The gcp-osm from aerospaceresearch

improving the position marker positioning

Line 22 in fb79c6a

def get_position_marker_coordinates(bounding_box):

use another function to get the center of the qe-code position marker, aka bobble.

use contours or a center of gravity approach

Robust detection of rotated QR codes

The OpenCV QRCodeDetector seems to have its problems with rotated QR codes. We must analyze this further and maybe try something else for better detection rates.

handle the (java) exceptions for non-images

gcp-osm/main.py

Lines 16 to 18 in 1078220

    
           reader = zxing.BarCodeReader() 
        
           barcode = reader.decode(file) 
        
           # does not handle exceptions yet, when the file is NOT an image

when non-image files are given to the barcode reader, it thows java exceptions.

handle the exceptions, or filter out non-image files.

if there is no qr-code in a photo, the barcode reader returns "None". Maybe also return "None" here to use the same functions later own.

Checking the OpenDroneMap GCP format!

gcp-osm/main.py

Lines 69 to 78 in 46377aa

    
           # OpenDroneMap GCP 
        
           ## header 
        
           if odm_gcp_header == 0: 
        
               # this is still to understood, why ODM just allows one utm zone and that is in the header! 
        
               f.write("WGS84 UTM " + str(location_utm[2]) + str(location_utm[3]) + "\n") 
        
               odm_gcp_header = 1 
        
           ## line by line saving the coordinates 
        
           f.write((str(location_utm[0]) + " " + str(location_utm[1]) + " "  + item[3] + " " 
        
                    + str(points[1][0]) + " "  + str(points[1][1]) + " "  + file.split(os.sep)[-1])  + "\n")

how to handle GCPs in different, neighbouring utm zones?
is the separator within the line a "space" or a "tab"?
shall we allow an input from my_gcp_list without altitude?

src: https://github.com/OpenDroneMap/ODM/wiki/Running-OpenDroneMap

Supporting multiple QR codes per image

The current implementation only supports one QR code per image. It would be great to not just stop at the first found QR code but instead take them all into account.

Support analysis of directories

The prototype currently does only support analysis of single files. It would be nice to have an argument that allows providing a directory that contains images gcp-osm should analyze.

Position pattern detection accuracy

The positioning pattern is currently approximated in the easiest way possible.
It might make sense to use OpenCVs contours to detect the position pattern with more accuracy.

adding conversion from Base64 to Integer 64 for OSM ID

gcp-osm/main.py

Lines 65 to 75 in 00cec7f

    
           if parsed.find("/n/") > -1: 
        
               print("type indicator: n/ = OSM Node ID with Basis 64") 
        
           if parsed.find("/w/") > -1: 
        
               print("type indicator: w/ = OSM Way ID with Basis 64") 
        
           if parsed.find("/a/") > -1: 
        
               print("type indicator: a/ = OSM Area ID with Basis 64") 
        
           if parsed.find("/r/") > -1: 
        
               print("type indicator: r/ = OSM Relation ID with Basis 64")

adding the conversion from Base64 with ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789_~ to integer 64bit to convert to OSM id.
More here https://github.com/aerospaceresearch/gcp-osm/blob/master/README.md

What format shall we use?

gcp-osm/my_gcp_list.txt

Lines 1 to 3 in 46377aa

    
           5;50.0;10.0;0.0* 
        
           23;51.1;11.1;111.1* 
        
           42;52.2;12.2;222.2*

current format line by line
ID = any ascii
latitude = degrees in float
longitude = degrees in float
altitude = meters above ground

separator between is semicolon ;
ending the line with asterisk *

shall we use it?

Check what will happen when multiple QRs are in one image!

gcp-osm/main.py

Lines 44 to 50 in 00cec7f

    
           if os.path.isdir(args.file): 
        
               print("loading in all files in folder:", filename) 
        
               processing_files = gcposm.utils.get_all_files(filename) 
        
           elif os.path.isfile(args.file): 
        
               print("loading in this file:", filename) 
        
               processing_files = gcposm.utils.get_one_file(filename)

it is now working for one QR per image.
please check if it works when two or more QR codes are in one image.

aerospaceresearch / gcp-osm Goto Github PK

gcp-osm's People

Contributors

Stargazers

Watchers

Forkers

gcp-osm's Issues

improving the position marker positioning

Robust detection of rotated QR codes

handle the (java) exceptions for non-images

Checking the OpenDroneMap GCP format!

Supporting multiple QR codes per image

Support analysis of directories

Position pattern detection accuracy

adding conversion from Base64 to Integer 64 for OSM ID

What format shall we use?

Check what will happen when multiple QRs are in one image!

Recommend Projects

React

Vue.js

Typescript

TensorFlow

Django

Laravel

D3

Recommend Topics

javascript

web

server

Machine learning

Visualization

Game

Recommend Org

Facebook

Microsoft

Google

Alibaba

D3

Tencent

	reader = zxing.BarCodeReader()
	barcode = reader.decode(file)
	# does not handle exceptions yet, when the file is NOT an image

	# OpenDroneMap GCP
	## header
	if odm_gcp_header == 0:
	# this is still to understood, why ODM just allows one utm zone and that is in the header!
	f.write("WGS84 UTM " + str(location_utm[2]) + str(location_utm[3]) + "\n")
	odm_gcp_header = 1

	## line by line saving the coordinates
	f.write((str(location_utm[0]) + " " + str(location_utm[1]) + " " + item[3] + " "
	+ str(points[1][0]) + " " + str(points[1][1]) + " " + file.split(os.sep)[-1]) + "\n")

	if parsed.find("/n/") > -1:
	print("type indicator: n/ = OSM Node ID with Basis 64")

	if parsed.find("/w/") > -1:
	print("type indicator: w/ = OSM Way ID with Basis 64")

	if parsed.find("/a/") > -1:
	print("type indicator: a/ = OSM Area ID with Basis 64")

	if parsed.find("/r/") > -1:
	print("type indicator: r/ = OSM Relation ID with Basis 64")

	if os.path.isdir(args.file):
	print("loading in all files in folder:", filename)
	processing_files = gcposm.utils.get_all_files(filename)

	elif os.path.isfile(args.file):
	print("loading in this file:", filename)
	processing_files = gcposm.utils.get_one_file(filename)