 IMAGE ENHANCEMENT
In many imaging applications the user will have a dark object on a bright background,
many dark and bright spots or shadows, or the light will not be sufficient, so the resulting
image will have a low contrast, and/or a very low dynamic range. To improve the image
quality in such conditions, BOBCAT offers a set of image enhancing features –
thresholding and multi-point image correction. The processing function is applied to the
entire image unless AOI8 is enabled as “Processing ROI”. In this case the processing
function will apply only to the selected ROI.
3.2.1 Threshold Operation
In many applications the binary images are much simpler to analyze that the
original gray scale one. The process, which converts the regular gray scale image
to binary, is called “Thresholding”. Thresholding is a special case of intensity
quantization (binarisation) where the image can be segmented into foreground and
background regions, having only two gray scale levels “white” and “black”.
Selecting the threshold value is very critical for the binary image quality, and it is
to a great extend scene dependent. If a threshold level is chosen correctly, this will
produce a well-defined boundary of the object, which is essential. In some cases it
is desirable if part of the image is binary and some is grayscale image. The
camera has built in four thresholding modes:
3.2.1.1
Single Threshold Binary
If the image is a high contrast scene and has well defined bright and dark
regions a simple binarisation technique can be used for thresholding –
Formula 3.1. The binary image output is converted to “white” for all gray
level values higher or equal to the selected threshold point X1, and to
“black” for all gray levels lower than X1. The user can set the optimal
threshold value. Figure 3.1 shows the original and the processed image
with single threshold.
Output signal =>
“WHITE” if (input signal
X1)
“BLACK” if (input signal < X1)
(3.1)
