Introduction to nine conventional vision software

1. Open source OpenCV

OpenCV (Open Source Computer Vision Library: http://opencv.org) is an open-source BSD-licensed library that includes hundreds of computer vision algorithms. The OpenCV 2.x API in the documentation describes the C++ API, and there is also an OpenCV 1.x API based on the C language, the latter of which is described in the documentation opencv1.x.pdf.

OpenCV has a modular structure, which means that the development package contains multiple shared libraries or static libraries. Here are the modules that can be used:

Core functionality - A compact module that defines basic data structures, including dense multidimensional Mat arrays and basic functions used by other modules.

Image processing - An image processing module that includes linear and nonlinear image filtering, geometry transformation (reset size, radial and perspective deformations, general base table reset mapping), color space transformation, histogram, etc.

Video - An image analysis module that includes motion determination, background reduction, and target tracking algorithms.

3D Calibration (calib3d) - Multi-view based geometry algorithm, flat and stereoscopic camera calibration, object pose determination, stereo matching algorithm, and reconstruction of 3D elements.

Features2d - Determination of raised features, feature description, and comparison of feature descriptions.

Object Detect - Targets and predefined types of instantiated detection (e.g., faces, eyes, cups, people, cars, etc.).

HighGui - An easy-to-use user function interface.

Video Input and Output (Videoio) - An easy-to-use video capture and video decoder.

GPU - GPU acceleration algorithms from different OpenCV modules.

… Some other auxiliary modules such as FLANN and Google's test packages, Python bindings, and others.

OpenCV is very suitable for low-level scientific researchers, with low cost and very convenient use, and the hub is open source.

2. VisionPro 7.0 system

Cognex's VisionPro system combines world-class machine vision technology with fast and powerful application system development capabilities. VisionPro QuickStart accelerates application prototyping by dragging down the shift. This result can be applied throughout the application development cycle. Develop application systems by using COM/ActiveX-based VisionPro machine vision tools and graphical programming environments such as Visual Basic and Visual C++. Combined with the MVS-8100 Series frame grabbers, VisionPro enables manufacturers, system integrators, and engineers to quickly develop and configure powerful machine vision applications.

Rapid prototyping and easy integration

VisionPro's two-layer software structure makes it easy to prototype and integrate. The interactive layer uses drag-down gear and ActiveX controls to accelerate application development; At the program level, prototype applications are developed into user solutions. COM/ActiveX technology makes it easy to integrate VisionPro applications with third-party utilities (e.g., graphical functions) and provides easy integration of COM-based applications for the entire machine (e.g., I/O, robot control, factory communications).

Improve the machine vision software of its predecessors

Cognex's Vision Tool Library provides a set of vision software programs for measurement, detection, guidance, and recognition. These tools are proven to be highly reliable in even the most challenging vision applications.

Hardware flexibility

VisionPro users have a wide range of options for MVS-8100 Series frame grabbers to develop vision applications. Tested and validated by VisionPro software, these frame grabbers provide the host with high-speed image transfer for image processing and display for fast manipulation of vision applications. Multi-camera inputs, high speed, and support for high-resolution cameras improve the acquisition flexibility of the VisionPro system.

Technology has the upper hand

Rapidly develop powerful PC-based vision applications

Simplify the integration of vision systems with other master control programs

Compatible with a wide range of Cognex MVS-8100 Series frame grabbers

Accelerate prototyping with QuickStart dragging shifts

Combine with a visual tool library for high performance

Software structure of VisionPro

VisionPro provides easy-to-apply prototypes, developments, and applications. At the interactive layer, the VisionPro QuickStart prototype environment accelerates the development of powerful machine vision systems by dragging down the work equipment in series. Users can quickly define connections between tools, test tool behavior, and valid running parameters. ActiveX controls are accessible through QuickStart or Visual Basic. Therefore, at the program level, prototype applications can be enhanced with Visual Basic or Visual C++ to develop personalized solutions. This feature will significantly reduce the development time of the application system. VisionPro is structured so that users can start their application development from any layer of QuickStart, ActiveX, or COM objects.

Cognex Corporation announces the launch of VisionPro7.0 machine vision software. VisionPro 7.0 introduces InspectionDesigner, a new concept that simplifies the development of inspection applications and field repairs. VisionPro 7.0 also adds the industry's best barcode reading tool, 1DMax, optimized for all-round and harder-to-read 1D barcodes, as well as several new tools and enhancements.

The introduction of Inspection Designer — for the first time for general inspection applications — provides vision system integrators and ultimate users with features that simplify the specification, development, and repair of inspection applications. Inspection Designer has three main advantages:

Defining defects. A new image grading utility that enables integrators and ultimate users to easily grade product images and specify different defect types in individual images. The Image Rating utility creates a library of graded images for use in the next steps.

Verification. Validation tools help developers quickly confirm that their vision system is getting the results they need by comparing the inspection results of hundreds or even thousands of retained and graded images.

On-site repairs. The validation tool also promises the ultimate user to retest the overhaul tool at any time against the rated image library to confirm that factory floor adjustments will not affect the previous inspection application. It also promises users to add new images to the rated artifacts library.

"The introduction of Inspection Designer provides a common communication tool for system integrators and their end users by defining a formal scale for defining good and bad artifacts," says Markku Jaaskelainen, Business Manager of Vision Software. ”

3. LabVIEW is used for machine vision

NI's application software LabVIEW machine vision software is the fastest programming speed. LabVIEW is a graphical programming language based on program code. It provides a large number of image preprocessing, image segmentation, image understanding function libraries and development tools, and users only need to use the icon connector at the top of the flowchart to connect the required sub-VIs (VirtualInstruments LabVIEW development programs) to complete the target task. Any VI has 3 departments: an interactive user interface, flowcharts, and icon connectors. LabVIEW is simple to program and has a high accuracy of identifying the workpiece.

4. MVTecHALCON in Germany

HALCON is a complete set of machine vision algorithm packages developed by MVtec in Germany, with a widely used machine vision integrated development environment. It saves product costs and shortens software development cycles – HALCON's flexible architecture facilitates rapid development of machine vision, medical imaging and image analysis applications. It is already recognized by the European and Japanese industries as the best machine vision software.

HALCON originates from academia and is different from commercial software packages in the market. In fact, it is an imageprocessing library consisting of more than a thousand independent functions and an underlying data governance core. It includes basic geometry and image computing functions such as various filtering, color and geometry, mathematical conversion, morphological calculation and analysis, correction, classification recognition, shape search, etc., because most of these functions are not designed for specific tasks, so as long as you can use the image processing place, you can use HALCON's powerful computational analysis capabilities to complete the work. The application range is almost unlimited, covering medicine, remote sensing detection, monitoring, to various types of automated testing in industry.

HALCON supports Windows, Linux and Mac OS X manipulation environments, which guarantees the effectiveness of the investment. The entire library can be accessed in a variety of common programming languages such as C, C++, C#, Visual basic, and Delphi. HALCON provides interfaces for a large number of image acquisition devices, ensuring hardware independence. It provides interfaces for more than 100 industrial cameras and frame grabbers, including GenlCam, GigE, and IIDC 1394.

HALCON function

In line with MVTec's ongoing collaboration with academia, the latest HALCON 13 features the following new features:

1. Technological innovation

HALCON 13 enables true target recognition. Sample-based recognition methods can distinguish a large number of target objects. This technique enables the identification of practiced targets based solely on features such as color or texture, eliminating the need for special imprints such as 1D or 2D codes for target identification.

2. Strong three-dimensional visual processing

A new technology offered by HALCON 11 is 3D surface comparison, which compares the surface profile measurement results of a 3D object with the expected shape. All 3D technologies offered by HALCON, such as multi-eye stereoscopic vision or sheet of light, can be used for surface reconstruction; It also supports 3D reconstruction directly from off-the-shelf 3D hardware scanners. In addition, the photometric stereo vision method has been improved for special applications in surface inspection. In addition, HALCON now supports many 3D target processing methods, such as point cloud calculation and triangulation measurement, feature calculation such as shape and volume, and point cloud segmentation through sectional surfaces.

3. High-speed machine vision experience

Automatic Operator Parallel Processing (AOP) technology is a unique feature of HALCON. HALCON 11 supports more than 75 operators for machine vision algorithms using GPU processing, more than any other software development kit offers. In addition, there is a significant acceleration in depth fromfocus-based image acquisition (FFT), fast Fourier transform (FFT), and local deformation matching for HALCON. HALCON 11 will bring users a faster machine vision experience.

4. Machine learning

The latest version, Halcon 13, adds machine learning capabilities to further enhance the use area of Halcon.

5. Other new features

1) Aztec code recognition;

2) Micro QR code recognition;

3) Automatically select features for classification;

4) Efficient coding analysis using HDevelop functional evaluation tool;

5) Support Mac OS X 10.7 control system;

6) Revise the HALCON/C++ interface;

7) Rapid visualization of 3D data;

8) Telecentric lens stereo vision;

9) Improve camera calibration technology;

10) HDevelop OCR assistant with practice file explorer;

11) GS1 terminology for 1D and 2D code recognition;

12) Serialized HALCON/. .NET and HALCON/C++;

13) Easy-to-use measuring tools;

14) Support JPEG XR and others.

5. MATLAB-related toolbox

Image Processing Toolbox

Computer Vision System Toolbox

Image Acquisition Toolbox

Some other toolboxes:

A

1。 Name: Machine Vision Toolbox. （Machine Vision Toolbox ）

2。 Developer: Peter Corke. Release Date: 1999

3。 Institution: Head of the Robotics and Automation Research Group, Brisbane, Australia.

4。 Download interface:

http://www.cat.csiro.au/cmst/staff/pic/vision-tb.html

5。 Introduction: This toolbox is a collection of programs in the authors' research areas: photometry,

Image measurement, colorimetry, filtering, feature extraction, image reading and writing, filtering, segmentation, etc. Combining Matlab and workstations can provide convenient support for machine vision algorithm research.

B

1。 Name: Auditory Toolbox. （Auditory Toolbox）

2。 Developer: Malcolm Slaney. Release date: 1998

3。 Institution: Interval Research Corporation, Canada.

4。 Download interface:

http://rvl4.ecn.purdue.edu/~malcolm/interval/1998-010/

5。 Introduction: The Auditory Toolbox implements a typical auditory model, and it can also help you validate your auditory model hypothesis and reveal how the human auditory system perceives sound.

C

1。 Name: A physiological simulation benchmark experiment (PHYSBE)

2。 Developer: Kevin Kohrt. Release date: Theoretical model proposed in 1966, available for download in 1999.

3。 Institution: ____.

4。 Download interface:

http://www.mathworks.com/products/demos/simulink/physbe/

5。 Introduction: A classic human reincarnation system simulation model that can be used to simulate oxygen, nitrogen, nutrients, heat, and chemical tracers in the bloodstream.

D

1。 Name: MetMAP is a Matlab Toolbox for Metabolical Modeling, Analysis and oPtimization

2。 Developer: Julio Vera. Release date: 2003.

3。 Institution: La Laguna University, Spain.

4。 Download interface: http://webpages.ull.es/users/sympbst/pag_ing/pag_metmap/index.htm

http://webpages.ull.es/users/sympbst/pag_ing/pag_metmap/Downloads.htm

5。 Introduction: We will help you study a metabolic system from the perspective of a theoretical model and analyze the structure and kinetic properties of metabolic pathways. In addition, an interference interference measure can be optimized designed.

E

1。 Name: BrainStorm (BrainStorm is an integrated Matlab Toolbox dedicated toMagnetoencephalography (MEG) and Electroencephalography (EEG) datavisualization and processing. ）

2。 Developer: Felix Darvas. Updated: January 2005!

3。 Institution: USC, Los Angeles.

4。 Download interface: http://neuroimage.usc.edu/brainstorm/Downloads.php

Please log in with my account. Username:tenhospital@hotmail.comPassword: UsPQ7t5G

5。 Introduction: Ability to visualize and analyze electroencephalography and magnetoencephalography. It is a very important shared piece in this regard.

6. Maxtor Image library in Canada

The MIL software package is a hardware-independent, scale-based 32-bit image library. It has a complete set of instructions for image processing and special manipulation, including: spot analysis, image calibration, calibration, two-dimensional data reading and writing, measurement, pattern recognition and optical symbol recognition and manipulation. It also supports basic graphics devices. MIL is capable of handling binary, grayscale, or color images.

This software package is designed for rapid application development and ease of use. It has a fully transparent governance system that follows virtual data object manipulation rather than physical data object manipulation, and promises platform-independent applications. This means that a MIL application can run on any VESA-compatible VGA board or Matrox image board in different environments (Win98/Me/NT/2000). MIL uses the concept of the system to identify hardware boards, and a single application can control more than one hardware board. MIL can run on the host alone, but it is more efficient to use dedicated acceleration Matrox hardware.

The MIL package has the following features:

· Get up to 16-bit grayscale or color plots

· Works with 1, 8, 16, and 32-bit integer or floating-point images

· According to the manipulation of the color image, each layer of the color image will be processed independently. Color processing is not supported for statistical analysis, speckle analysis, measurement, pattern matching, optical symbol recognition, and code manipulation.

· Displays 1, 8, or 16-bit grayscale or color images (if supported by the platform)

7. eVision machine vision software

The eVision machine vision software package is a set of machine vision software development SDK launched by the Belgian company euresys, compared with other machine vision development packages such as cognex visionlibrary and Matrox Imaging library, it seems to provide more options than these two in terms of SDK function classification, and this feels that Matrox Imaging library is particularly poorly done and looks particularly pale.

All the code of the evision machine vision software development package is optimized by mmx instructions, the processing speed is very fast, and it feels like a fight with Intel's IPP (of course, it is still slightly inferior to IPP, but after all, this is Intel's own development), but it provides much more machine vision functions than IPP, such as OCR, OCV, image quality detection based on image comparison, Barcode and MatrixCode recognition. And my machine vision software development in the past two years is basically based on evision development.

8. HexSight

HexSight, developed by Adept, is a high-performance, comprehensive vision software development package that provides machine vision underlying functions for uncluttered, reliable, and correct positioning and inspection of parts. Its powerful locator tool accurately identifies and locates objects, regardless of whether they are rotated or change in size scale. HexSight delivers reliable results even in the harshest operating environments.

9. RVB machine vision and image processing algorithm software

Levy Machine Vision Application Software Development Kit (Real ViewBench, abbreviated as RVB) is a professional machine vision and image processing algorithm software package dedicated to the field of automation, and is a highly competitive and priced professional algorithm software package in the machine vision industry.