Studio 4

intuitive powerful adaptable

software for machine vision engineers

Adaptive Vision 4
1
 Introduction

Adaptive Vision Studio

Adaptive Vision Studio software is the most powerful graphical
environment for machine vision engineers. It is based on
dataflow and comes with a comprehensive set of powerful,
ready-for-use image analysis filters.

Its unique strength lies in its focus on professional users – it

allows you to create typical applications easily, but at the same
time makes it possible to efficiently develop highly customized
and large-scale projects.

Features

Intuitive

Drag and drop You can see everything HMI designer

All programming is done by choosing filters and Inspection results are visualized on multiple You can easily create custom graphical user
connecting them with each other. You can focus configurable data previews; and when a interfaces and thus build the entire machine
all your attention on computer vision. parameter in the program is changed, you can vision application using a single software
see the previews updated in real time. package.

Powerful


1000 effective filters Hardware acceleration Loops and conditions

There are over 1000 ready-for-use machine The filters are aggressively optimized for the Without writing a single line of code, you
vision filters tested and improved in hundreds SSE technology and for multicore processors. can create custom and scalable program
of applications. They have many advanced Our implementations are ones of the fastest in flows. Loops, conditions and subprograms
capabilities such as outlier suppression, the world. (macrofilters) are realized with appropriate
subpixel precision or any-shape region-of- data-flow constructs in the graphical way.
interest.

Adaptable
GigE vision and GenTL User filters C++ and .NET generators

Adaptive Vision Studio is a GigE Vision You can use user filters to integrate your Programs created in Studio can be exported to
compliant product, supports the GenTL own C/C++ code with the benefits of visual C++ code or to .NET assemblies. This makes
interface, as well as a number of vendor- programming. It takes 30 seconds to build first it very easy to integrate your vision algorithms
specific APIs. Thus, you can use it with most user filter and 5 seconds to update one after with applications created in C++, C# or VB
cameras available on the market. recompilation. programming languages.

Overview Adaptive Vision 4

2
 Workflow

Design the algorithm 1

Creating vision algorithms consists in repeating three
intuitive steps:

1 Drag & drop filters from the toolbox to the program

editor.

2 Drag & drop connections between the filters or set

constant input values.

3 Drag & drop filter outputs to data preview panels for

interactive analysis.

The only expertise users need to have is about the core art
of computer vision – which filters to connect and how. And
yet another great feature of Adaptive Vision Studio is that
you can develop this competence quickly through intuitive
experimentation.

Create a custom HMI 2

Creating a custom graphical user interface is just as easy:

1 Drag & drop controls from the control catalog to the HMI
panel.

2 Set the controls’ properties.

3 Drag & drop connections between the filters and the

controls.

Note that you can also easily integrate your algorithm

with an HMI created in C++ or C#.

Deploy 3
When the program is ready, you can export it to a runtime
executable and deploy it on a PC-based industrial computer
or on a smart camera such as CORSIGHT from NET or
NEON from Adlink.

Adaptive Vision 4 Workflow

3
 Data Flow Programming

Introduction Macrofilters
Adaptive Vision Studio is a drag and drop environment Macrofilters provide a means for building large real-life
designed for machine vision professionals. Experience in programs. They are reusable subprograms with their own
low-level programming is not required. Nevertheless, it is inputs and outputs, and as a whole they can be used just as
a  fully fledged visual programming language and you have to any regular filter. Moreover, macrofilters can perform iterative
understand its 4 core concepts: Data, Filters, Connections and computations with state registers, thus providing an elegant
Macrofilters. way to create programs of any complexity.

Data & Filters

Filters are the basic data processing elements in the data flow
driven programming. In a typical machine vision application
input images are tranformed by a sequence of filters into
regions, contours, geometrical primitives and then into final
results such as a pass/fail indication. Adaptive Vision Studio
also supports data collections (arrays) and filters that store
information between consecutive iterations to compute
aggregate results.

Connections
Different types of connections between filters support: basic
flow of data , automatic conversions , for-each
processing and conditional processing . The
connection types are inferred automatically on the do what
I  mean basis.

More information can be found on our web page and in the

user manual.

Data Flow Programming Adaptive Vision 4

4
Tutorial
Here is a simple example of
a data-flow driven program

1 An image is acquired from a camera:

2 The image is thresholded into a region (a run-length

encoded binary image) and then split into connected
components:

3 ClassifyRegions filter analyzes the elongation feature of

each component:

{ 19.554, 19.433, 3.866, ...}

4 ...and splits the array of regions into arrays of accepted

and rejected elements accordingly to the specified range:

5 Finally the bounding circles of the accepted objects are

drawn:

Adaptive Vision 4 Data Flow Programming

5
 Capabilities

Filters
There are over 1000 filters encompassing both basic transforms and specialized machine vision tools.

Image Processing Camera Calibration Corner Detection Edge Template Matching

Blob Analysis Fourier Analysis 1D Profile Analysis Histogram Analysis

Contour Analysis Hough Transform 1D Measurements OCR

Planar Geometry Barcode Reading 2D Measurements Machine Learning

Shape Fitting Data Code Reading Gray Template Matching GigE Vision and GenTL

ANOTHER OPEN
Performance FILTER AVS 4 PRODUCT LIBRARY

Image negation 0.055 ms 0.103 ms 0.055 ms

The filters of Adaptive Vision Studio are highly optimized for
Image difference 0.050 ms 0.125 ms 0.055 ms
modern multicore processors with SSE2 technology. Speed-up
factors that can be achieved are however highly dependent on Rgb to Hsv conversion 0.632 ms 1.243 ms 0.586 ms
the particular operator. Simple pixel-by-pixel transforms after
the SSE-based optimizations already reach memory bandwidth Gauss filter 3x3 0.063 ms 0.290 ms 0.439 ms
limits. On the other hand, more complex filters such as gauss
Gauss filter 5x5 0.071 ms 0.323 ms 0.661 ms
smoothing can achieve even 10 times lower execution times
than with C++ optimizations only. 3.132 ms
Gauss filter 21x21 0.559 ms 1.815 ms

Image erosion 3x3 0.055 ms 0.127 ms 0.156 ms

Speed Up Factors
Image erosion 5x5 0.059 ms 0.128 ms 0.209 ms
SmoothImage_Gauss SSE SSE + 4 Cores
Threshold to region 0.061 ms 0.032 ms

SmoothImage_GaussFast Bilinear image resize 0.205 ms 0.272 ms 0.114 ms

Splitting region into blobs 0.039 ms 0.082 ms

RotateImage

The above results correspond to 640x480 resolution, 1xUINT8 on an

ThresholdImage
Intel Core i7 - 3.4 GHz (4 cores) machine. To assure the most precise
results big images were tested and the results were normalized.
Note also that the functions from the different libraries do not always
0 1 2 3 4 5 6 7 8 9 produce exactly the same output data.

Capabilities Adaptive Vision 4

6
 More than Visual Programming

Large-Scale Projects
One of the most distinguishing features of Adaptive Vision
Studio is its applicability for large-scale projects. It has been
used to build algorithms composed of hundreds of tools
and in systems that required maintenance of many different
product versions. Creating such systems is possible due to the
feature of modules, hierarchical view of the program structure,
version control friendly file format and support for test-driven
development.

User Filters
User filters allow C++ programmers to integrate their existing or
new code with the benefits of the rapid graphical development.
It works with Microsoft Visual C++ (including Express edition).

To add a new user filter click File / Add New Filter to have a
sample C++ filter ready and functional in less than 30 seconds.
Then on you can work side by side in both enviroments and
have your filters automatically reloaded each time you build
the C++ project.

Adaptive Vision Library

Adaptive Vision Library is a tool for C++ and .NET programmers

available as a single DLL file with appropriate headers. The
design is simple – for each filter of Adaptive Vision Studio
there is a corresponding function in Adaptive Vision Library
with several input and output parameters corresponding to the
input and output ports of the filter.

The C++ code or a .NET assembly can also be generated

automatically from algorithms created with Adaptive Vision
Studio. This creates a unique possibility to create a program
in a graphical environment and then compile it to a native
Windows executable.

Adaptive Vision 4 More than Visual Programming

7
 About Adaptive Vision
Adaptive Vision is a reliable supplier of high quality machine
vision software and libraries.

Our team has been concentrating on creating computer vision

software designed for specific applications for the last 16
years. Initially, we have built our expertise in projects related to
intelligent video surveillance, biometric systems and medical
imaging, which were delivered under individual contracts with
companies from the UK. In 2007 we have moved our focus
to industrial vision systems and today we deliver standard
software products for this market.

Looking for Developers?

We also operate as a team of programmers supporting
machine builders and system integrators in developing
bespoke machine vision algorithms and solutions. Moreover,
as a part of Future Processing Sp. z o. o., one of the leading
European software outsourcing companies, we can also offer
you highly scalable services in general software engineering.

Design by:
Adaptive Vision
Future Processing Sp. z o. o.

ul. Bojkowska 37A

44-100 Gliwice, Poland
Tel: +48 32 461 23 30
Fax: +48 32 461 23 09

info@adaptive-vision.com
www.adaptive-vision.com

More than Visual Programming Adaptive Vision 4

8