Hacker News new | comments | show | ask | jobs | submit login
Pixel Perfect map of china (o.cn)
233 points by daleharvey on Mar 3, 2011 | hide | past | web | favorite | 62 comments

There's quite a lot of map companies offering similar "SimCity-style" maps:

Edushi (http://edushi.com/), as others have mentioned --- IIRC they were one of the first to market

Baidu rolled out their own offering in 2010 (go to http://maps.baidu.com/, pick a city from the menu on the right, and then click the "三维" button)

City8 (http://city8.com/) don't have 3D maps, but they have a Streetview equivalent which they've been working on since 2008 in Shanghai, Beijing, and other cities: http://en.city8.com/search/search.aspx

Couple of articles which may be interesting reading (in Chinese) --- touching on China-specific privacy, legal/data licensing issues, etc. surrounding these maps:



This is neat. For example:

Shanghai skyscrapers: http://j.map.baidu.com/-FNB Beijing Olympic stadium: http://j.map.baidu.com/TjNB

Edit: for some reason, the links don't work in Chrome. Firefox works though.

The Baidu maps are powered by o.cn--you can see "o.cn" in the watermark.

Or for example chachaba.com (in Shenzhen):


They had problems with providing a mapping service in China without a license though. They were forced to shutdown for a very long time + redesigned their website to survive. Make sure you get the right license first in China!

For "Ultima-style" city maps, check out 8-Bit Cities:


A couple of notes from a resident of Beijing:

I wonder how frequently this will be updated. I notice several skyscrapers built over the past few months that aren't displayed yet (they only show the construction huts). So it's at least several months out of date.

This clearly has the most detail for zooming and browsing that I've seen, but for my day-to-day use I'll still be using map.sogou.com, which already has pretty good building and landmark decomposition, but has the best path-finder / location search of everyone in the market by far. I just made several queries on bj.o.cn that fell flat. For really hard-to-find places, I suspect I'll do my searching on sogou first, then pull up the closest landmark on o.cn and scroll over to where I want to be for a good visual description of where I'm going.

"I notice several skyscrapers built over the past few months that aren't displayed yet"

that's a crazy pace !, compared to seeing the same unmistakable window view at San Jose/ SF every single year.

o.cn doesn't seem to have very good coverage of Beijing. It looks like everything outside of the 3rd ring is missing.

I've mentioned this before, but entering the "Chinese" internet (or internet of any foreign culture for that matter) is like entering a whole new world. It's amazing how the cultural aspects impacts small things like design sense and advertising colors.

This is mind-blowing. How was this created? The detail is incredible: http://i.imgur.com/gmsu4.png

The buildings are meticulously 3-D modeled and texture mapped--you can see the texture repetition artifacts. I've read some research about automatic modeling techniques from street view images+, but I'd suspect this was done by hand.

+ http://web.mit.edu/jxiao/Public/publication/2009/TOG/paper_h...

Looks like China's cheap labor is bridging the gap to AI.

This is essentially the same as a traditional animated movie: gigantic sum of work to do, essentially infinitely parallel, and capable of being process-optimized such that two adjacent frames require much less work than freehanding them. They're probably using a large library of templated buildings (how many silhouettes are there at that resolution, really?), brushes, and then just applying detail work to the significant ones.

It's a big job, and it's impressive, but it isn't "We had enough people to freehand a scale map of the entire country, bwahaha."

Perhaps they'll perform Searle's "Chinese room" experiment IRL.

Props for the reference, but aside from matching /Chin/ the Chinese room has absolutely nothing to do with this approach. It is staffed by a single guy and it is critically important to the thought experiment that he does not understand Chinese.

(Brief sketch of the Chinese room: there is a locked room with a slit which permits paper to come in and paper to go out. Inside the room is a man who does not speak Chinese. He receives paper with Chinese symbols on it, consults a vast library of books with rules on what to do in response to particular symbols, laboriously copies his response onto paper, and pushes it out through the slit. The response is intelligible as Chinese responsive to the input Chinese. Searle argues that the man can't understand Chinese. Personal opinion: it's navelgazing that only matters to philosophy, but I think the man and books together constitute a system which speaks Chinese, in the same way that people bidding in an auction together constitute an efficient price discovery mechanism even if none has expert knowledge of the "true value" of all items at auction.)

aside from matching /Chin/ the Chinese room has absolutely nothing to do with this approach. It is staffed by a single guy and it is critically important to the thought experiment that he does not understand Chinese.

It raises an alternate formulation: What if the operation of the room was crowdsourced? What if the individuals of the crowd could communicate and organize, or what if not? Would that change the properties of the thought experiment in any interesting ways?

Obviously, if the room included Chinese people the language would have to be different. We could call it the "English room" thought experiment.

The point is not whether the system understands Chinese, the point is that such an "algorithm" or "system" does not produce human-like consciousness.

> The point is not whether the system understands Chinese

From early in Searle's paper: "Partisans of strong AI claim that in this question and answer sequence the machine is not only simulating a human ability but also 1. that the machine can literally be said to understand the story and provide the answers to questions, and 2. that what the machine and its program do explains the human ability to understand the story and answer questions about it. Both claims seem to me to be totally unsupported by Schank's work, as I will attempt to show in what follows."

And, just after describing the "Chinese room" scenario: "Now the claims made by strong AI are that the programmed computer understands the stories and that the program in some sense explains human understanding. But we are now in a position to examine these claims in light of our thought experiment."

> human-like consciousness

No, Searle is not at all only concerned with "consciousness". Searle again:

""" "But could something think, understand, and so on solely in virtue of being a computer with the right sort of program? Could instantiating a program, the right program of course, by itself be a sufficient condition of understanding?" This I think is the right question to ask, though it is usually confused with one or more of the earlier questions, and the answer to it is no. """

(The focus on "consciousness" is a more recent development, and I cynically suspect it's motivated by a recognition that as far as anything we can observe goes, computers are in fact likely to be able to do everything humans can in the not too distant future -- so best to concentrate on something conveniently unfalsifiable, such as the claim that computers couldn't really be "conscious" even if they behaved in every respect exactly as if they were.)

Searle talks about "understanding" throughout. He occasionally makes reference to other mental capabilities, including "consciousness" once or twice, but "understanding" is much the most frequent.

"understand" is an overloaded term. It could mean a few things:

- Human like understanding, i.e. awareness, consciousness

This is what the Chinese room experiment is designed to dispute

- Ability to produce appropriate output

We often use "understand" to mean this. e.g. "I wrote a parser that understands Ruby code and compiles it to C".

Of course the Chinese room "understands" Chinese in the second sense, but not the first sense.

You first quote is describing what I consider to be awareness/consciousness. Maybe Searle didn't use the same word, but I believe he's describing the same notion.

Think of it this way: a C compiler doesn't really "understand" C code in the same way that a human does. For instance, it can't make changes to the code. If it could, it would replace the programmer.


From wikipedia: http://en.wikipedia.org/wiki/Chinese_room

> The experiment is the centerpiece of Searle's Chinese Room Argument which holds that a program cannot give a computer a "mind" or "understanding", regardless of how intelligently it may make it behave.

The real question is, is this more useful than the photographic approach? i.e. does a user identify these landmarks better in this slightly abstracted hand drawn rendering than in a photo rendering of the same?

There's less cognitive overhead in processing these renderings because there's less stuff to pay attention to. A photograph tends to record everything, which can be overwhelming if you're looking for select landmarks.

Also, the real world tends to be more drab and polluted than these cheery 100% saturation renderings.

Is it fully 3D rendered? or just an isometric view like the old simcity? You can't rotate the viewpoint, as far as I've seen (but as I don't know Chinese I might be wrong).

Seriously by hand? How many mechanical turks? Why do you say so? The textures and all seem to suggest otherwise. I am very much interested in the actual technology. Perhaps the textures were superimposed on the 3d models and not a direct input of the 3d modelling.

The cost of labor in China is still low enough for this to happen.

The research done by your fellow in MIT while in HKUST is also very impressive. Just that the result is not the same kind found in here - the one here is simply artistic model, not a realistic one.

This map only covers Guangzhou, actually. But I look forward to full coverage of China in a few hundred years!

You'll probably be waiting longer than that. One of the reasons China opted for the hand drawn approach is it lets them control what is and (more importantly) isn't shown.

I enjoy a good sinophobic innuendo as much as the next guy, but how is it any easier to omit something from a hand-drawn map than to blur it out in a satellite photo?


I have no opinion on what is true in this context, but completely in the abstract: when drawing things it's possible to put in fake data (residential area or forest where a military facility or nuclear reactor is located) without it being possible to be detected. The blurring at least shows that something is there.

Heh... It's possible to fool aerial viewers as well:


That is wicked cool, thanks for the link. It must be awesome to be assigned a job like this, or being part of the crew implementing it.

nah, there is a city switcher in the top left.

And here's the one from Shanghai: http://sh.o.cn/

It's hard to explore o.cn in detail because everything is in Chinese, but whatever they have is quite impressive. Not only is the 3D view really well rendered (i.e. easy on the eyes), they have incredible attention to detail, down to using billboards as ad space.

Reminds me so much of SimCity, only you can't raise taxes.

How can an image be "pixel perfect" if the original is not made of pixels? I can understand pixel-perfect image scaling or the like, but a pixel-perfect map of the real world is impossible.

This is amazing. I wonder how they do this? Surely it has to be automated somehow, I just can't imagine anyone building this by hand, even in China.

I've seen similar maps from 100+ years ago of my hometown in the UK:


And of London:


It's amazing how our idea of what is unachievable by hand has been skewed by computer graphics!

Not only can they build rendered cities at breakneck pace, they can do it for the real thing:


Actually that's in fact possible. The models must be done by hand for landmarks, the rest may be repetitively. Similar efforts has been done by OSS games (http://wiki.openttd.org/City_Buildings_%28New_Graphics%29)

Obviously this is not what was done here, but how difficult is it to take a satellite photo from the side, instead of from directly above?

Not very. Most photos are taken from some angle (it's easier to reorient a satellite than to reposition its orbit). Course, you get the most quality when shooting close to straight down.

Wow... just... wow. The obvious question is... how was this done? Mechanical Turk?

There's another company eDushi created this kind of map in 2007 or earlier. The company is in Hangzhou. They cover more cities than o.cn. I've checked the details of architecture. Quite accurate. And there is a lot daily life related information hidden in the map.

Here's the map of Hangzhou. http://hangzhou.edushi.com/

this is also very impressive. I like how they add small buses and signs to indicate bus stops. We can also see the name of big shopping centers

I checked how they did that. Basically they provide a easy way to let a user(Small buses owner) to add info. In other words it's a kind of "social mapping".Probably with some control.

I found an English version of the company behind Edushi: http://www.aladdincn.com/en/index.html

The technical/theoretical side of it, the how do you scale up the (automatic) creation of those Sim-Cities ... yes, VERY fascinating and interesting.

The other side of the medal: those maps are creating a cute, sleek, clean and well-behaved view on the areas of high(est) population density ... they are I think heavily distorting the perception of the reality in those areas. I consider them a very strange way to do propaganda.

The reality is hidden by this means, which is the opposite of what maps/sat maps should be providing...

"Pixel Perfect" nails it quite good. They are unrealistically perfect. Constructed down to every pixel presented.

Event the most horrible sweatshops look cute in this renderings.

If this had/has an API it would make for a great game-board for MMOG's/on-line-&-offline-games or even strategy-games. Imagine adding textures for units and/or craters (though I'd guess this latter would be banned pretty soon).

I assume they're doing this because of some government restriction on photographing the city? Otherwise, I can't imagine that this kind of effort could be profitable...

Meaning "pixel-art-style map of part of one city in China with many copy/pasted buildings."

You can change cities by clicking on "切换城市". Of course it's still not the whole country but it's still more than one city.

Baidu has something similar as well: http://j.map.baidu.com/hg2l

Are they doing this to get around some restrictions on showing real images?

there's no restriction, most of google earth works. personally i feel that these sim city maps are easier to view compared to live ariel photos.

i went to edushi to discuss a business opportunity, they were telling me how these were updated, i heard it's weekly. they also have a program with a school that teaches 3d studio max so they have a pipeline of people to help render these sim city like maps.

Baidu is using the o.cn images itself as explained by the watermarks.

And... it was out of date the minute it went live...

Can't read the language. What is this? The SimCity version of China?

It's Chinese, which is a three-dimensional map of GuangZhou city, the labels are the names of places. and as rywang mentioned above, there is a 3d map http://sh.o.cn/ for ShangHai too

Sorry. I did understand what was being shown, even though I couldn't read the language. I was just making a joke about how the map looked like a SimCity game.

Actually, I was making two jokes, the second being how the map was out of date the instance it was published, because of how quickly a city's buildings and streets can change in China.

Still, it is very nicely done!

Sorry for my broken English, thanks for the elaboration.

I wonder if the whole city is a vector....

Damn. Now I want to live in one of those cities... they're so cuteee!

Where are the 1.3 billion people?

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact