My goal is to find a way to download (with Python) satellite images given coordinates describing a rectangle. I've never really found a precise and free solution (no business here, just school stuff).
At first I tried Google Maps' API, which worked perfectly but turned out to be paying after a certain time. I then considered using OpenStreetMap, but again I had a lot of trouble finding information on how to obtain those.
Can you please help me with a simple solution?
OpenStreetMap only provides map data. OpenStreetMap doesn't have aerial imagery and thus also no satellite imagery API.
If you are looking for free aerial imagery then take a look at OpenAerialMap.
Related
I have recently been looking at using Bing Maps API to return images of properties and was very interested in the oblique imagery provided by Bing.
It seems that the API does offer this imagery and there are examples of this working. However, when I go to implement it myself, I find that the birdseye view does not return the building of interest. It returns a view off centre of the point put into the url.
An example of this is the place below, which shows the URL for looking at the Leadenhall building in City of London:
aerial_url = http://dev.virtualearth.net/REST/V1/Imagery/Map/Aerial/51.5138,-0.0821/18?&key={api_key}
birdseye_url = http://dev.virtualearth.net/REST/V1/Imagery/Map/BirdsEyeV2/51.5138,-0.0821/18?&key={api_key}
You will notice that the build is top right of the second image, rather than centre as in the first image.
Is anyone able to help me resolve this, as ideally I would want it to be centre of the image?
Thanks
I took a look at it, and it looks to be centered correctly. It's hard to tell comparing at the same zoom level. Setting the aerial image to zoom level 17 helps provide a better overview so you can better match the images. Note that the coordinates are based on the ground, not building heights. Aerial images are captured looking straight down while birdseye are captured at approximately 45 degree angle.
I have a pair of coordinates (lat, long).
I need to generate an image of displaying these coordinates on the map.
And then generate such images with other coordinates in the future without the Internet.
Please tell me whether there are solutions that allow you to display coordinates offline?
Upd: is there any opportunity to download maps offline , eg: gps tracker maps or something like that?
thank you
This is not possible while offline. To generate an image of the coordinate location you would most likely be using
os.system("open \"\" https://www.google.nl/maps/place/" + location)
and then generating a image of the location that is popped up. This is impossible to do while offline, I am very sorry.
The question is too broad to give a good answer. However:
There are several companies, such as TeleAtlas and NavTeq, that sell map data. I have no idea what buying the world from them at 1:1M resolution would cost, but I'd guess several thousand USD.
You could download data, or pre-rendered rasters, from Natural Earth. However, they don't have quite the resolution required for good 1:1M maps.
You could download data from OpenStreetMap. The data is free (as in beer, and as in speech), but using it is a major undertaking.
There are companies that offer pre-rendered maps in various formats from OpenStreetMap data. OpenMapTiles is the one I happen to have at the top of my head, but here are others.
I am building a app to see the progress of deforestation. Over time i would like to take a satellite image from a location and see what percentage of that image contains forest.
I have attempted google's vision API, it does not have this functionality.
Is this something that can be done in OpenCV or must I do this from scratch with semantic segmentation or something similar?
So far I could see in the documentation there doesn't seem to be any pattern/texture recognition for the API. My belief is that you could try to do a dominant color recognition. if your image data has enough data of differentiable colors, I think you should be able to get an acceptable analysis.
PD: Having some experience with satellital imagery processing, as additional info, I can comment that the usual way to find out the status of the land for plants, forest and general crop development and health is though color analysis.
Nonetheless, Satellite/drone images are mostly multispectral and several UV bands are extensively used as biomass behaves very different with season/health/development status with the combination of visible and UV electromagnetic bands.
Have you tried to look at the satellite image recognition Kaggle competitions? There are a lot of discussions as well as available scripts for tasks similar to yours:
Links: https://www.kaggle.com/c/dstl-satellite-imagery-feature-detection
Example script: https://www.kaggle.com/arpandhatt/satellite-image-classification
What I want to do is to generate a static image (e.g. a png) using python and using openstreetmap tiles as a background.
Mathplotlib and Basemap is almost what I'm looking for. The problem is being able to use OSM tiles as background. I'm not pleased by the approach suggested in http://stevendkay.wordpress.com/2010/02/24/plotting-points-on-an-openstreetmap-export/
The closest I found is in this answer but using R, and not python Plotting points from a data.frame using OpenStreetMap
Did I miss any obvious and easy solution?
Thanks for your help
EDITÂ : this questions suggests many tools, but none seems to match my needs How can I display OSM tiles using Python?
You overlooked the "Export" tab at the OSM website, which is capable of generating a static image with the dimensions and map extents you want. Have a look at http://wiki.openstreetmap.org/wiki/Export
Please be advised that generating static images is a resource-intensive process, and the OSM sysadmins will frown upon you if you do a large number of requests or abuse this feature. Unfortunately this means you'll have to find another solution if you're trying to do lots of images.
By the way, the data you're plotting on top is properly projected into EPSG:3857 and not just raw lat/lon coordinates, right? Raw lat/lon data will look distorted at large zoom levels.
I have been asked to publish a complete book online similar way Google Books does? i.e. it's viewable and printable but not download-able.
Is the process is basically "high quality scanning"? are there any open source solution to "mass generation" of "watermark" on those high quality images. Suppose you have an original image. and when the user views it online, I re-create the image add watermark and some other text on top of the image "on-the-fly" are there such library exist in python off course :)
Any tips? If you have done this before please share.
Thanks
Unfortunately Google uses a patented technique for scanning it's books, so you will probably have to stick to traditional methods.
Google created some seriously nifty
infrared camera technology that
detects the three-dimensional shape
and angle of book pages when the book
is placed in the scanner. This
information is transmitted to the OCR
software, which adjusts for the
distortions and allows the OCR
software to read text more accurately.
No more broken bindings, no more
inefficient glass plates.
Basically you will need to scan the book using an OCR application (tesseract is good), then I would generate a PDF/image from the scanned text, and finally add the watermark on top. The Python Imaging Library would seem to be the best tool for this.
Don't know much about Google Books, but Python Imaging Library can do watermarking (there's ASPN recipe for that).
See the slashdot question on reproducing Google's photo + laser grid technique.