Interference Mitigation

Vaishnavi worked on Median-based interference identification and removal. The basic idea is to treat the time-freq map of a given baseline as a 2-D array. See the above diagram for such an example, where the red arrow indicates one interference location.

We then take a small section of this 2-D array (say 32x32 matrix), where the 3rd axis is the intensity (or amplitude). We compute the median and standard deviation of the section (32x32 matrix). We then put some criterion of (median+7*sigma) for genuine data.

Amplitude > (median + 7*sigma) is treated as interference. This appears to identify interference quite reasonably. Check the following image, where black pixel indicates interference. Compare that with the top image, where most of the interference is identified.

There is some data loss due to over-correction. Even so, total data flagged is about 15%, which is quite good.

Interference Removal from Time-Frequency Data

We start with a flat time-frequency data, much like shown in the image on the left. This is VLA data (one baseline) observed at 325 MHz. The arrows indicate high-level and low-level interference in the data.

The most striking quality of interference, is that it is mostly confined to high 'spatial' frequencies. Here, I refer to spatial grey-level variation frequencies.

There are two approaches to remove / isolate such an interference:

1. To take an FT of the image itself.
2. Remove the high-frequency features (those belonging to fast-changing features.
3. For all pixels, use a large (21x21 pixel) filter to compute mean, median and rms of grey levels.
4. If the pixel value is greater than (mean+5*sigma) or (median + 5 sigma), then one can safely substitute median for the pixel value.
5. One could compute rms over the entire image to avoid being dominated by local features.

2008: Plans and Actions

Order not so crucial, but wish list is quite doable, right?

1. Get SAX paper out of the door.
2. Finish analysing MHD cube, how many chimneys will i lose due to sensitivity and resolution limits? Can we count holes in HI for a large number of MHD simulations?
3. B0809+74, what is the circulation time?
4. B0834+06, what is the problem with the circulation time at 35 MHz? How does the polar pattern look at higher frequencies?
   
5. 2-antenna interference on 3/4 major sources: from amplitude/ phase calibration of fringes, can we tell the sizes of sources with fringes observed?
   
6. Can we make crude images with even 10 different u-v tracks with a 2-antenna telescope?
7. Can we use the Moon for occultation studies?
8. Astrophysics - II course.
9. Meditate in Leh.
10. Maha working towards her higher studies abroad.
    

X'mas mass in Goa Velha

I attended the midnight mass on 24th December night with Angela and her family, along with Devart (Bablu) Rana. The main attraction was to see the localization of the global X'mas tradition: the celebration of the birth of Jesus.

Some curious points:
1) Everyone, irrespective of where they were born, celebrate some birth of a man thousands of Km away. They do not know most of the names (Bethelhem, Jerusalem, sea of Mediterrainean) and the times involved.
2) Continuing the above, all think Jesus was white. No mention was made to Judaism and the roots of christianity in Judaism.
3) For common folk, there were skits presented, where common public problems (drinking, gambling, etc.) were shown to have a common cure (christianity). No efforts were made to understand the psychological problems possible, how healing is possible without invoking the Lord, or without any reference to His greatness.
4) Most the examples above were very crude. Indeed, how can theft (possible social and economic factors being dominant) be bracketed with smoking or drugs (where personal factors may dominate)? To localize such problems and separate the different factors behind the cause of such 'illness' is vital to find a cure.
5) Political statements were inserted slyly, for example, SEZs were bad. No reasons were given, except vague statements about nature. How economics matters and how it can sideline sustainable growth is never openly discussed. Perhaps I am expecting too much from an organized religion.


On the other hand:

6) Hindu way of 'pooja' (prayers) was adopted, priests were using 'agarbatti' (incense sticks) to make circular movements in front of the Lord and were chanting. This is very similar to our Ganapati pooja, except the 'arati' is different.
7) Songs were in all local languages, Marathi, Konkani, and (most amusingly) English and Portuguese. English is the medium of choice for the upwardly mobile (like yours truly). Portuguese was displaced from that position after 'independence'...
8) Occasion was treated as a way to show-case local fashions by many. In fact, some local people explained that " 'too many reveling' clothes by front-bencher women kept the priests from falling asleep". The one in the rear rows promptly closed the eyes and caught on their sleep.

After this foreword, here are a couple of songs in Konkani

1) San Kiteak ailai balla?
Sang konnem tuka dhadd-la?
Sorg soddun koso denvloi?
Khoim Mhonn tthikann kortoloi?

Patkantlim soddounk ailam
Devjivit vattunk tumkam
Devmogan tumkam bhoronk.
Bapache vengent vhorunk.


2) Mogan thevlem monxeachea hatant
Devalem dan (3) ball Jezu rupan (2)
Dan tem omolik, kitlem tem vichitr
Guttlailolem soddvonne ghuttant

Natalm disa ailo denvum mog
Amchea kallzamni ravunk dovrunk tog,
khuxalkaie-dan sonvsarbhor vanttunk
pavla konnkonim chol-ia moganuch.

Natalam disa ailo denv'n sontos
Dukh-khoddamnin bhorunk visvas vhodd,
Bhorvanxachem dan jivitant vosunk
Niraxiponnam dium-ia soddun.

Processing HI cube - II

A) We now have a cube V_los(x,y,z)

1. Compute V_min and V_max (extrema) of V_los matrix,
2. Create new array V_obs, where velocity values (V_los) range from V_min to V_max, in steps of 1 km/s,

B) We want an output image matrix, OUT(x,y,z), where x_i corresponds to V_obs(i)

1. Scan the V_los(x,y,z) matrix; say, V = V_los(x,y,z)
2. If V1 < V < V2, then OUT(v,y,z) += rho(x,y,z)
   
3. Repeat a & b for energy and pressure.

Processing HI cube - I

Computing the orientation of each cell w.r.t. the central pixel

We would like to keep the simulation cube at a distance d (element: X=0, YMAX/2, ZMAX/2 is at a distance d).

Now, we compute angle subtended by all cells, and compute differential rotation velocity as per Brand and Blitz (1991).



1) Read velocity cubes Vxx,Vyy,Vzz

a) Compute projects of components of V_rot (due to galactic rotation) along the los.
b) Compute projection of velocities Vxx, Vyy & Vzz along the line of sight (los)
c) V_los = V_rot + components from step a


Formulae

a)
tan(theta) = y / (d+x)
(D * D) = (d+x)^2 + (y*y)

R' = sqrt ( D^2 + R^2 + 2*R*D* cos(theta+L)

Compute V(R') using Brand & Blitz (1991)

R" = D sin(theta + L)
phi = 90 - (theta+L)
phi" = arccos(R"/R)

V_rot = V(R') * cos(90 - (phi+phi") ) = V(R') * cos(theta + L - phi)

b)
theta = arctan(y / (d+x) )
V"yy = Vyy cos(theta)
V"xx = Vxx sin(theta)

chi = arctan(z/ (d+x) )
V"zz = Vzz cos(chi)

c) v_los (x,y,z) = V"xx + V"yy + V"zz + V_rot

Fractals and image characterisation

Some links are in order

1. Fractal Dimension: Wikipedia
   
2. A course on Fractals in Yale U
3. A course on Fractal dimension from images: Munich U
4. Fractal Dimension explained
   

So, once you know about fractal dimensions, come to read the stuff on the right (Conti, 2001)


One can treat the image 3-D object. Compute the total number of occupied boxes in X-Y-I dimension box, as a function of size of the box. D = ln(number)/ ln(radius).

It is a little bit more complicated. Check the paper by Conci, a PPT talk can also be found.

How to distinguish between landscape and portrait pictures?

1. Perhaps we can search for a large number of pixels with same natural colors: green, blue and black (shadows). look if a large fraction of pixels contain the same 'Hue' and 'Saturation'.
2. Another try: Look at the Fourier spectra of images, and mark radii of 60%, 90%, 99%, 99.9% power. they should be distinct for landscape images and facial portraits or nearby objects.
   
3. Human objects have a lot more symmetry than the natural objects. In fact, there could be some fractal pattern seen over the different length scales of an image of a natural scenery. Try to capture 'fractal' properties of pixels.

speaking of the last one: one could look at fractal dimension of a picture pixel values. How? Perhaps in the next blog post...

PHY GC 471 : Astrophysics, a first course

Course Description

This is a first course in astronomy aimed at a wide section of audience. We will introduce concepts with minimal use of mathematical tools. We would like to give a broad perspective of science to students, from regular night-sky astronomy to fascinating space astronomy. We will discuss historical context of some ideas and how they have evolved today. Keeping the engineering background of student population in mind, we will also emphasize on techniques, some challenging instrumentation, and details of space exploration.

Scope and Objectives

This being a first course to a variety of students from engineering background, we prefer to teach fundamental ideas about astronomy, and defer a detailed astrophysical treatment to another higher-level course. Astronomers study properties of light emitted by various sources in the sky. We will start by studying concepts of brightness, flux measurement and spectral-line observations, followed by instruments used these measurements. We will then discuss various astronomical objects, starting from our solar system, stars, to galaxies. In each case, we discuss basic astrophysical mechanisms to understand nature of light emitted by objects under consideration. We then discuss how to estimate physical parameters of stars, galaxies and the cosmos from the properties of light detected in various wavebands.

Textbook

IGNOU study material PHE-15 (Astronomy and Astrophysics), Indira Gandhi National Open University, 2006.

Astronomy Discussion Meetings: Proposal to the Astronomical Society of India

Over the past few years astronomy meetings have not been held regularly in India. Astronomical Society of India (ASI) only organizes bi-annual meetings, which are woefully short on discussions. To increase their frequency, topical discussion meetings involving a small number of active participants were proposed. Such meetings will allow a close interactions between astronomers, mainly from India. They will have an added benefit of smaller number of participants and of being focussed on topics of LOC's interests. The costs and local organization efforts will also be reduced significantly.

The ASI has accepted the proposal in principle. Here are the basic tentacles of such meetings. The ASI has granted a sum of 1 Lakh towards expenses of the meeting. One could further approach DST/ISRO/UGC for more funding. Given Ranjan being in DST committee, it would be easier from them.

There will typically be sessions over 2 and half days, with about 15 presentations. The emphasis is on discussions. Hence, followed by one 45 minute talk, there is 15 minute's break for Q/A session. At the end of each session, there will be an hour of open debate on possible new work, or comments on on-going work. This is place for on-board calculations, and some laptop demos, etc.

The meeting crucially depends on 4/5 resource persons. These will chair sessions, seed discussions and direct the course of the meeting. Their participation is vital, as is the willingness of the public to discuss issues and settle them on the spot, rather than defering to person-to-person discussion.

There will be about 10-15 faculty and 10-15 PhD scholars. Most members will stay during the entire two and half days of the meeting. There are no frills, except perhaps one institute buffet and some drinks from the LOC. The number is restricted to a number of 25, and is by invitation only. Most participants will not be provided travel support, and the local travel support will be provided on pay-per-use basis.


All these are thought to make the arrangement work of the LOC minimal, allowing for more groups from India to host meetings. This will allow a larger number of meetings, wider set of meetings, and dispersing meeting venues.


Amen to that!