algorithm.rst 6.3 KB
Newer Older
BO ZHANG's avatar
tweaks    
BO ZHANG committed
1
2
SLS pipeline
============
BO ZHANG's avatar
BO ZHANG committed
3
4
5
6
7


Instrument correction
---------------------

Wang Xia's avatar
Wang Xia committed
8
Description
Wang Xia's avatar
Wang Xia committed
9
``````````````````
Wang Xia's avatar
Wang Xia committed
10
`csst_ms_sls_instrument` 程序包是应用于CSST无缝光谱模块的仪器效应改正程序,无缝光谱模块包含12块9k×9k探测器,每4块探测器对应1个波段,每个波段的部分仪器参数如下:
Wang Xia's avatar
Wang Xia committed
11

Wang Xia's avatar
Wang Xia committed
12
13
14
    - GU: 660×660″ field of view from 255-400nm with a plate scale of 0.074″/pixel
    - GV: 660×660″ field of view from 400-620nm with a plate scale of 0.074″/pixel
    - GI: 660×660″ field of view from 620-1000nm with a plate scale of 0.074″/pixel
Wang Xia's avatar
Wang Xia committed
15
`csst_ms_sls_instrument` 将顺序执行一系列的探测器效应改正,生成单次曝光图像预处理后的数据产品。该程序是用Python语言实现,代码地址:code_。
Wang Xia's avatar
Wang Xia committed
16

Wang Xia's avatar
Wang Xia committed
17
.. _code: https://csst-tb.bao.ac.cn/code/csst-l1/sls/csst_ms_sls_instrument
Wang Xia's avatar
Wang Xia committed
18
19
20
Input
``````````````````
`csst_ms_sls_instrument` 将从csst_common.CsstMsDataManager获取0级数据和参考文件。
Wang Xia's avatar
Wang Xia committed
21

Wang Xia's avatar
Wang Xia committed
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
1. 0级数据: 包含一个扩展的fits文件,头文件中存放着观测天区、曝光信息、探测器信息等,具体关键字说明参见 DataModel_,扩展的数据单元存放着原始观测图像。
#. 参考文件: 仪器效应改正中所需的参考文件如下表所示。

+-----------------+----------------------------+
| Reference file  | Description                |
+=================+============================+
| Gain map        | gain array(9k×9k)          | 
+-----------------+----------------------------+
| Superbias       | superbias file(9k×9k)      | 
+-----------------+----------------------------+
| Superdark       | superdark file(9k×9k)      |
+-----------------+----------------------------+
| Superflat       | super flatfield file(9k×9k)|
+-----------------+----------------------------+
| Badpixel table  | Bad/hot Pixel table or map |
+-----------------+----------------------------+
Wang Xia's avatar
Wang Xia committed
38
39

.. _DataModel: https://csst-tb.bao.ac.cn/code/csst-l1/csst-l1doc/-/blob/main/docs/source/sls/data_model.md
BO ZHANG's avatar
BO ZHANG committed
40

Wang Xia's avatar
Wang Xia committed
41
Output
Wang Xia's avatar
Wang Xia committed
42
``````````````````
Wang Xia's avatar
Wang Xia committed
43
`csst_ms_sls_instrument` 生成的数据产品是包含三个扩展的fits文件,扩展内容如下表所示:
Wang Xia's avatar
Wang Xia committed
44

Wang Xia's avatar
Wang Xia committed
45
46
47
+-----------------+---------+-------------------+
| Extension name  |  Bunit  | Comment           |
+=================+=========+===================+
Wang Xia's avatar
Wang Xia committed
48
| SCI             | e-/s    | Science image     |
Wang Xia's avatar
Wang Xia committed
49
+-----------------+---------+-------------------+
Wang Xia's avatar
Wang Xia committed
50
| ERR             | e-/s    | Error array       |
Wang Xia's avatar
Wang Xia committed
51
+-----------------+---------+-------------------+
Wang Xia's avatar
Wang Xia committed
52
| DQ              | unitless| Data quality array|
Wang Xia's avatar
Wang Xia committed
53
+-----------------+---------+-------------------+
Wang Xia's avatar
Wang Xia committed
54

Wang Xia's avatar
Wang Xia committed
55
第一个扩展(SCI)的数据单元存放着仪器效应改正后的光谱图像,第二扩展(ERR)的数据单元存放着光谱图像误差,第三个扩展(DQ)的数据单元存放着光谱图像每个像素的数据质量标志位,具体参见 DQFlags_,DQFlags应包含下表的三项。仪器效应改正过程中,同时会记录下处理时间、状态信息等关键字,具体参见 DataModel_。
Wang Xia's avatar
Wang Xia committed
56

Wang Xia's avatar
Wang Xia committed
57
.. _DQFlags: https://?
Wang Xia's avatar
Wang Xia committed
58

Wang Xia's avatar
Wang Xia committed
59
60
61
62
63
64
65
66
67
+----------------+---------+---------------------------------------------+
| TTYPE          | TFORM   |  Description                                |
+================+=========+=============================================+
| BIT            | integer | The bit number, starting at zero            |
+----------------+---------+---------------------------------------------+
| VALUE          | integer | The equivalent base-10 value of BIT         |
+----------------+---------+---------------------------------------------+
| DESCRIPTION    | string  |  A description of the data quality condition|
+----------------+---------+---------------------------------------------+
Wang Xia's avatar
Wang Xia committed
68

Wang Xia's avatar
Wang Xia committed
69
70
71
Data Calibration Steps
``````````````````
仪器效应改正包括以下步骤:
Wang Xia's avatar
Wang Xia committed
72

Wang Xia's avatar
Wang Xia committed
73
**DQ Initialization**
Wang Xia's avatar
Wang Xia committed
74

Wang Xia's avatar
Wang Xia committed
75
class: csst_ms_sls_instrument.steps.DQIstep
Wang Xia's avatar
Wang Xia committed
76

Wang Xia's avatar
Wang Xia committed
77
78
reference file: Badpixel table(map)、Saturation file

Wang Xia's avatar
Wang Xia committed
79
DQ Initialization实现两部分内容,一是利用参考文件Badpixel table或是Badpixel map对DQ扩展进行标记,该参考文件记录着存在问题的像素点,可能是探测器的坏点、热像素。二是根据Saturation file对观测数据进行判定,并在DQ扩展中标记。Saturation file的形式需要根据后续测试结果来定,如果能测出每个像素的饱和值即可做成map,现阶段用的是常值。DQ标记的详情参见 DQFlags_。
Wang Xia's avatar
Wang Xia committed
80

Wang Xia's avatar
Wang Xia committed
81
82
83
84
**Bias correction**

class: csst_ms_sls_instrument.steps.BiasCorrStep

Wang Xia's avatar
Wang Xia committed
85
86
reference file:Superbias

Wang Xia's avatar
Wang Xia committed
87
Bias correction通过减去superbias的操作,去除掉原始科学数据中的探测器本底。Superbias的文件格式与输出文件一致,为包含三个扩展的fits。该步会对superbias的误差和DQ进行传递。
Wang Xia's avatar
Wang Xia committed
88

Wang Xia's avatar
Wang Xia committed
89
**To electrons**
Wang Xia's avatar
Wang Xia committed
90

Wang Xia's avatar
Wang Xia committed
91
class: csst_ms_sls_instrument.steps.ToElectronsStep
Wang Xia's avatar
Wang Xia committed
92

Wang Xia's avatar
Wang Xia committed
93
94
95
96
reference file: Gain map

数据乘以增益,SCI数据单元和ERR数据单元的单位由ADU转换为电子。

Wang Xia's avatar
Wang Xia committed
97
98
99
100
**Uncertainty Initialization**

class: csst_ms_sls_instrument.steps.UncertaintyInitStep

Wang Xia's avatar
Wang Xia committed
101
102
103
104
reference file: readnoise file

ERR数据单元构造误差数据,该步误差公式:

Wang Xia's avatar
Wang Xia committed
105
.. math::
Wang Xia's avatar
Wang Xia committed
106
107
108

 \sigma = \sqrt{(SCI-bias)+{\sigma_bias}^2 + readnoise^2)} 

Wang Xia's avatar
Wang Xia committed
109

Wang Xia's avatar
Wang Xia committed
110
111
112
113
**Dark correction**

class: csst_ms_sls_instrument.steps.DarkCorrStep

Wang Xia's avatar
Wang Xia committed
114
115
116
117
reference file:Superdark

Dark correction通过减去superdark的操作,去除掉原始科学数据中的探测器暗电流。Superdark的文件格式与输出文件一致,为包含三个扩展的fits。该步会对superdark的误差和DQ进行传递。

Wang Xia's avatar
Wang Xia committed
118
119
120
121
**Flatfield correction**

class: csst_ms_sls_instrument.steps.FlatCorrStep

Wang Xia's avatar
Wang Xia committed
122
123
124
125
reference file:Superflat

Flatfield correction通过除flatfield的操作,去除掉像素不均匀性。Superflat的文件格式与输出文件一致,为包含三个扩展的fits。该步会对superflat的误差和DQ进行传递。

Wang Xia's avatar
Wang Xia committed
126
127
128
129
**CR rejection**

class: csst_ms_sls_instrument.steps.CRrejStep

Wang Xia's avatar
Wang Xia committed
130
131
reference file:cr model 

Wang Xia's avatar
Wang Xia committed
132
CR rejection基于无缝光谱图像训练的模型,利用 deepCR_ 对单次曝光的科学数据进行宇宙线检测,在DQ中进行标记,标记详情见 DQFlags_ 。
Wang Xia's avatar
Wang Xia committed
133
134
135

.. _deepCR: https://deepcr.readthedocs.io/en/latest/

Wang Xia's avatar
Wang Xia committed
136
137
138
139
**To electrons/sec**

class: csst_ms_sls_instrument.steps.CPSStep

Wang Xia's avatar
Wang Xia committed
140
141
142
143
144
reference file:无

数据除以曝光时间,SCI数据单元和ERR数据单元的单位由e转换为e/s。


Wang Xia's avatar
Wang Xia committed
145

BO ZHANG's avatar
BO ZHANG committed
146
147
148
149
150
151
152
153
154
155
Position calibration
---------------------

`csst_ms_sls_position` package.


API
---

TODO