Page 151 - Kaleidoscope Academic Conference Proceedings 2020

Industry-driven digital transformation

through which information about plants in the nature doing the task, the moderator was observing without assisting
reserve are accessed. the user. The second test required the user to re-launch
the application, scan the marker, use other functions of the
• Output: The output part is retrieved from the database
application, and close it afterward. The performance results
after image processing has been completed or when
of the user were noted down by the moderator.
points of interest are identified. The overlaid information
will be displayed on the user’s mobile phone screen as
4.1.3 Device Performance Test
output.
3. SYSTEM IMPLEMENTATION Two categories of tests were carried out. The first was a
device compatibility test, which was used to validate that
An agile methodology was used in developing the application. the application behaves as expected across different mobile
Pictures of various plants in the Cape Flats Nature reserve devices running the Android operating system. Application
were taken and uploaded to an online database. Our functions and features were also tested as components might
AR markers were designed using AR.js an open source behave differently on different devices. The last test case was
platform for designing AR-markers [9]. Vuforia SDK, Unity to check device compatibility with regard to installation in
and Adobe Stack [10],[11] were used to develop the AR order to check whether the application could be installed on
mobile application. The scope of this project only focused other devices..
on Android-enabled mobile phones and developed using
The second category of tests was performance testing. Here
Android Studio.
we tested the speed of the application by checking how
The Samsung Galaxy S8 mobile phone [16] was used for
quickly the application is launched, how long scans took,
testing purposes. The S8 was chosen because of its high
the length of time it took to move across pages and to exit
screen resolution (2960x1440), 12MP main camera with focal
the application. We also tested the application’s scalability,
length of 35mm and an aperture of f/1.7. With its Snapdragon
to determine the maximum number of users the system could
835 processor coupled with 4 GB of RAM, it was sufficiently
support concurrently.
capable of running the AR app smoothly with minimal delays
or processing bottlenecks.
4.2 Test Reports
4. SYSTEM TESTS AND RESULTS DISCUSSIONS
4.2.1 Unit Testing
A number of tests were carried out, including a unit test,
[12],[13], user acceptance test and device performance tests. While the test was running, the status bar at the top of the Test
For the unit test the focus was on the code and was carried Explorer window in Microsoft Visual Studio was animated.
out using Microsoft Visual Studio. For user acceptance, the At the end of the run, the bar turned green if all the test
focus was on user experience of the mobile app and to ensure methods pass, or red if any of the test failed. In the event of the
that the app met usability and accessibility expectations [14]. test failing, we would select test explorer to view the detail at
Users were allowed to use the application and qualitative the bottom of the window. All detected errors were corrected
data was subsequently collected in a bid to improve the forall four classes. This processinvolved repeated debugging,
users’ experience. The final test was geared towards device recompiling and reruns until all required expectations are met
compatibility and performance. For this test, the application and pass the test.
was installed on a number of phones to running Android 4.1
(or higher) to verify compatibility. The various test cases are Table 1 – Unit testing
discussed in the following subsection.
Iteration Test Results Bug fixes
4.1 Test Cases First Compilation error Modified the
method "Create
4.1.1 Unit Testing Single Pixeltexture"
Second Long compilation Replaced if
This test was performedby the developer using amethodology time statements with
called code first whereby a thorough testing of the nested loops
Third Inconsistent results modified
functionality of the code was made to make sure the code
"OnRenderObject"
performs as it is required to. All the methods and classes that
Fourth Success No change required
are used in the application were tested here to make sure they
pass this test before deployment.
4.2.2 User Acceptance Test
4.1.2 User Acceptance Test
The results for tests carried out are summarized in tables 1 to
Two tests were carried out, in the first test users were given the
3. In terms of user experience, the users were happy with the
mobile phone and allowed to search through the phone, find
application as it was user-friendly and easy to navigate.
the application, open and then close it. While the user was

– 93 –