Skip to content
体验新版
项目
组织
正在加载...
登录
切换导航
打开侧边栏
Pinoxchio
apollo
提交
c2153ccb
A
apollo
项目概览
Pinoxchio
/
apollo
与 Fork 源项目一致
从无法访问的项目Fork
通知
2
Star
0
Fork
0
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
0
列表
看板
标记
里程碑
合并请求
0
Wiki
0
Wiki
分析
仓库
DevOps
项目成员
Pages
A
apollo
项目概览
项目概览
详情
发布
仓库
仓库
文件
提交
分支
标签
贡献者
分支图
比较
Issue
0
Issue
0
列表
看板
标记
里程碑
合并请求
0
合并请求
0
Pages
分析
分析
仓库分析
DevOps
Wiki
0
Wiki
成员
成员
收起侧边栏
关闭侧边栏
动态
分支图
创建新Issue
提交
Issue看板
前往新版Gitcode,体验更适合开发者的 AI 搜索 >>
提交
c2153ccb
编写于
9月 17, 2017
作者:
J
Jun Zhu
提交者:
FangzhenLi-hust
9月 17, 2017
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
[perception] modify document on tracking (#1795)
上级
fbf490ac
变更
1
隐藏空白更改
内联
并排
Showing
1 changed file
with
15 addition
and
15 deletion
+15
-15
modules/perception/README.md
modules/perception/README.md
+15
-15
未找到文件。
modules/perception/README.md
浏览文件 @
c2153ccb
...
...
@@ -302,11 +302,10 @@ HM Object Tracker
The HM object tracker is designed to track obstacles detected by the
segmentation step. In general, it forms and updates track lists by
associating current detections with existing track lists, deletes the
old track lists if it no longer persists, and spawns new track lists if
it identifies new detections. The motion state of the updated track
lists will be estimated after association. In HM object track, the
old track lists if it no longer persists, and spawns new track lists if new detections are identified. The motion state of the updated track
lists will be estimated after association. In HM object tracker, the
Hungarian algorithm is used for detection-to-track association, and a
Robust Kalman Filter is adopted for motion
and velocity
estimation.
Robust Kalman Filter is adopted for motion estimation.
### Detection-to-Track Association
...
...
@@ -317,20 +316,21 @@ detection-to-track matching with minimum cost (distance).
**Computing Association Distance Matrix**
In the first step, an association distance matrix is established. The
distance between a given detection and track is calculated according to
distance between a given detection and
one
track is calculated according to
a series of association features including motion consistency,
appearance consistency, etc. Some features used in HM tracker’s distance
computing are shown as below:
|location_distance |Evaluating motion consistency |
|--------------------- |-----------------------------------|
|direction_distance |Evaluating motion consistency |
|bbox_size_distance |Evaluating appearance consistency |
|point_num_distance |Evaluating appearance consistency |
|histogram_distance |Evaluating appearance consistency |
|Association Feature Name |Description |
|-------------------------|----------------------------------|
|location_distance |Evaluating motion consistency |
|direction_distance |Evaluating motion consistency |
|bbox_size_distance |Evaluating appearance consistency |
|point_num_distance |Evaluating appearance consistency |
|histogram_distance |Evaluating appearance consistency |
Besides, there are important parameters of distance weights which are
used for combining the above-mentioned
distance
features into a final
Besides, there are
some
important parameters of distance weights which are
used for combining the above-mentioned
association
features into a final
distance measurement.
**Bipartite Graph Matching via Hungarian Algorithm**
...
...
@@ -339,7 +339,7 @@ Given the association distance matrix, as shown in figure 5, Apollo
constructs a bipartite graph and uses Hungarian algorithm to find the
best detection-to-track matching via minimizing the distance cost. It
solves the assignment problem within O(n
\^
3) time complexity. To boost
its computing performance, Hungarian algorithm is implemented after
its computing performance,
the
Hungarian algorithm is implemented after
cutting original bipartite graph into subgraphs, by deleting vertices
with distance greater than a reasonable maximum distance threshold.
...
...
@@ -390,7 +390,7 @@ A high-level workflow of HM object tracker is given in figure 6.
<div
align=
center
>
Figure 6 Workflow of HM Object Tracker
</div>
1) Construct tracked objects and transform them into world coordinates.
1) Construct t
he t
racked objects and transform them into world coordinates.
2) Predict the states of existing track lists and match detections to
them.
...
...
编辑
预览
Markdown
is supported
0%
请重试
或
添加新附件
.
添加附件
取消
You are about to add
0
people
to the discussion. Proceed with caution.
先完成此消息的编辑!
取消
想要评论请
注册
或
登录