Daisy : The Video Capture Library of the LV-SDS

Daisy is a Software Development Kit for programmers who want to integrate

PCI Frame Grabbers into their own applications. 

Features

- C++ Class library (16bit DLL) for Windows 3.x and Windows 95

- C++ Class library (32bit DLL) for Windows NT4.0

- Supports Borland C++ 5.x for both Daisy 16bit and 32bit

- Supports Microsoft Visual C++ 1.5x for Daisy 16bit

- Supports Microsoft Visual C++ 4.x for Daisy 32bit

- Realtime video capture supported

- Twain driver included

- Camera Editor to define non standard cameras

- Plug and Play connector handling

- Reference Manual as Windows Help file and printable Word file

- Automatic firmware update included in library 

General Description

Camera configuration

Predefined setting of camera configurations for the most popular cameras are available.

A camera definition is composed out of other predefined settings which are

VIDEO TYPE, COLOR CODING, ASPECT RATIO, VIDEO TIMING, SIGNAL COUPLING and 

SYNCHRONISATION. Each of these items is composed by specific settings, which can be 

adapted by the programmer. Like this, a user can compose his own camera setting for 

special cameras.

 
Leutron Vision offers a Camera Editor which assists you in defining your own cameras.

The Camera Editor produces a datafile with the detailed camera description. 

Daisy itself will read during startup this camera description and set the 

hardware according to the definitions.

Camera connection and activation

Depending on the camera cable and the actual Hardware, several cameras can be 

connected to the Framegrabber at the same time. There are functions available 

to control the proper use of the available physical connectors and for activating

one of the connected cameras. 

Video input parameters

Depending on the COLOR CODING, the following items can be set: 
 

 - Brightness  
 - Contrast  
 - Hue  
 - Saturation  
 - InputLut 
 

For all cameras the following items can be set: 

- Mirror  
- Scaling or Clipping 

Region of Interest configuration (ROI)

The coordinates of the live video image and its pixeldepths on screen can be defined

trough the ROI class. The ROI handling covers the whole video memory area,

therefore an image may be acquired in the off screen area too. 

The Windows API covers only the on screen area of the video memory, 

the ROI class offers a possibilty for the user to access images outside

the memory controlled by Windows.

Utility functions

- Automatic palette handling for greyscale images (display set 256 colors)  
- Synchronisation to camera or monitor timing  
- Optocoupler control  
- Version control of the library  
- Automatic firmware update 

Class Overview

LvCameraNodeObjects belonging to the LvCameraNode class represent the input

equipment used to transfer in the video memory images from the outer world. 
  

LvConnectorEach Camera can be connected to the grabber through an Input Connector;

depending on the type of connector used, it will be possible to connect to the

grabber a different number of camera types. 

Each available connector is represented by a Handle, that uniquely identifies

the connector in the system; the user application, when connecting input devices

should specify where the equipment will be physically connected in order to allow

Daisy to perform the correct settings. 
  

LvGrabberNodeObjects belonging to the LvGrabberNode class represent a video

grabber installed in the system. 

More than one grabber can be installed in the system, Daisy will provide

different LvGrabberNode objects and each of them will act only on the physical

board it represents.

Handles and connections

External equipment, like monitors and cameras, can be connected to a grabber;

in its base concept a grabber supports more than one input and more than one

output connections; so, in order to ask Daisy to operate on the correct

equipment an application has to use Handles. 

All the handles representing external equipment are of type HCONNECTION;

for easier code readibility are also defined the HCAMERA and HMONITOR typedefs.

Through a handle the application can still have a reference to the object it

represents, so it can directly use its interface.

Monitors

More than one monitor could be connected to the hardware, of course if the hardware

itself can support them; the actual implementation of the hardware can support

only one monitor connection, so these methods are kept for future use;

to maintain future compatibility also the user application are recommended to

use the query methods instead of assuming a one-monitor system.

Actually Daisy takes care of connecting the monitor and of activating it during

its initialization, so the application has not to deal with this system setting. 

Input Connectors

In order to allow the end user to choose the input where the Camera will be

connected to, Daisy offers query methods to retrieve data about the connectors

still free in the system; these information can be then presented to the end user

that will choose the desired input; the connector descriptor maintains also a

ASCII string that will match the description on the physical cable

(i.e RGB1, YC_C1 and so on). 

Cameras

The LvGrabberNode allows the user application to connect a Camera to the grabber,

activate or deactivate it. Each camera is represented in Daisy by a Handle that

uniquely identifies it in the system; the user can use this handle to refer to

a specific input equipment when accessing other grabber methods. Knowing its handle,

a user application can retrieve a pointer to each LvCameraNode object. 
  

LvROIA Region Of Interest, also called ROI, is a rectangular part of video

memory formatted with a specific color coding, that can be different from the

actual output formatting, depending on the use the application wants to do of this

memory; for example,

an incoming image is placed in a ROI and also the output screen is a ROI.

The user application can retrieve from Daisy information about the ROI currently active,

those related with external equipment connected to the grabber, so that it can access

that part of memory interpreting in the right way the data stored in it.

A ROI is defined by its position in the Video memory, by its dimensions and by the

color coding used to generate it. A user application can generate and use its own ROI,

in order to store in the Video Memory application related data.

A ROI is defined in the system by the LvROIInfo structure; since exists a strict

relationship among the content of its fields, Daisy provides the LvROI class that can

be used by the application to manage the LvROIInfo data in order to be sure that the 

content of their fields is coherent. All the LvROIInfo attributes remain 'public', so

that the reading access to them can be a little bit faster, but the user application

is strongly recommended to use the methods provided by LvROI to set those attributes in

order to have them coherent.

DirectCapture, the Video Capture Library 

DirectCapture is a Software Development Kit for programmers who want to

integrate Leutron Vision PCI Frame Grabbers into their own applications. 

Features

- C library (32bit DLL) for Windows NT4.0 and Windows 95  
- implemented with Microsoft's Direct Draw  
- language independent  
- sample code for Microsoft Visual C, Visual Basic, Borland C and Delphi  
- Realtime video capture supported  
- Camera Editor to define non standard cameras  
- Plug and Play connector handling  
- Reference Manual as Windows Help file and printable Word file  
- Automatic firmware update included in library 

General Description

Introduction

DirectCapture is a language independent software library for the control of

Leutron Vision image processing boards. DirectCapture is a marriage of Leutron's

hardware related DAISY software library and Microsoft's Direct Draw with the goal

of simplifying the development of user applications.

DirectCapture uses a window paradigm to facilitate the control of the image processing

boards. The library allows the creation of On Screen Window and Off Screen Window

objects. The On Screen Window objects enable the display of live video on the desktop

from the image processing boards with scaling, clipping, mirroring, and overlaid text

and graphics. The Off Screen Window objects provide many of these features while 

directing the video stream directly to system memory instead of the windows desktop.

On Screen Windows

The On Screen Window object is provided for the display of live video on the 

Widows NT desktop. The On Screen Window objects is created by passing the 

handle (HWND) of a Windows NT window into the LvDcCreateOnScreenWin function.

DirectCapture takes the Windows NT window handle and subclasses it to take control

of the window. Subclassing is the technique where by all windows message generated

for a window will now be handled by the software which subclassed the window, in

effect taking over control of the window. The LvDcCreateOnScreenWin will pass back

a handle (HDCVIDEOWIN ) of the newly created On Screen Window object. This handle

must be used in all future to calls to the DirectCapture library to manipulate the

On Screen Window. 

Off Screen Windows

The Off Screen Window object is provided to direct live video to system memory.

The Off Screen Window Object is created by calling the LvDcCreateOffScreenWin function

which will pass back a handle to the newly created window (HDCVIDEOWIN). Like the On

Screen Windows all function calls which manipulate the Off Screen Window require the

handle returned from the create call.

DirectCapture Components

DirectCapture consists of two separate DLL's, the DSYCTL.DLL and the CAMDLG.DLL.

The DSYCTL.DLL contains all the functions for creating and using the On and Off 

Screen Window objects and the CAMDLG.DLL contains the routines to select grabbers

and cameras to connect to the window objects. The C source code for the CAMDLG.DLL

library is provided so that the users can customize the dialog as they wish. 

Camera configuration

Predefined setting of camera configurations for the most popular cameras are

available.Leutron Vision offers a Camera Editor which assists you in defining your

own cameras. The Camera Editor produces a datafile with the detailed camera description.

DirectCapture itself will read during startup this camera description and set the

hardware according to the definitions. 

Function Overview

The following is a list of features that are supported by the On and Off Screen 

windows and a description of each.

- Scaling

  Both the On and Off Screen Windows support scaling. When scaling is enabled the cameras

  field of view will be scaled to fit the size of the window on screen or in memory.

  For example, if the size of the window is set to 150 x 150 pixels and a video source

  with a field of view of 576 x 452 is being used the entire image will be scaled down

  to 150 x 150 for display or capture.

- Clipping

  Both the On and Off Screen Windows support clipping.With clipping only the portion of

  the video that will fit within the size of window will be displayed or captured 

  to memory. For example, if the size of the widow is set to 150 x 150 pixels and a 

  video source with a field of view of 576 x 452 is being used only a 150 x 150 pixel

  window will be displayed or captured.

- Mirroring

  Both the On and Off Screen Windows support mirroring along the vertical, horizontal

  or both axis.

- Overlay Text

  The On Screen Window supports overlaid text.Overlaid text allows the user to 

  overlay individual character strings on the live video. The font, size, color and

  position of the character string can be specified and changed through the

  DirectCapture library.

- User Masks

 
  Both the On and Off Screen Windows support user masks. A user mask is a one bit

  plane deep bit map that can be overlaid on the video. The mask can be created and

  manipulated via standard WIN32 function calls. Also the user can specify that the 

  mask should be stretched to match the size of the video.

- Copy to Clipboard

  Both window types support the copying of the video to the Window NT clipboard.

  For On Screen windows when the image is copied to the clipboard the current video card

  image format is used to determine the format of the image, i.e. when the desktop is at

  1280 x 1024 by 8 bits the image will be copied to the clipboard in 8bit color format.

  For off screen windows the format of the image copied to the clipboard can be set 

  using the LvDcSetColorFormat function. If there is a user mask set and enabled for 

  either window it will also be copied to the clipboard. Currently overlaid text is not

  copied to the clipboard.

  
- Get Image As Bitmap 
  

  Both windows support the retrieval of the video as a windows BMP. In the case of 

  the On Screen Windows the color format of the BMP will match the current Windows NT

  desktop setting. For Off Screen Windows the format of the image retrieved can be set

  using the LvDcSetColorFormat function. As with LvDcCopyToClipboard any user mask that

  has been enabled will be copied but overlaid text will not.  

- Adjusting Picture Controls 
  

  Both windows support the adjustment of brightness, contrast, hue and saturation.

- Events and Triggers

  On Screen Window objects can be configured to respond to Windows NT user interface

  actions and Off Screen Window objects can be configured to trigger on certain On

  Screen Widow actions.

On Screen Window objects can be configured to enter the live video mode when the

video window receives the desktop focus or only when it is the top most window.

Like wise it can be configured to freeze the video when it loses focus or is covered

by other windows. Also Off Screen Windows can be configured to go into live mode after

an On Screen Window freeze.  

- Window Association 

  Off Screen Window objects can be associated with On Screen Window objects. This allows

  a Off Screen Window to be configured to track the resize mode, color format, mirror 

  mode, live state, and freeze state of a selected On Screen Window . This will give the

  user the ability to have the video in system memory track the on screen video as 

  closely as possible.