Subscribe to this thread
Home - General / All posts - New video: Georeference a Scanned Paper Map
Dimitri


7,413 post(s)
#04-Mar-21 12:13

There's a new video on the videos page: Georeference a Scanned Paper Map

Description:

In only five minutes of actual work we use Manifold Release 9 to georeference a 157 MB scanned paper map so it can be used as a layer in GIS. The scanned map is a historic map showing Davy Crockett National Forest in Texas, downloaded from the Library of Congress website.

Compare the speed and ease of use of georeferencing in Manifold to the slowness and inconvenience of georeferencing in ArcMap or ArcGIS Pro. A YouTube video from an ESRI user shows this exact same scanned paper map being georeferenced in ArcMap. Compare Manifold workflow to ESRI workflow and you'll agree it's a lot easier, quicker, and less confusing in Manifold. Fast GIS is fun GIS! All this works in the free Manifold Viewer too.

dchall8
1,008 post(s)
#04-Mar-21 22:22

You're definitely a trooper for trying to pronounce names of communities in Texas. I'm not familiar with the Crockett Park area to be of any help, but you're taking your chances when trying to speak local. For example

Gruene = green

Boerne = BER nee

Palestine = PAL iss teen

New Berlin = new BER lin

Waxahatchie = woks uh HATCH ee

Mexia = muh HEE uh

Leakey = LAKE ee

Burnet = BURN it

Lake Buchanan = lake buk AN un

And for the ESRI guy, it's CROK it, not crow KET. At the same time, please don't ask me to pronounce any community names in South Asia.

RonHendrickson
283 post(s)
#06-Mar-21 14:43

I agree with all except Mexia - should be muh HAY uh

If I may be permitted, the Palestine name reminds me of an old Texas joke in my Texas church:

A man asked a minister if he could join their church. The minister said yes if you can answer a few questions. He asked the man "What geographical area did most of the Bible happen in?"

The man said he didn't know and asked the minister for the correct answer. The minister said "Palestine".

The man exclaimed "I knew it was somewhere in East Texas!"

dchall8
1,008 post(s)
#05-Mar-21 05:55

More seriously, I've taken a stab at georeferencing a map of San Antonio from 1909. Most of the roads have the same name, but it seems many of the important road names have been changed. Some entire neighborhoods have disappeared to urban renewal. In fact the neighborhood I originally tried to locate is now a mix of freeway and freeway tailings piled up into a hill.

At first I relied on railroad tracks to be my "truth", my unchanging base. Well, it turns out tracks can be abandoned and, essentially moved. Some of the original road intersections were imperfectly aligned as two T intersections. Over the years those intersections have been corrected into a + intersection by bending something. The thin-plate spline method of georeferencing will push, pull, warp, stretch, or buckle the original map to match the control points. If something unexpected has changed in an area of interest, you'll see it in the curved roads around the area.

hphillips31 post(s)
#05-Mar-21 15:52

Its unlikely you would have missed this as a resource for your San Antonio map but I see that USGS has historical USGS 24K topos circa 1953 as GeoTIFF, San Antonio East and San Antonio West that would be helpful for georeferencing the 1909 map you are working on, more useful than current maps perhaps.

Attachments:
san_antonio_usgs.jpg

dchall8
1,008 post(s)
#05-Mar-21 18:00

Thanks for that. I didn't do any research beyond the Texas General Land Office. Your link might have answered my question directly, since most of the urban renewal has taken place after the early 50s.

dchall8
1,008 post(s)
#05-Mar-21 23:55

We're having fun now!

The USGS maps don't have street names, but that's not hard to figure out. It turns out they map pretty close to the current Bing and Google imagery, but not close enough for my OCD. There are four quads that cover what is now San Antonio. As it happens, the east and the west maps meet just about where I wanted to draw in the 1909 neighborhood, but again, not hard to figure that out. They all come in as polyconic projections as is stated on the maps themselves. Here's a picture of where the east and west maps meet.

The red/yellow box/dot in the middle of the picture is the location where the corners of the two maps are supposed to meet, according to the notation on the maps. The location to the south is the NE corner of the SA West map as it imported. I suspect Tim would recognize this as a problem with the original projection. I'm not that clever with recognizing those things, plus I was looking for an excuse to use the georegistration tools, so that's what I did. As we say, there is more than one way to skin a cat...although there is probably a preferred way. At first I adjusted the lat/long locations as described in the video about georegistering the entire Earth. When that wasn't perfect, I used monuments on the ground, and that worked very well. At most I needed 5 points including four at the outer corners and one on the interior.

These 1954 maps do align much better to today than the 1909 map did. I suspect the surveying equipment and skills had improved. In 1954 they used aerial surveys which were not available in 1909.

Attachments:
Errors in USGA San Antonio GeoTif Maps.jpg

hugh
200 post(s)
#06-Mar-21 21:38

You may be interested in these two aerial photos, some of many images georeferenced in this file:

mf5_m9.mxb

The one above shows a small area near El Portal in northeast Miami-Dade County Florida near Biscayne Bay. The gray area on the left is an extension of the Everglades with vegetation growing in shallow water (1 to1 1/2 meter depth, water surface ~ 2 meters above sea level). That ended around the middle of the photo and the water drained through rapids into Biscayne Bay. Before the end of this century, the water level will be back to 2 meters with sea-level rise.

The second 1928 one below (lower resolution with more artifacts and some streets apparently drawn on the photo) was taken after a canal was dredged draining this part of the Everglades.

I did the georeferencing of the images in the file and many others using M8. Most of the historical ones do not have to be very precise, but at some point soon I will be doing some tweaking in M9.

This work is part of a citizen science effort that is a part of the Miami-Dade SLR adaptation process that was announced last week: NYT article It probably won't succeed since it requires an economic transition to mostly workforce housing when SLR risk ends the speculative real estate market.

Attachments:
1924.jpg

hugh
200 post(s)
#07-Mar-21 04:47

Sorry I just talked about what these images were of and not all the georeferencing problems I had to figure out. But that was with M8 and have not started working on it in 9. You will probably hear from me then on this wonderful forum where Manifolders at any skill level know they will will get serious help. -- Hugh

Manifold User Community Use Agreement Copyright (C) 2007-2021 Manifold Software Limited. All rights reserved.