Description
Smart camera arrangement and a method thereof
This invention relates to a smart camera application, in
particular to the testing and deploying of said smart camera
application.
A smart camera is an integrated machine vision system which,
in addition to image capture circuitry, includes a processor,
which can extract information from images without need for an
external processing unit, and interface devices used to make
results available to other devices. Due to recent
advancements, processing and memory capabilities of smart
cameras have increased considerably enabling complex
algorithms to be run on them. Smart camera networks are
popularly deployed in applications where cameras are required
to be geographically distributed since a single camera cannot
track mobile objects. Further, applications running on the
smart cameras will require inputs from other smart cameras in
the network to analyze images cohesively. Therefore, smart
cameras should co-ordinate among themselves much like
computation nodes in a distributed environment. Distributed
smart cameras find applications in video surveillance and
traffic control. To develop distributed applications the
developer will have to take care of aspects which are very
important in a distributed environment, like fault tolerance
and concurrency. In addition, developing smart camera
applications involves interfacing with the camera for image
acquisition, developing the business logic to process these
images and then sending the meta-data obtained from the
processing phase over the network as inputs to other smart
cameras. To accomplish these tasks the application developer
will have to build distributed system aspects in the
applications and also interface with the hardware elements of
the embedded environment that the smart camera is using.
Therefore the development is very specific to a distributed
network and the particular smart camera being used.

In the current scenario a pre-defined sequence of images are
used to verify distributed smart camera applications. The
individual clocks of the cameras in the network are also
synchronized. Testing and verifying a distributed smart
camera application, in the present scenario, is very
difficult because the emulators that are available currently
in Integrated Development Environments (IDE) emulate only the
core or the processor and do not emulate peripherals like the
camera and the network. Therefore, to test and verify all
functional modules of a distributed smart camera application
completely, the application will have to be deployed on the
target hardware. If any errors or bugs are found then they
have to be resolved in the development environment and then
the application is redeployed on the actual hardware for
functional testing. Further, in the current setup
applications are developed to run on a single device and then
if necessary, the applications are extended to run on a
network of devices. The application developer has to develop
high level application concepts over primitive facilities
that the target hardware provides. In a typical scenario, the
developer first develops the application using dedicated
development environments on a commercial off-the-shelf
Operating System. Then the application has to be deployed on
the target hardware. The developer will have to be aware of
the features that the hardware manufacturer provides to
interface effectively with the hardware. In effect,
development is done partly on the target hardware itself
making porting to other smart camera hardware difficult. The
complicated development environment and the lengthy
development cycle make developing smart camera applications
in this environment cumbersome and error prone.
It is an object of the present invention to improve testing
and deploying distributed smart camera applications.
The above object is achieved by an arrangement for testing
and deploying an application for a plurality of target smart
cameras in a network comprising:

an emulator module for testing and debugging of the
application said emulator module adapted to simulate an
environment comprising a plurality of virtual smart
cameras connected over at least one virtual network and
each virtual smart camera running at least one image
processing algorithm corresponding to an image
processing algorithm in the respective emulated target
smart camera; and
a framework executable on the target smart cameras, said
framework comprising modules adapted to provide an
interface to the application being deployed on said
target smart cameras.
The above object is achieved by a method for testing and
deploying an application for a plurality of target smart
cameras in a network comprising:
performing testing and debugging of the application by
an emulator module adapted to simulate an environment
comprising a plurality of virtual smart cameras
connected over at least one virtual network and each
virtual smart camera running at least one image
processing algorithm corresponding to an image
processing algorithm in the respective emulated target
smart camera; and
executing a framework on the target smart cameras
wherein said framework (34) interfaces the application
being deployed on said target smart cameras.
The suggested approach is to use a simulated environment to
develop, test and debug distributed smart camera
applications. A distributed smart camera emulator module
emulates a network of smart cameras (generally on a single
machine). The emulator has virtual cameras and a virtual
network. A virtual camera simulates the behavior of a smart

camera. The emulator module thus enables the developer to
test and verify the functional requirements of a distributed
smart camera application. Further, the emulator allows
different kinds of network to be formed. The emulator reduces
the overall time required to test and verify a distributed
smart camera application. The arrangement also comprises a
framework which encapsulates all modules that are common when
an application is developed over a particular hardware. Each
target smart camera has to have the framework running on it
so as to abstract away all hardware details and utilize the
distributed computing algorithms. This framework helps in
deploying the application on the target application. This
arrangement helps an application developer to develop on a
commercial off-the-shelf operating system, a more
sophisticated environment that the developer is comfortable
with, rather than an embedded environment. The developer is
not required to know about the camera and network primitives,
that each embedded environment exposes, in detail because the
framework interfaces with them and allows the developer to
use them in a consistent manner.
In a preferred embodiment of the invention the modules within
said framework further comprise a sensor interface module
adapted to capture images. This module provides a common
interface enabling the capturing of images on any camera in a
uniform manner.
In an alternative embodiment the modules within said
framework further comprise a core module adapted to integrate
the image processing algorithms into the target smart camera.
This is the module where the main business logic i.e. the
image processing algorithm is plugged in. The image
processing algorithms that are supposed to run on the smart
camera are integrated into this module. This ensures that the
application when deployed on the target hardware will
function in the pre-ordained manner i.e. the developer will
not have to make any further changes for the code to be
deployed on the target hardware.

In another alternative embodiment the modules within said
framework further comprise a peer module adapted to manage
the target smart cameras in the network using at least one
distributed algorithm. This module enables to manage the
nodes or smart cameras that are available on the network. The
distributed algorithms required to manage the network are
made available in this module. This considerably reduces the
time taken to develop, test and verify a distributed smart
camera application.
In another alternative embodiment the emulator module is
adapted to use stored images as input for the virtual smart
cameras to generate data to communicate across the virtual
networks. This enables the testing of a wide range of images
prior to deployment of the application on the target cameras.
The emulator module is used to test the functional
correctness of the application by using the stored images
(which could be a recorded or synthesized images) thereby
ensuring proper simulated environment prior to actual
deployment.
The present invention is further described hereinafter with
reference to illustrated embodiments shown in the
accompanying drawings, in which:
FIG 1 is a block diagram illustrating the arrangement
according to an embodiment of the present invention,
FIG 2 illustrates the components of a smart camera framework,
and
FIG 3 is a flowchart illustrating the testing and deploying
of the smart camera application according to an embodiment of
the present invention.
The invention aims at developing distributed smart camera
applications on a commercially available off-the shelf

operating system using standard software development tools
and environment. The application developer does not have to
port the application manually because the framework running
on the target hardware ensures that the application functions
in a consistent manner. The emulator, by simulating smart
cameras connected over the network, provides an environment
to test and verify the application. The time taken to
develop, test and verify a distributed smart camera
application is considerably reduced because the time taken to
port the application is eliminated almost completely.
FIG 1 is a block diagram illustrating the arrangement
according to an embodiment of the present invention. This
arrangement is used for testing and deploying an application
2 for a set of target smart cameras 4, 6, 8, 10 in a network
12. The arrangement comprises of an emulator module 14
adapted to run in a computer machine for testing and
debugging of the application 2. The emulator module 14
simulates an environment comprising a set of virtual smart
cameras 16, 18, 20, 22 connected over a virtual network 24.
Each virtual smart camera runs at least one image processing
algorithm 26, 28, 30, 32 corresponding to an image processing
algorithm running in the respective emulated target smart
camera. The image processing algorithm could all be the same
or different depending on the required application and the
cameras used. The target cameras are shown running a
framework 34. The time required to develop a distributed
smart camera application using the framework is reduced
because the time required to familiarize with the hardware is
eliminated. Further, testing can be accomplished in the same
real environment which gets simulated.
FIG 2 illustrates the components of a smart camera framework.
The framework 200 comprises modules adapted to provide an
interface to the application 205 and the hardware 210 being
deployed on said target smart cameras. The virtual cameras
for example, can run on a high level COTS OS because there is
a software abstraction in the form of a smart camera

framework that is being provided. The framework encapsulates
all modules that are common when an application is developed
over a particular hardware. It has modules which interface
with the hardware to take care of image acquisition,
communication and distributed computing issues. The framework
200 is shown to comprise of a sensor interface module 215,
where said module interfaces with the Software Development
Kit provided by the camera manufacturer to provide a common
interface enabling the developer to obtain images on any
camera in a uniform manner. The framework 200 further
comprises a core module 220 where the business logic is
plugged in. The image processing algorithms that are supposed
to run on the smart camera are integrated into this module.
The framework further has a communication module 230 that
enables transmission of data across the network. The
communication module 230 interfaces with the network stack
provided along with the hardware 210. The framework has a
peer module 225 which manages the nodes or smart cameras that
are available on the network. The distributed algorithms
required to manage the network are made available in this
module. Each smart camera has to have the framework running
on it so as to abstract away all hardware details and utilize
the distributed computing algorithms that reside in the
framework. The framework can be installed and configured for
any hardware environment. So the application developer can
develop applications, for example on a COTS OS, test them
using the emulator and then the code can be directly ported
to a smart camera network without making any changes. The
smart camera framework therefore enables the application
programmer to use the hardware facilities, irrespective of
the development environment, in a consistent manner.
FIG 3 is a flowchart 300 illustrating the testing and
deploying of the smart camera application according to an
embodiment of the present invention. At step 305 an emulator
module is launched which will simulate an environment where
there are several smart cameras connected over the network,
each running the image processing algorithms. At step 310 the

simulated environment is tested. At step 315 any occurrence
of an error is detected and said detected error gets debugged
at step 320. The simulated environment is further tested for
possible errors if required. When all tests are cleared then
the application is ported to the target hardware at step 325
without making any changes because the framework is already
running on the hardware. The time required to develop a
distributed smart camera application using the framework is
reduced because the time required to familiarize with the
hardware is eliminated. Further, testing can be accomplished
in the same environment. Therefore deployment can be done
after full functional verification of the application. The
steps like cross compiling and deployment of said application
on the hardware several times for testing and for each and
every error detected could be avoided by using said emulator
module and the framework.
Summarizing, the present invention relates to an arrangement
for testing and deploying an application 2 for a plurality of
target smart cameras 4, 6, 8, 10 in a network 12. The
arrangement comprises of an emulator module 14 for testing
and debugging of the application 2. The emulator module 14 is
adapted to simulate an environment comprising a plurality of
virtual smart cameras 16, 18, 20, 22 connected over at least
one virtual network 24. Each virtual smart camera runs at
least one image processing algorithm 26, 28, 30, 32
corresponding to an image processing algorithm in the
respective emulated target smart camera. The arrangement
further comprises of a framework 34 executable on the target
smart cameras, said framework comprising modules adapted to
provide an interface to the application being deployed on
said target smart cameras.
Although the invention has been described with reference to
specific embodiments, this description is not meant to be
construed in a limiting sense. Various modifications of the
disclosed embodiments, as well as alternate embodiments of
the invention, will become apparent to persons skilled in the

art upon reference to the description of the invention. It is
therefore contemplated that such modifications can be made
without departing from the spirit or scope of the present
invention as defined.

We claim,
1. An arrangement for testing and deploying an application
(2) for a plurality of target smart cameras (4, 6, 8, 10) in
a network (12) comprising:
an emulator module (14) for testing and debugging of the
application (2) said emulator module (14) adapted to
simulate an environment comprising a plurality of
virtual smart cameras (16, 18, 20, 22) connected over at
least one virtual network (24) and each virtual smart
camera running at least one image processing algorithm
(26, 28, 30, 32) corresponding to an image processing
algorithm in the respective emulated target smart camera
(4, 6, 8, 10); and
a framework (34, 200) executable on the target smart
cameras, said framework comprising modules adapted to
provide an interface to the application (2) being
deployed on said target smart cameras (4, 6, 8, 10).
2. The arrangement according to claim 1, wherein the modules
within said framework further comprise a sensor interface
module (215) adapted to capture images.
3. The arrangement according to any of the preceding claims,
wherein the modules within said framework further comprise a
core module (220) adapted to integrate the image processing
algorithms (26, 28, 30, 32) into the target smart camera.
4. The arrangement according to any of the preceding claims,
wherein the modules within said framework further comprise a
peer module (225) adapted to manage the target smart cameras
(4, 6, 8, 10) in the network (12) using at least one
distributed algorithm.
5. The arrangement according to any of the preceding claims,
wherein the emulator module (14) is adapted to use stored

images as input for the virtual smart cameras (16, 18, 20,
22) to generate data to communicate across the virtual
networks (24).
6. A method for testing and deploying an application (2) for
a plurality of target smart cameras (4, 6, 8, 10) in a
network (12) comprising:
performing testing and debugging of the application by
an emulator module (14) adapted to simulate an
environment comprising a plurality of virtual smart
cameras (16, 18, 20, 22) connected over at least one
virtual network (24) and each virtual smart camera
running at least one image processing algorithm (26, 28,
30, 32) corresponding to an image processing algorithm
in the respective emulated target smart camera; and
executing a framework (34, 200) on the target smart
cameras (4, 6, 8, 10), wherein said framework (34)
interfaces the application (2) being deployed on said
target smart cameras (4, 6, 8, 10).
7. The method according to claim 6, further comprising
executing a sensor interface module (215) within said
framework (34, 200) for capturing images, and
executing a core module (220) within said framework (34,
200) for integrating the image processing algorithm (26,
28, 30, 32) to said target smart cameras (4, 6, 8, 10).
8. The method according to any of the claims 6 or 7, further
comprising executing a peer module (225) within said
framework (34, 200) having at least one distributed algorithm
for managing the target smart cameras (4, 6, 8, 10) in the
network (12).
9. The method according to any of the claims 6 to 8, wherein
the emulator module (14) use stored images as input for the

virtual smart cameras (16, 18, 20, 22) to generate data to
communicate across the virtual network (14).
10. An arrangement or method substantially as herein
described and illustrated in the figures of the accompanying
drawings.

The present invention relates to an arrangement for testing and deploying an application (2) for a plurality of target smart cameras (4, 6, 8, 10) in a network (12). The arrangement comprises of an emulator module (14) for testing and debugging of the application (2). The emulator module (14) is adapted to simulate an environment comprising a plurality of virtual smart cameras (16, 18, 20, 22) connected
over at least one virtual network (24). Each virtual smart camera runs at least one image processing algorithm (26, 28, 30, 32) corresponding to an image processing algorithm in the respective emulated target smart camera. The arrangement further comprises of a framework (34) executable on the target smart cameras, said framework comprising modules adapted to provide an interface to the application being
deployed on said target smart cameras.