Halcon 6.1 C++编程手册：入门与实战指南

需积分: 5 133 浏览量更新于2024-07-17 收藏 614KB PDF 举报

Halcon 6.1 C++ 用户手册是由 MVTec Software GmbH 发布的一份官方文档，专为那些想要在自己的 C++ 项目中集成 HALCON 6.1.4 图像分析工具的开发者设计。该手册详细介绍了如何有效地将 HALCON 的功能与 C++ 结合使用，帮助初学者和有经验的开发者深入了解并利用这一强大的机器视觉库。此版本的历史更新涵盖了从 HALCON 5.1 到 6.1.4 的过程，包括了 1997 年 7 月的初始版本，直到 2005 年 2 月的 6.1.4 版本。MVTec Software GmbH 是该软件的版权所有者，强调了对知识产权的保护，禁止未经许可的复制、存储、检索或任何形式的传输。手册中包含了重要的版权信息，指出 1997 至 2005 年间的内容受法律保护，并提到了与微软相关的商标，如 Microsoft Windows、Windows NT、Windows 2000、Windows XP 和 Visual C++，这些都是 MVTec Software GmbH 所认可的商标或注册商标。 Halcon 6.1 C++ 用户手册的核心内容可能包括以下几个部分： 1. **入门指南**：为新用户提供了安装和配置 HALCON 6.1.4 的步骤，以及如何创建基本的 C++ 程序以调用 HALCON 函数。 2. **API 文档**：详尽地解释了 HALCON 在 C++ 中的接口，包括函数定义、参数类型、返回值和用法示例，使开发者能快速理解和使用 HALCON 的各种功能，如图像处理、物体识别、特征检测等。 3. **编程示例**：提供了实际的代码片段和项目案例，展示如何在 C++ 中实现常见的图像分析任务，以便于开发者在实践中学习和应用。 4. **错误处理和调试**：针对可能出现的问题和错误，手册可能包含如何诊断和解决在 C++ 环境中使用 HALCON 时遇到的技术问题。 5. **兼容性和系统需求**：明确列出了 HALCON 6.1.4 对 C++ 编译器（如 Visual C++）和操作系统（如 Windows）的要求，以及可能的硬件支持。 6. **升级和维护**：指导用户如何从早期版本升级到 6.1.4 版本，以及如何获取后续的补丁和更新。 7. **许可证条款**：确保开发者了解 HALCON 的使用许可协议，如商业使用限制、分销条款等。 Halcon 6.1 C++ 用户手册是开发人员不可或缺的参考资料，它不仅提供了一个完整的工具集，还为在 C++ 环境中高效使用 HALCON 提供了详尽的教程和实践支持。无论是初学者还是高级开发者，都可以从中找到所需的信息来提升自己的机器视觉项目能力。

Chapter 2

Basics of the HALCON/C++ Interface

In fact, the HALCON/C++ interface provides two different approaches to use HALCON’s func-

tionality within your C++ program: a procedural and an object-oriented approach. The proce-

dural approach corresponds to calling HALCON operators directly as in C or HDevelop, e.g.:

Hobject original_image, smoothed_image;

::read_image(&original_image, "monkey");

::mean_image(original_image, &smoothed_image, 11, 11);

Note that, in comparison with HALCON/C or HDevelop, the operator names are preﬁxed with

a double colon. In the following, we will leave out this preﬁx when referencing operators in the

text.

In addition to the procedural approach, HALCON/C++ allows to call HALCON operators in

an object-oriented way, i.e., via a set of classes. For example, the code from above can be

“translated” into:

HImage original_image("monkey"), smoothed_image;

smoothed_image = original_image.MeanImage(11, 11);

This simple example already shows that the two approaches result in clearly different code: Be-

sides the different operator names (procedural: small letters and underscores; object-oriented:

capitals), the operator calls differ in the number and type of parameters. Furthermore, func-

tionality may be available in different ways; for example, images can be read from ﬁles via a

constructor of the class HImage. In general, we recommend to use the object-oriented approach.

Note, however, that HDevelop can export programs only into procedural C++ code. Section 2.4

shows how to combine procedural with object-oriented code.

In the following sections, we take a closer look at various issues regarding the use of the HAL-

CON/C++ interface; chapter 3 describes the provided classes in more detail.

2.1 Calling HALCON Operators

How a HALCON operator can be called via the HALCON/C++ interface is described in detail

in the HALCON/C++ reference manual. As an example, ﬁgure 2.1 shows parts of the entry for

the operator mean image.

Below, we take a closer look at the parameters of an operator call, describe how to call operators

via classes, and explain another special HALCON concept, the tuple mode.

6 CHAPTER 2. BASICS OF THE HALCON/C++ INTERFACE

Herror ::mean image ( Hobject Image, Hobject *ImageMean,

const HTuple &MaskWidth, const HTuple &MaskHeight )

HImage HImage::MeanImage ( const HTuple &MaskWidth,

const HTuple &MaskHeight ) const

HImageArray HImageArray::MeanImage ( const HTuple &MaskWidth,

const HTuple &MaskHeight ) const

Image (input object) . . (multichannel-)image(-array) ;

Hobject: HImage(Array)

( byte /

int2 / uint2 / int4 / real / dvf )

ImageMean (output object) . . (multichannel-)image(-array) ;

Hobject * : HImage(Array)

(

byte / int2 / uint2 / int4 / real

/ dvf )

MaskWidth (input control) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . extent.x ;

HTuple.long

MaskHeight (input control) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . extent.y ;

HTuple.long

Figure 2.1: The head and parts of the parameter section of the reference manual entry for

mean image.

2.1.1 A Closer Look at Parameters

HALCON distinguishes two types of parameters: iconic and control parameters. Iconic pa-

rameters are related to the original image (images, regions, XLD objects), whereas control

parameters are all kinds of alphanumerical values, such as integers, ﬂoating-point numbers, or

strings.

A special form of control parameters are the so-called handles. A well-known representative of

this type is the window handle, which provides access to an opened HALCON window, e.g., to

display an image in it. Besides, handles are used when operators share complex data, e.g., the

operators for shape-based matching which create and then use the model data, or for accessing

input/output devices, e.g., frame grabbers. Classes encapsulating handles are described in detail

in section 3.2.3.

Both iconic and control parameters can appear as input and output parameters of a HALCON

operator. For example, the operator mean image expects one iconic input parameter, one iconic

output parameter, and two input control parameters (see ﬁgure 2.1); ﬁgure 2.2 shows an oper-

ator which has all four parameter types. Note how some parameters “disappear” from within

the brackets if you call an operator via a class; this mechanism is described in more detail in

section 2.1.2.

An important concept of HALCON’s philosophy regarding parameters is that input param-

eters are not modiﬁed by an operator. As a consequence, they are passed by value (e.g.,

Hobject Image in ﬁgure 2.1) or via a constant reference (e.g., const HTuple &MaskWidth).

This philosophy also holds if an operator is called via a class, with the calling instance acting as

an input parameter. Thus, in the following example code the original image is not modiﬁed by

the call to MeanImage; the operator’s result, i.e., the smoothed image, is provided via the return

value instead:

HALCON / C++, 2005-02-01

2.1. CALLING HALCON OPERATORS 7

Herror ::find 1d bar code ( Hobject Image, Hobject *CodeRegion,

const HTuple &BarCodeDescr, const HTuple &GenericName,

const HTuple &GenericValue, HTuple *BarcodeFound,

HTuple *BarCodeElements, HTuple *Orientation )

HRegion HImage::Find1dBarCode ( const HTuple &BarCodeDescr,

const HTuple &GenericName, const HTuple &GenericValue,

HTuple *BarcodeFound, HTuple *BarCodeElements, HTuple *Orientation )

const

HRegion HBarCode1D::Find1dBarCode ( const HImage &Image,

const HTuple &GenericName, const HTuple &GenericValue,

HTuple *BarcodeFound, HTuple *BarCodeElements, HTuple *Orientation )

const

Image (input object) . . . . . . . . . . . . . . . . . . . . . . . . . . . image ;

Hobject: HImage

( byte / uint2 )

CodeRegion (output object) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . region ;

Hobject * : HRegion

BarCodeDescr (input control) . . . . barcode 1d-array ;

HTuple.const char *

/ long / double

GenericName (input control) . . . . . . . . . . . . . . . . . . . . . . . .string(-array) ;

HTuple.const char *

GenericValue (input control) . . . . . . . . . . . . . . . . . . number(-array) ;

HTuple.double

/ long

BarcodeFound (output control) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . integer ;

HTuple.long *

BarCodeElements (output control) . . . . . . . . . . . . . . . . . . . . .number-array ;

HTuple.double *

Orientation (output control) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . angle.rad ;

HTuple.double *

Figure 2.2: The head and parts of the parameter section of the reference manual entry for

find 1d bar code.

HImage original_image("monkey"), smoothed_image;

smoothed_image = original_image.MeanImage(11, 11);

In contrast to input parameters, output parameters are always modiﬁed, thus they must be passed

by reference. Note that operators expect a pointer to an already declared variable! For exam-

ple, when calling the operator Find1dBarCode as in the following lines of code, variables of

the class HTuple are declared before passing the corresponding pointers using the operator &.

HImage image("barcode/ean13/ean1301");

HBarCode1D barcode("EAN 13", 13, 13);

HRegion code_region;

HTuple isfound, elements;

code_region = barcode.Find1dBarCode(image, HTuple(), HTuple(),

&isfound, &elements, (HTuple*) _);

The above example shows two other interesting aspects of output parameters: When calling

operators via classes, one output parameter becomes the return value (see section 2.1.2 for

more details); in the example, Find1dBarCode returns the bar code region. Secondly, a spe-

cial variable called was speciﬁed for the last parameter of Find1dBarCode. This variable is

pre-declared; it tells the operator that it does not need to output the corresponding parameter.

HALCON 6.1.4

剩余61页未读，继续阅读

lxhkkll

粉丝: 0
资源: 7

Halcon 6.1 C++编程手册：入门与实战指南

halcon-11.0-windows.part1.rar

halcon-17.12.0.0-windows-images-deep-learning.exe

halcon-17.12.0.0-windows-runtime.exe

halcon-18.11下载

halcon-18.11.2.0

halcon-19.11.0.0 32位 dll文件

halcon-18.05.0.2-windows安装

C:\Program Files\MVTec\HALCON-10.0\include\cpp\HRegion.h(126,21): error C2059: 语法错误:“const”

ubuntu halcon

最新资源