当前位置:网站首页>[TA frost wolf \u may- hundred talents plan] 1.2.3 MVP matrix operation
[TA frost wolf \u may- hundred talents plan] 1.2.3 MVP matrix operation
2022-07-01 03:47:00 【zczplus】
【TA- Frost Wolf _may-《 Hundred people plan 》】1.2.2 Matrix computing
1.2.3 MVP Matrix computing
1.2.3.1 Learning goals
- take 3D The object is transformed into 2D Plane
- Prepare for the use of various spaces
1.2.3.2 MVP Definition of matrix
MVP The matrices are :
- Model Model
- View Observe
- Projection Projection
Five coordinates :
Vertex coordinates start in local space (Local Space), This is called local coordinates (Local Coordinate), It will then change to world coordinates (World Coordinate), Observation coordinate (View Coordinate), Cut coordinates (Clip Coordinate), And finally in screen coordinates (Screen Coordinate) The form ends .
1.2.3.3 M: Model space to world space
Model space takes itself as the origin , World space has an origin independent of the model itself .
There are three steps to transform from model space to world space ( The order cannot be changed ):
- The zoom
- rotate
- translation
The corresponding matrix transformation is carried out in turn to obtain the transformation matrix .
M Representation of matrix ( The rotation here only includes the rotation around y The rotation of the shaft , Do you need to add information about other axes ?):
1.2.3.4 V: From world space to visual space
Visual space : A space coordinate system centered on a camera .
From world space to camera space
↓
Pan the entire viewing space , Make the camera origin coincide with the elder martial sister's coordinate origin , And make the coordinate axes coincide
↓
The camera rotates first in world space , Retranslated
↓
In order to make the camera coincide with the world coordinates , Using inverse transformation
The transformation process :
- Pan the world space
- Rotate world space
- z The component is reversed ( The reason of left-handed coordinate system )
According to the above method, the transformation matrix is obtained by matrix transformation
V A matrix representation of a matrix :
1.2.3.5 P: Visual space to crop space
Be careful :
- Not a real projection , Prepare for projection
- Purpose : Determine whether the vertex is within the scope of the courseware
- P matrix : Yes x, y, z Scale the component , use w The component is the range value . If x,y,z It's all in range , Then the point is in the clipping space .
There are two ways of projection : - Perspective projection
- Orthographic projection
1.2.3.6 P: Perspective projection
P A matrix representation of a matrix :
The meaning of each parameter of perspective projection is shown in the figure below :

Use the above P matrix , The coordinates of the transformed visual cone are shown in the figure :
It can be obtained. , If a vertex is within the visual cone , Then its transformed coordinates must meet :x,y,z Both in -w To w Within the scope of .
1.2.3.7 P: Orthographic projection
Relevant schematic diagram of orthogonal projection and its parameter interpretation :
Orthographic projection P A matrix representation of a matrix :
summary
Model space 、 World space 、 The difference of visual field space :
- Model space is centered on the model itself , The origin of the coordinate axis is the origin of the coordinate , By scaling 、 rotate 、 translation **(Model matrix )** Set the model coordinates (Model Coordinate) Change to world coordinates (World Coordinate)
- World space is centered on the whole scene , By translation and rotation **(View matrix )**, as well as z The inverse of the component will set the world coordinates (World Coordinate) Convert to observation coordinates (View Coordinate)
- The visual field space takes the camera as the origin center , Through the projection matrix **(Projection matrix )** Observe the coordinates (View Coordinate) To crop coordinates (Clip Coordinate)
边栏推荐
- jeecgboot输出日志,@Slf4j的使用方法
- Quickly filter data such as clock in time and date: Excel filter to find whether a certain time point is within a certain time period
- 【TA-霜狼_may-《百人计划》】1.3纹理的秘密
- Ouc2021 autumn - Software Engineering - end of term (recall version)
- TEC: Knowledge Graph Embedding with Triple Context
- 318. 最大单词长度乘积
- 【EI检索】2022年第六届材料工程与先进制造技术国际会议(MEAMT 2022)重要信息会议网址:www.meamt.org会议时间:2022年9月23-25日召开地点:中国南京截稿时间:2
- Jeecgboot output log, how to use @slf4j
- Idea plug-in backup table
- pytorch nn. AdaptiveAvgPool2d(1)
猜你喜欢

Review column - message queue

复习专栏之---消息队列

Valentine's Day is nothing.

衡量两个向量相似度的方法:余弦相似度、pytorch 求余弦相似度:torch.nn.CosineSimilarity(dim=1, eps=1e-08)

【TA-霜狼_may-《百人计划》】1.3纹理的秘密

Appium自动化测试基础 — APPium基本原理

RSN:Learning to Exploit Long-term Relational Dependencies in Knowledge Graphs

Access denied for user ‘ODBC‘@‘localhost‘ (using password: NO)

jeecgboot输出日志,@Slf4j的使用方法

SEM of C language_ Tvariable type
随机推荐
访问阿里云存储的图片URL实现在网页直接预览略缩图而不直接下载
FCN full Convolution Network Understanding and Code Implementation (from pytorch Official Implementation)
[深度学习]激活函数(Sigmoid等)、前向传播、反向传播和梯度优化;optimizer.zero_grad(), loss.backward(), optimizer.step()的作用及原理
10、Scanner. Next() cannot read spaces /indexof -1
Golang multi graph generation gif
【TA-霜狼_may-《百人计划》】2.2 模型与材质空间
Ouc2021 autumn - Software Engineering - end of term (recall version)
【TA-霜狼_may-《百人计划》】2.1 色彩空间
衡量两个向量相似度的方法:余弦相似度、pytorch 求余弦相似度:torch.nn.CosineSimilarity(dim=1, eps=1e-08)
Explain spark operation mode in detail (local+standalone+yarn)
Take you through a circuit board, from design to production (dry goods)
Addition without addition, subtraction, multiplication and division
Pyramid Scene Parsing Network【PSPNet】论文阅读
165. compare version numbers
Idea plug-in backup table
【TA-霜狼_may-《百人计划》】1.4 PC手机图形API介绍
208. implement trie (prefix tree)
IPv4和IPv6、局域网和广域网、网关、公网IP和私有IP、IP地址、子网掩码、网段、网络号、主机号、网络地址、主机地址以及ip段/数字-如192.168.0.1/24是什么意思?
RSN:Learning to Exploit Long-term Relational Dependencies in Knowledge Graphs
复习专栏之---消息队列