| Railtrack
Mapping and Monitoring |
|
Introduction
This
railtrack multi-camera system can be configured for almost any
coarse mapping or monitoring operation within the rail industry.
This system can be used to record location of trackside assets
(points, signal lights, boxes), infrastructure (buildings, bridges)
and other rail corridor objects (work sites, vandalism, rubbish)
using low-cost, state of the art spatial video technology.
The
railtrack video mapping systems is
Inexpensive
Rapid mobilisation
Easy to install and use
Advanced data processing enabling fast turn around time of data
|
Image
depicting positioning of object in front of locomotive
|
Positional
accuracy is typically within the 3m to 10m range depending on
navigation used. Image quality will depend on camcorder quality
and speed of acquisition but expect reasonable quality images
(768*576) acquired at speeds of 55 mph. Video cameras can be calibrated
and so allow objects and features within the image to be positioned
and measured. This calibration is based on a monoscopic camera
model using a perspective transformation algorithm. |
| There
are three stages when carrying out a railtrack video mapping project;
acquisition, processing and browser development.
|
| Acquisition
This
system is constructed using three camcorders and one CamNav
unit. Most high-street camcorders are suitable (Video8, S-VHS,
Hi8, DV) as long as the camcorder has an external mic input
to enable the GPS signal to be encoded on one of the audio tracks
(normally two on standard camcorders). Alternatively, higher
quality optical sensors together with DV VCR can be used. These
have the advantage of higher quality images and longer play
tapes.
Power
consumption for this multi-camera configuration
is usually quite low and can be usually operated using
batteries or in-cab cigarette-lighter. A mount is normally constructed
to maintain the cameras in a stable position inside the cab
window. This mount contains separate screw-in plates for each
individual cameras. These can be adjusted enabling the camera
to be rotated in pitch, roll and yaw. GPS encoded data is recorded
onto each camcorder. A green light indicates that the GPS fix
is good and recording is operating correctly.
|
Schematic
of multi-camera equipment for rail corridor mapping
|
|
Processing
The
encoded data is processed using a frame-grabbing card and a
suite of software located on a standard PC machine. The tape
containing the encoded GPS is now decoded using the Camcorder,
which is connected to the CamNav. Two output cables containing
the video and GPS data are then connected to the frame-grabbing
card and RS-232 serial port respectively.
This
is usually a two stage process.
Processing
- Stage One
Frame-grabs
frames at a rate chosen by the operator. These frames are tagged
with GPS time and saved onto hard disk. At the same time, the
GPS data is logged at the serial RS-232 port and saved in a
file.
|
|
Screen-grab
of spatial video processing system.
|
|
Processing
- Stage Two
The
second stage involves de-interlacing the digital video images
and producing digital images 768*576 in size. An ESRI 'shape'
navigation file is constructed from the logged GPS data. This
file contains all the navigation data together with references
to image names. It will take approximately 1.5hr of computer
time to process one hour of spatial video data. Most of this
time is required by the frame-grabbing module, which can only
operate at normal video speeds.
|
| Browsing
The
browser integrates the various video, GIS, database, measurement
modules into an easy to use lightweight application. The main
application window is split into video (right), GIS (left) and
database (bottom left) displays. The user can load up any survey
using the pull-down menus. A metadata layer is usually constructed
beforehand and enables various surveys to be loaded with minimum
fuss.
|
|
Screen-grab
of a ‘basic’ browser system
|
| The
user can navigate through the spatial video data using the video
controls, or by double-clicking on some point on the map or by
clicking on a record within the database module. The video plays
in the main video display as well as the three thumbnail displays.
The position of the locomotive on the track is dynamically displayed
on a moving map. The usual pan/zoom/identify functions are available.
Replay speed can be adjusted and various views can be turned on
or off. Various left, middle, right views can be chosen by clicking
on the any one of the thumbnails below the main video display.
Positions of trackside objects can be measured using a perspective
transformation algorithm
(based on a monoscopic camera model). This feature together
with a data input module enables railtrack asset to be entered
into a database. Full spatial and imaging geometry of the object
or feature are also stored. This allows the object to be quickly
displayed later once the user double-clicks on the database record.
This database together with associated video images can be easily
exported to other mapping/imaging system and/or databases.
|
|
CamNav Mapper™
Multi Camera Services
|
|
Blueglen
prefer to work with the client to devise the most cost-effective
and 'fit for purpose' solution for client's application.
Acquisition
The client
can acquire the CamNav Encoder/Decoder box and carry out the
data acquisition themselves. Alternatively,
BlueGlen can carry out any type of survey on your behalf.
Regardless, BlueGlen has a wealth of experience in choosing
cameras, recorders, mounts, power distributionunits and can
help the client decide on best options.
Processing
The multi
camera processing suite can be purchased, rented or BlueGlen
can undertake the processing on behalf of the client.
BlueGlen would provide remote or on-site support should the
client purchase the processing module.
It is
possible to setup a number of processors in parallel, to process
multiple videotapes at the same time.
Browser
A single
browser licence is included with a multi camera purchase.
Any number of extra licenses can be obtained or a corporate
licence might be a more suitable option.
Normally
a browser requires some customisation.
Level
1 Brow ser
Simple
digital interactive map and GPS encoded video viewing system
that allows the client to view data using PC based video control
buttons and a 'moving map' display.
Level
2 Browser
Same
as level 1 plus a database interface for data input, in-frame
measurement, automated report generation, image processing functions.
|
