-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathindex.html
executable file
·392 lines (353 loc) · 26 KB
/
index.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head><meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<meta name="viewport" content="width=800">
<meta name="viewport" content="width=device-width, initial-scale=1">
<meta name="generator" content="HTML Tidy for Linux/x86 (vers 11 February 2007), see www.w3.org">
<style type="text/css">
/* Color scheme stolen from Sergey Karayev */
a {
color: #1772d0;
text-decoration:none;
}
a:focus, a:hover {
color: #f09228;
text-decoration:none;
}
body,td,th,tr,p,a {
font-family: 'Lato', Verdana, Helvetica, sans-serif;
font-size: 14px
}
.hp-photo{ width:240px; height:240px; border-radius:240px; -webkit-border-radius:240px; -moz-border-radius:240px; }
strong {
font-family: 'Lato', Verdana, Helvetica, sans-serif;
font-size: 14px;
}
heading {
font-family: 'Lato', Verdana, Helvetica, sans-serif;
font-size: 24px;
}
papertitle {
font-family: 'Lato', Verdana, Helvetica, sans-serif;
font-size: 15px;
font-weight: 700
}
name {
font-family: 'Lato', Verdana, Helvetica, sans-serif;
font-size: 32px;
}
.one
{
width: 160px;
height: 160px;
position: relative;
}
.two
{
width: 160px;
height: 160px;
position: absolute;
transition: opacity .2s ease-in-out;
-moz-transition: opacity .2s ease-in-out;
-webkit-transition: opacity .2s ease-in-out;
}
.fade {
transition: opacity .2s ease-in-out;
-moz-transition: opacity .2s ease-in-out;
-webkit-transition: opacity .2s ease-in-out;
}
span.highlight {
background-color: #ffffd0;
}
</style>
<title>Xinjie Zhou</title>
<!--<link rel="stylesheet" type="text/css" href="/imgs/css" >-->
<!-- <link rel="icon" type="image/png" href="./images/zju.png"> -->
<link rel="icon" type="image/png" href="./images/hit.png">
<!-- <link rel="shortcut icon" type="image/vnd.microsoft.icon" href="./images/zju.png"> -->
<!-- <script type="text/javascript" async
src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.7/MathJax.js?config=TeX-MML-AM_CHTML">
</script>
<script type="text/x-mathjax-config">
MathJax.Hub.Config({
tex2jax: {inlineMath: [['$','$'], ['\\(','\\)']]},
});
</script> -->
</head>
<body>
<table width="800" border="0" align="center" cellspacing="0" cellpadding="0">
<tbody>
<tr>
<td>
<!--SECTION 1 -->
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20">
<tbody><tr>
<td width="68%" valign="middle">
<p align="center"><name>Xinjie Zhou (周新杰)</name></p>
I am a Master student (Sep. 2020 - Jul. 2022) in the <a href="http://sa.hit.edu.cn/">School of Astronautics</a>
at <a href="http://http://www.hit.edu.cn/">Harbin Institute of Technology</a>,
supervised by <a href="http://homepage.hit.edu.cn/gaohuijun"> Prof. Huijun Gao</a>.
I also obtained a B.Eng from <a href="http://en.hit.edu.cn/"> Harbin Institute Of Technology </a>.
</br></br>
<!-- <strong><font color="red">Looking for a 2022 PhD position in computer vision / robotics perception! Welcome to contact me! </font></strong> -->
</br>
</td>
<td align="right"> <img class="hp-photo" src="./images/zxj.jpg" style="width: 150; height: 210;"></td></tr>
</tbody>
</table>
<!--SECTION 2 -->
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20">
<tbody><tr>
<td>
<heading>Research Interests</heading>
<p align="justify">I'm interested in <strong>SLAM</strong> and robotics. Currently I'm working on <strong>vehicle localization technology based on multi-sensor fusion</strong>.
I hope to build high quality point cloud map and get robust and precised ego-motion estimation. My research goal is to help the robot to
realize <strong>robust and precise localization</strong> with the help of <strong>multi-sensor</strong>,including <strong>LiDAR,camera,IMU,GNSS</strong>.
<!--</br></br>-->
<!--<span class="highlight"><strong>Internship Position: </strong> If you're interested in ...</span> -->
</p>
</td></tr>
</tbody>
</table>
<!-- SECTION 3 -->
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20">
<tbody><tr>
<td><heading>Research Platform</heading>
</td>
</tr></tbody>
</table>
<!--SECTION 5 -->
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20">
<tbody><tr>
<td width="20%"><img src="./images/platform/platform.png" alt="PontTuset" width="250" style="border-style: none"></td>
<td width="80%" valign="middle">
<papertitle>An autonomous driving car equipped with LiDAR, camera, IMU and RTK</papertitle>
</p><p></p>
<p align="justify" style="font-size:13px"> <strong>Car Platform:</strong> Trumpchi</p>
<p align="justify" style="font-size:13px"> <strong>LiDAR:</strong> RS-LiDAR-32*1, RS-LiDAR-16*2</p>
<p align="justify" style="font-size:13px"> <strong>Monocular Camera:</strong> A5201M/CG50</p>
<p align="justify" style="font-size:13px"> <strong>Integrated Navigation System:</strong> PwrPak7</p>
<a href="./videos/platform/ad_car_demo.mp4">Autonomous Driving Demo Video</a>
</td>
</tr>
<!-- SECTION 3 -->
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20">
<tbody><tr>
<td><heading>Research Experience</heading>
</td>
</tr></tbody>
</table>
<!--SECTION 5 -->
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20">
<tbody><tr>
<td width="20%"><img src="./images/mapping/mapping_10s.gif" alt="PontTuset" width="250" style="border-style: none"></td>
<td width="80%" valign="top">
<p align="justify" style="font-size:13px"> <strong>------------------------------------------------------------------------------------------</strong> </p>
<papertitle>Mapping framework based on multi-sensor fusion with LiDAR, IMU and GNSS</papertitle>
</p><p></p>
<p align="justify" style="font-size:13px"> 2020.11-2021.02 </p>
<p align="justify" style="font-size:13px"> <strong>【数据预处理】:</strong> 对各种传感器数据进行<strong>时间戳同步</strong>，基于<strong>LiDAR</strong>与<strong>RTK</strong>外参将<strong>RTK</strong>位姿转换到<strong>LiDAR坐标系</strong>，作为位姿参考</p>
<p align="justify" style="font-size:13px"> <strong>【前端】:</strong> 维护一个固定帧数的<strong>局部地图</strong>，采用NDT对当前帧激光点云做<strong>scan-to-map</strong>的匹配，获取<strong>前端激光里程计</strong></p>
<p align="justify" style="font-size:13px"> <strong>【回环检测】:</strong> 基于手写<strong>Scan Context</strong>方法查找当前帧的候选回环帧，并求取粗略位姿初值，进而用ICP求取回环约束</p>
<p align="justify" style="font-size:13px"> <strong>【后端】:</strong> 利用<strong>G2O</strong>引入<strong>前端里程计帧间约束</strong>、 <strong>IMU帧间约束</strong>、 <strong>RTK对单帧的位置约束</strong>、 <strong>回环约束</strong>，对<strong>关键帧位姿</strong>进行优化</p>
<p align="justify" style="font-size:13px"> <strong>【建图】:</strong> 执行<strong>全局位姿图优化</strong>后，根据<strong>优化后的关键帧位姿</strong>拼接每个<strong>关键帧</strong>对应的的激光点云，得到完整的<strong>点云地图</strong></p>
<p align="justify" style="font-size:13px"> <strong>【验证】:</strong> 在公开的<strong>KITTI数据集</strong>和<strong>UrbanNav香港数据集</strong>上测试，本算法获取的<strong>点云地图</strong>可以很好地支持后续的<strong>定位需求</strong></p>
<p align="justify" style="font-size:13px"> <strong>【TO DO/DOING】:</strong></p>
<p align="justify" style="font-size:13px"> ① <strong>前端改进：</strong> </p>
<p align="justify" style="font-size:13px"> <strong>前端</strong>采用
<strong>多度量线性最小二乘ICP (MULLS-ICP)</strong>提取<strong>点、线、面特征</strong>，
提高构建的<strong>约束</strong>的合理性，并对<strong>动态物体</strong>进行一定程度滤波，
保证建图效果不受<strong>动态物体</strong>影响;
也可考虑引入用点云<strong>恢复深度信息</strong>的<strong>视觉特征点</strong>构成的<strong>约束</strong>。</p>
<p align="justify" style="font-size:13px"> ② <strong>退化问题处理：</strong></p>
<p align="justify" style="font-size:13px"> 对于关于$6$维优化变量$\boldsymbol{x}$的优化问题
${\rm arg}$ $\mathop{{\rm min}} \limits_{\boldsymbol{x}}||\boldsymbol{Ax} - \boldsymbol{b}||_{\boldsymbol{P}}^2$，
构建矩阵$\boldsymbol{H}=\boldsymbol{A}^\mathrm{T}\boldsymbol{P^{-1}A}$，
$\boldsymbol{g}=\boldsymbol{A}^\mathrm{T}\boldsymbol{P^{-1}b}$，
计算$\boldsymbol{H}^\mathrm{T}\boldsymbol{H}$的<strong>特征值</strong>${\lambda}_1,...,{\lambda}_6$与<strong>特征向量</strong>$\boldsymbol{v}_1,...,\boldsymbol{v}_6$，
对于小于<strong>某一经实验测定得到的阈值</strong>的<strong>特征值</strong>${\lambda}_1,...,{\lambda}_m (m \le 6)$，
构建如下矩阵：
<!-- </p> -->
<!-- <p align="justify" style="font-size:13px"> -->
<!-- $$\boldsymbol{V}_p = [\boldsymbol{v}_1,...,\boldsymbol{v}_m,0,...,0],$$
$$\boldsymbol{V}_u = [0,...,0,\boldsymbol{v}_{m+1},...,\boldsymbol{v}_6],$$
$$\boldsymbol{V}_f = [\boldsymbol{v}_1,...,\boldsymbol{v}_m,\boldsymbol{v}_{m+1},...,\boldsymbol{v}_6],$$ -->
$$
\begin{aligned}
&\boldsymbol{V}_p = [\boldsymbol{v}_1,...,\boldsymbol{v}_m,0,...,0]^\mathrm{T},\\
&\boldsymbol{V}_u = [0,...,0,\boldsymbol{v}_{m+1},...,\boldsymbol{v}_6]^\mathrm{T},\\
&\boldsymbol{V}_f = [\boldsymbol{v}_1,...,\boldsymbol{v}_m,\boldsymbol{v}_{m+1},...,\boldsymbol{v}_6]^\mathrm{T},\\
\end{aligned}
$$
<strong>退化方向</strong>$\boldsymbol{v}_1,...,\boldsymbol{v}_m$
使用<strong>预测值</strong>$\boldsymbol{x}_p$，
<strong>良态方向</strong>$\boldsymbol{v}_{m+1},...,\boldsymbol{v}_6$
使用<strong>优化问题</strong>求解出的<strong>更新量</strong>$\boldsymbol{x}_u$。
最终的<strong>解</strong>为：$$\boldsymbol{x}_f = \boldsymbol{x}_{p}^{’} + \boldsymbol{x}_{u}^{’}$$
其中，$\boldsymbol{x}_{p}^{’} = \boldsymbol{V}_f^{-1}\boldsymbol{V}_p\,\boldsymbol{x}_p，
\boldsymbol{x}_{u}^{’} = \boldsymbol{V}_f^{-1}\boldsymbol{V}_u\,\boldsymbol{x}_u$
</p>
<strong></strong>
</td>
</tr>
<!--SECTION 5 -->
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20">
<tbody><tr>
<td width="20%"><img src="./images/localization_initialization/localization_init_on_map.gif" alt="PontTuset" width="250" style="border-style: none"></td>
<td width="80%" valign="top">
<p align="justify" style="font-size:13px"> <strong>------------------------------------------------------------------------------------------</strong> </p>
<papertitle>Global pose initialization on pre-built point cloud map</papertitle>
</p><p></p>
<p align="justify" style="font-size:13px"> 2021.03 </p>
<p align="justify" style="font-size:13px"> <strong>【基于GNSS】:</strong> </p>
<p align="justify" style="font-size:13px"> 记录<strong>建图时</strong><strong>第一帧</strong>有效点云数据对应的<strong>GNSS数据</strong>，在<strong>地图上初始化</strong>时通过调用<strong>GeographicLib库</strong>将<strong>该GNSS数据</strong>设为<strong>原点</strong>，<strong>初始化</strong>时的GNSS数据通过<strong>GeographicLib库</strong>将<strong>GNSS数据</strong>转化为<strong>ENU坐标系</strong>下的数据，从而得到<strong>初始化</strong>时<strong>载体相对于地图坐标系</strong>的<strong>位姿信息</strong></p>
<p align="justify" style="font-size:13px"> <strong>【基于Scan Context场景识别】:</strong> </p>
<p align="justify" style="font-size:13px"> <strong> 建图时</strong>存储关键帧对应的<strong>Scan Context</strong>与<strong>位姿信息</strong>，<strong>初始化</strong>时查找<strong>最相似的关键帧Scan Context</strong>并进行位姿匹配进而得到<strong>初始化</strong>时位姿估计的初值，进而进一步进行点云配准得到<strong>初始化</strong>时<strong>位姿</strong>的精确估计值</p>
</td>
</tr>
<!--SECTION 6 -->
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20">
<tbody><tr>
<td width="20%"><img src="./images/ESKF_localization/ESKF_HK_10s.gif" alt="PontTuset" width="250" style="border-style: none"></td>
<td width="80%" valign="top">
<p align="justify" style="font-size:13px"> <strong>------------------------------------------------------------------------------------------</strong> </p>
<papertitle>Localization based on pre-built point cloud map with Error-State Kalman Filter</papertitle>
</p><p></p>
<p align="justify" style="font-size:13px"> 2021.03-2021.04 </p>
<p align="justify" style="font-size:13px"> <strong>【地图上定位初始化】:</strong> 基于<strong>GNSS</strong>或<strong>Scan Context</strong>方法(使用初始scan查询<strong>Scan Context</strong>索引)完成<strong>基于点云地图定位的初始化</strong></p>
<p align="justify" style="font-size:13px"> <strong>【ESKF预测步】:</strong> 通过<strong>IMU数据</strong>利用<strong>中值积分</strong>方法完成<strong>惯性解算</strong>，
<strong>解算结果</strong>作为当前帧<strong>位置、姿态、速度状态误差</strong>的<strong>先验值</strong></p>
<p align="justify" style="font-size:13px"> <strong>【ESKF校正步】:
</strong> 有<strong>观测</strong>（<strong>当前帧点云</strong>和<strong>点云地图</strong>匹配获取位姿）时
求解<strong>误差状态</strong>的<strong>观测值</strong>，求解<strong>卡尔曼增益</strong>、<strong>后验协方差</strong>
和<strong>后验误差状态值</strong>；无<strong>观测</strong>时，<strong>后验值</strong>直接等于<strong>先验值</strong></p>
<p align="justify" style="font-size:13px"> <strong>【求解状态量后验值】:</strong> 根据<strong>后验误差状态量</strong>的值更新<strong>位姿、速度、IMU bias</strong>的<strong>后验值</strong></p>
<p align="justify" style="font-size:13px"> <strong>【验证】:</strong> 在公开的<strong>KITTI数据集</strong>和<strong>UrbanNav香港数据集</strong>上测试，本算法基于<strong>点云地图</strong>可以获取很高的定位精度</p>
</td>
</tr>
</table>
<!--SECTION 7 -->
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20">
<tbody><tr>
<td width="20%"><img src="./images/sliding_window_localization/SlidingWindow_10s.gif" alt="PontTuset" width="250" style="border-style: none"></td>
<td width="80%" valign="top">
<p align="justify" style="font-size:13px"> <strong>------------------------------------------------------------------------------------------</strong> </p>
<papertitle>Localization based on pre-built point cloud map with Sliding Window and Graph Optimization</papertitle>
</p><p></p>
<p align="justify" style="font-size:13px"> 2021.05-2021.08 </p>
<p align="justify" style="font-size:13px"> <strong>【地图上定位初始化】:</strong> 基于<strong>GNSS</strong>或<strong>Scan Context</strong>方法(使用初始scan查询<strong>Scan Context</strong>索引)完成<strong>基于点云地图定位的初始化</strong></p>
<p align="justify" style="font-size:13px"> <strong>【前端里程计】:</strong> 通过<strong>当前帧点云</strong>和<strong>地图</strong>匹配获取前端里程计</p>
<p align="justify" style="font-size:13px"> <strong>【基于滑动窗口的后端优化】:</strong> 维护一个固定长度的<strong>滑动窗口</strong>来保证定位的实时性，引入<strong>前端里程计的帧间约束</strong>、<strong>IMU预积分帧间约束</strong>和<strong>边缘化产生的约束</strong>对状态进行<strong>滑窗优化</strong>，获得<strong>实时</strong>的定位结果</p>
<p align="justify" style="font-size:13px"> <strong>【验证】:</strong> 在公开的<strong>KITTI数据集</strong>和<strong>UrbanNav香港数据集</strong>上测试，本算法基于<strong>点云地图</strong>可以获取较高的定位精度</p>
</td>
</tr>
</table>
<!--SECTION 7 -->
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20">
<tbody><tr>
<td width="20%"><img src="./images/slam_in_HIT/slam_in_HIT.gif" alt="PontTuset" width="250" style="border-style: none"></td>
<td width="80%" valign="top">
<p align="justify" style="font-size:13px"> <strong>------------------------------------------------------------------------------------------</strong> </p>
<papertitle>Mapping with LiDAR SLAM in HIT campus</papertitle>
</p><p></p>
<p align="justify" style="font-size:13px"> 2021.09-至今 </p>
<p align="justify" style="font-size:13px"> <strong>【设备】:</strong> RS-LiDAR-32 </p>
<p align="justify" style="font-size:13px"> <strong>【前端】:</strong> 对当前帧点云采用<strong>双阈值地面滤波</strong>分割出<strong>地面点云</strong>，
并采用<strong>PCA方法</strong>从当前帧<strong>非地面点云</strong>和维护的<strong>局部地图</strong>中提取<strong>点、线、面特征</strong>，
对<strong>同类别</strong>的<strong>点、线、面特征</strong>分别构建<strong>距离残差</strong>，进而构建<strong>线性最小二乘问题</strong>求解<strong>当前帧位姿</strong>，
求解过程<strong>收敛</strong>后将经过<strong>动态物体滤波</strong>后的<strong>特征点</strong>加入<strong>局部地图</strong></p>
<p align="justify" style="font-size:13px"> <strong>【后端】:</strong> 引入<strong>submap</strong>的<strong>相邻子图间位姿约束</strong>和<strong>回环约束</strong>，
采用<strong>TEASER</strong>作为<strong>回环帧</strong>间位姿估计的<strong>初值</strong>，
采用<strong>MULLS-ICP</strong>实现<strong>回环帧</strong>间<strong>位姿约束</strong>的<strong>精求解</strong>，
之后对<strong>submap</strong>中关键帧位姿进行优化</p>
<p align="justify" style="font-size:13px"> <strong>【验证】:</strong> 用自己在<strong>哈工大校园</strong>采集的数据进行<strong>激光建图</strong>，
该算法<strong>位姿估计</strong>以及<strong>建图效果</strong>明显优于开源的纯激光SLAM算法<strong>LeGO-LOAM</strong></p>
<a href="./images/slam_in_HIT/mapping_result_in_HIT.jpg">哈工大校园场景建图效果(VS LeGO-LOAM)</a>
</td>
</tr>
</table>
<!--SECTION 4 -->
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20">
<tbody><tr>
<td width="20%"><img src="./images/TVL-SLAM/TVL_SLAM.gif" alt="PontTuset" width="250" style="border-style: none"></td>
<td width="80%" valign="top">
<p align="justify" style="font-size:13px"> <strong>------------------------------------------------------------------------------------------</strong> </p>
<papertitle>Tightly coupled Visual-LiDAR SLAM</papertitle>
</p><p></p>
<p align="justify" style="font-size:13px"> 2020.03-2020.09 </p>
<p align="justify" style="font-size:13px"> 设计<strong>激光雷达</strong>和<strong>相机紧耦合</strong>的六自由度位姿估计算法，解决实际情况中基于<strong>激光雷达</strong>和基于<strong>相机</strong>的定位系统的不足</p>
<p align="justify" style="font-size:13px"> 1.设计<strong>点云-图像的数据关联</strong>算法，实现<strong>视觉特征点</strong>的提取、跟踪，并结合点云数据<strong>恢复</strong>出<strong>3D视觉特征点</strong></p>
<p align="justify" style="font-size:13px"> 2.优化<strong>地面分割算法</strong>，并在进行点云物体分割后提取激光特征点，以降低计算量，得到全局一致的轨迹和地图</p>
<p align="justify" style="font-size:13px"> 3.建立<strong>视觉特征点</strong>之间以及<strong>激光特征点</strong>之间的约束关系，使用L-M算法优化求解</p>
<p align="justify" style="font-size:13px"> 4.实现<strong>近邻</strong>的<strong>闭环检测</strong>和<strong>全局位姿图优化</strong>，得到<strong>全局一致</strong>的<strong>轨迹</strong>和<strong>地图</strong></p>
<p align="justify" style="font-size:13px"> <strong>【评价】:</strong> 此方法可以<strong>实时</strong>地估计出无人车地位姿，缓解激光雷达在<strong>结构特征缺失地场景</strong>下约束不足的问题，精度超过开源的A-LOAM、LeGO-LOAM</p>
<a href="./images/TVL-SLAM/TVL_comparison.png">本算法 VS LOAM & LeGO-LOAM</a>
</td>
</tr>
<!--SECTION 8 -->
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20">
<tbody><tr>
<td>
<heading>About Me</heading>
<p> <strong>Skills</strong>: C / C ++, Linux, ROS, OpenCV, PCL, g2o, ceres</p>
<p> <strong>Languages</strong>: Chinese: Native. English: CET-6: 534, CET-4: 614.</p>
</td></tr>
</tbody>
</table>
</td>
</tr>
</tbody>
</table>
<script>
if (typeof MathJax === 'undefined') {
window.MathJax = {
loader: {
load: ['[tex]/mhchem'],
source: {
'[tex]/amsCd': '[tex]/amscd',
'[tex]/AMScd': '[tex]/amscd'
}
},
tex: {
inlineMath: {'[+]': [['$', '$']]},
packages: {'[+]': ['mhchem']},
tags: 'ams'
},
options: {
renderActions: {
findScript: [10, doc => {
document.querySelectorAll('script[type^="math/tex"]').forEach(node => {
const display = !!node.type.match(/; *mode=display/);
const math = new doc.options.MathItem(node.textContent, doc.inputJax[0], display);
const text = document.createTextNode('');
node.parentNode.replaceChild(text, node);
math.start = {node: text, delim: '', n: 0};
math.end = {node: text, delim: '', n: 0};
doc.math.push(math);
});
}, '', false],
insertedScript: [200, () => {
document.querySelectorAll('mjx-container').forEach(node => {
let target = node.parentNode;
if (target.nodeName.toLowerCase() === 'li') {
target.parentNode.classList.add('has-jax');
}
});
}, '', false]
}
}
};
(function () {
var script = document.createElement('script');
script.src = '//cdn.jsdelivr.net/npm/mathjax@3/es5/tex-mml-chtml.js';
script.defer = true;
document.head.appendChild(script);
})();
} else {
MathJax.startup.document.state(0);
MathJax.texReset();
MathJax.typeset();
}
</script>
</body>
</html>