Crop, Vegetation and Change Monitoring
Download brochure
GENERAL PRINCIPLES
The ability of the CosmoVision's proprietary VisCMSE (Visual Control, Monitoring
and Simulation Environment) software to combine monitoring, image processing
and modeling within the same application may be effectively used for
classification of the soil conditions and for issuance the recommendations
for planning of the agricultural activities. The proposed approach is based
on already existing software, which supports identification of the similar
ground surface cover areas using their spectral signature, derived from
the multispectral satellite image, and on standard set of the M2C
capabilities, as described in the M2C MODELING, MONITORING AND CONTROL
TECHNOLOGY brochure. Within this methodology and considering the fact that
M2C applications are designed to monitor and trend large numbers of
parameters that are related to large number of objects the following
approach was be taken:
(1) The monitored area is be separated into a set of objects (actually,
separate fields) following existing Land Cadastral boundaries.
(2) Within each object the area, which is free from the boundary effects is
defined, and monitoring application which includes all the mentioned
objects, weather data, fertilizer data, and vegetation type data will be
created.
(3) The optimal satellite / image provider shall be selected based on the
following criteria:
· Maximum number of separate bands
· High enough resolution to receive reliable reflection data without
boundary
interference
· Low enough resolution to keep the cost under control
· Long enough remaining period of operation to maintain data continuity (5
years at least)
The likely candidate is USA Landsat7, but the issue has to be revisited to
be more certain.
(4) After the provider is selected, the image of the same area using the
same instruments has to be taken each (preferably), or every second revisit
(every 24-36 or 48-72 days).
(5) The application will standby and collect weather data (from NOAA GMES or
Meteosat satellites), constantly updating the history files for weather,
etc., during the time period between images are taken.
(6) After next image is accrued, it has to be georeferenced and
automatically processed during the nearest monitoring update of the M2C
application. The spectral data (averaged over the sampling area) are added
to the history file for every object.
(7) After data added to the history file they are automatically analyzed
using built-in analysis model, and report with recommendations is issued and
automatically distributed using M2C notification features.
SYSTEM OPERATION
New image processing
The system operates in semi-automated regime and has to be deployed as a
separate application (possibly within the same master VisCMSE application)
for every monitored area. VisCMSE is deployed on the server computer that
has to be accessible (over LAN or WAN) for the satellite ground station for
downloading newly received images. As soon as new image arrives VisCMSE
informs the operator that new image is available (Fig.1, orange tab; audio
notification is optional).
The operator has to decide on weather the
particular image shall be used by the given
application. The options are:
· to move the image to the archive without integration in the particular
application,
· to integrate image into application as image for particular object,
· to integrate image into application as a map for particular object,
· to integrate image into application as external resource for particular
object
without changing the map and the image of the object.
In any case the image itself remains archived, and only network address is
passed to the application for retrieving the copy of the image from the
ground station’s archive.
If either use option (excerpt for archiving) is selected, the application
brings up the georeferencing workplace where newly arrived image is
pre-positioned according to the coordinates that are supplied by the image
processing unit of the ground station. The operator may either leave it as
it is, or refine the positioning of the image manually. After operator is
satisfied with image positioning, image is integrated into the system by
single button click. The entire process of integration of the image into
system takes less than 10 minutes.
Ground surface conditions and change
detection
If image is integrated as external resource, it becomes immediately
accessible for ground cover monitoring applications and for change detection
applications which may be integrated into individual nodes of the system.
The analysis is performed via comparing the new and the previous messages by
band pixel-by-pixel (with proper calibration) for change detection, and via
deriving average by band
over the object’s footprint area for ground cover / vegetation monitoring.
In the case of ground cover analysis new data are added to the data log for
particular object. These logs are the used to generate 3D (Intensity of
reflection vs. time and bandwidth) chart of the spectrum and of the spectrum
change for each monitored object.
The absolute (calibrated by the 1st band) values of the spectral intensity
are also available. Correspondent plug-in is activated by pressing the
“resource selector” button (large button at the top of the state analysis
panel).
Area history
Change history
In the case of change monitoring, the new image is
compared to the previous one and data are processed and presented by the
special plug-in application, that (similarly to ground cover) can be brought
up by pressing on the “resource selector” button. The user interface for the
change detection shows enlarged area with change spots marked, and has an
option to review the history of changes for the monitored area graphically
and as visual or printed report, which may also be automatically mailed
through standard email interface of the underlying M2C application.
Local change detection
In general terms any object can carry unlimited number of external resource
references, and if more than one is available the operator has to pick the
option of interest from the selector menu, that appears after clicking on
the resource selector button.
|