Security problems

Commented text during writing post in the MD source file, like this

1

<!-- Hidden text will appear in html source code. -->

will still be accessible for readers.

It is insecure to hide sensitive information by commenting them out, in any circumstance.

VS Code Extensions for Hexo and Markdown

Special for Hexo-theme-next

Use this line to include image:

1

<img src="{% asset_path FPKE_history.png%}" width="400px" title="caption">

General HTML

Add color to text

1

<span style="color:red">TEXT_COLORFUL</span>

Add background color to text

1

<span style="background-color:green">BACKGROUND_COLORFUL</span>

My conclusions

Zotero can do everything Mendeley could, even more elegantly. Simple DO NOT use Mendeley.

Sadly the development of Docear stopped, so don’t use it

Use tags to organize papers, see below for reasons. Leave collection just literally a collection of something.

Set up your own rules ahead and then follow them strictly. Keep polishing your rules as needed.

Working flow

Collecting efficiently

Add a new collection, search and add using the Zotero connector plugin for Chrome.

Download paper using sci-hub-fy plugin, if necessary.

Make sure meta-data are correct at the very beginning.

Files should be managed by ZotFile, saved in Dropbox.

Reading on tablet

ZotFile + Dropbox + Boox Note+ or other tablets. Helps to concentrate on papers.

1. Beate Klinger, and Torsten Mayer-Gürr, “The role of accelerometer data calibration within GRACE gravity field recovery: Results from ITSG-Grace2016”, Advances in Space Research, vol. 58, Nov. 2016, pp. 1597–1609.

It should be noted that our calibration approach aims at removing effects of instrument imperfections (cf. Sections 4.2 and 4.3) on the gravity field recovery. Hence, the used calibration equation does not guarantee to model these imperfections (e.g. temperature-induced bias drifts, misalignment) in a physical correct way, and will probably also absorb other not-modeled effects.

In this paper we only show results for GRACE-A, as very similar results have been obtained for GRACE-B. 【间接说明两颗星是独立做的calibration】

…

(p.1058) This is the only link between what the instrument actually measures (e.g., radiance, in the form of digital counts) and what one wants to measure (e.g., radiance). (p.1058) Conversion to at-sensor spectral radiance and/or top-of-atmosphere (TOA) reflectance is the fundamental steps to compare products from different sensors.

Calibration 的分类

See TanDEM-X at eoportal

• Internal calibration
• Geometric calibration
• Antenna pointing
• Antenna model verification
• Radiometric calibration
  • see here: Radiometric Calibration and Corrections
    • “Radiometric correction is done to calibrate the pixel values and/ correct for errors in the values. The process improves the interpretability and quality of remote sensed data. Radiometric calibration and corrections are particularly important when comparing multiple data sets over a period of time.”
    • “The value recorded for a given pixel includes not only the reflected or emitted radiation from the surface, but also the radiation scattered and emitted by the atmosphere. In most cases we are interested in only the actual surface values. To achieve these values, radiometric calibration and correction processes must be applied.”

If you are looking at this post, it means you are also pretty much a newbie to TensorFlow, like me, as of 2020-07-29.

tensorflow.keras

Keras is already part of TensorFlow, so, use from tensorflow.keras import ***, not from keras import ***.

TensorFlow backend

Early stopping

EarlyStopping

model.fit(..., callbacks=[EarlyStopping(monitor='val_loss', patience=5, verbose=1, mode='min', restore_best_weights=True)], ...)

Reproducibility of results

TL;DR
Set all random seeds
Use tensorflow.keras instead standalone keras
Use model.predict_on_batch(x).numpy() for predicting speed.

Useful Commands

Some commands that are not easy to search for.

阅读 GPstuff 包中的函数的源代码，对其中注释做的笔记，在 official document 中有些内容没有写。

基础知识

网络上的大量的教程供学习，先去搜索学习一段时间（2天）后再来看这个教程。

When you search “Excluding self-citation in Google Scholar” on Google, the first item is [By Brinxmat]. But unfortunately it does not work anymore, at least for me.

However, it gives me some insights and then I fiddling with Google Scholar a little bit. Now we have an easy solution.

This is the original citation link of one of my paper:
https://scholar.google.com/scholar?oi=bibs&hl=en&cites=17562392294963478550
This link actually works:
https://scholar.google.com/scholar?hl=en&as_sdt=0%2C31&sciodt=0%2C31&cites=8618462917299343813&scipsc=1&q=-author%3A"H+Peng"&btnG=

As you see, the trick is still on the query except that it has been much more complicated than what Brinxmat found in 2010.

So, the manual solution is to:

A little explanation:

• After opening a cited by link, you will see a checkbox saying “Search within citing articles”, then check the box.
• Then put -author:"H Peng" in the search box, then press Enter.
  (It is a short hyphen there, not dash. Pay attention to your input method.)
• Then the above link address shows up.

用线条填充图

使用 Hatchfill2
配合 legendflex 对 lagend 进行控制

SPACEofPHD 上的推送v

优雅地作图——用Matlab画出比较漂亮的图

科研绘图 | 提取两个曲面构成的交曲面（两种方法）

科研绘图 | Matlab绘制精美的势能函数剖面图

link properties of graphic objects:
linkprop
linkaxes only links XLim or YLim or both。 内部是用 linkprop 实现的。

推荐：只用来展示时使用，每个axis还是通过命令单独设置，否则很难控制赋值逻辑，经常被其它函数重写。

问题：fill会重写之前的所有axis属性（as of 2019-05-01 18:17:59）

科研绘图 | Matlab绘制精美的势能函数剖面图

AngularRaDec(final GroundStation station, final Frame referenceFrame, ... 给出的是 topocentric Ra/Dec。

从theoreticalEvaluation()的代码中可以看出， Station-satellite vector expressed in inertial frame 是一个 vector (not a state)，转换到 Field transform from inertial to reference frame at station’s reception date 后，对得到的 state 求解 Ra/Dec。
实质上，等价于，把 reference frame 平移到 station’s location，然后在这个新的frame中求解 Ra/Dec，即 topocentric Ra/Dec。

另一方面，参考 Vallado1997 书中 6.4 Angle-only observations 的说明，angle-only observation 没有办法直接得到 range information，所以得到的只能是 topocentric Ra/Dec。