Abstract:
For visually impaired people, image base methods are a new approach for navigation problem solving.
Approach:
Electronic crane is a new approach for blind people using the environment represented as a weighted
topological graph.
Results:
The system gives advices for the blind person to select the right direction in indoor navigate depending on
weights and session.
Introduction:
The goal of blind people is that to navigate through unfamiliar spaces without the human guide help. In
order to achieve this aim, they are using many methods and devices. There are two essential points to
navigate the blind person in the environment; first is appropriate information about the travel path, and
the second, is recognition of objects through the travel path which follow. These points are a wide area of
interest for the researchers. We proposed a model that can help blind people to navigate anywhere,
depending on the weight between the captured and stored scenes with in the session of the vision, where
the new session of vision will be started with the relevant objects inside this session and the site will be
identified to the mobile computer and also the blind person will be informed this session with this idea we
have kitchen, bathroom, bedroom sessions. Navigating from one place to another, the blind person needs
an internal representation or model of its environment. It will be sad for the blind user, if the used system
delayed while retrieving the stored object in the database. Sessions are important to describe scenes and
matching or retrieving the stored scene much faster to convert it to the blind person as a voice scene. This
also gives intelligent behavior to the system and it will increase the speed of the system to remain in the
real time in other side the blind person will be happy of the fast query answer of the system. The most
important point, that we use the proposed system in real time application for blind person.
Related Works:
The indoor navigation is heavily researched and the investigation continuous in this field, most of the
research concentrate on the features of the scenes, the features are extracted in many ways, the common
ways are Harries and SIFT extraction of image features, we shall concentrate in our study on SIFT where
many navigations used SIFT feature descriptors as in, many papers proposed approaches to how to deal
with the features extracted to be useful in the next stage of the system. We established this by clustering
descriptors; descriptors in a large cluster are less distinctive than those in a small cluster. It is important to
our system recognize the scene correctly, to do this we took what we see is distinctive and discriminative,
so we don’t incorrectly recognize scenes that are common in our environment. The Bag of Words
algorithm has been applied to SIFT descriptors to identify discriminative combinations of descriptors. The
basic method used for object recognition is to encode some of the properties of the object as a descriptor
with it poses that can be stored in a database, this is used for localization process in the environment for
the blind person. Much work to remove ‘visually ambiguous’ scenes was needed and more complex
profiles were preferred to provide more discriminative features. The blind person needs to recognize the
object invariantly with occlusion, different views (or scenes), scales, orientations and perspective
transformations. The epipolar constraint does define an invariant feature, but this is defined by seven
feature matches (up to a small number of possibilities) so eight or more points are needed for this to
validate or invalidate a possible correspondence. In demonstrates real-time loop detection using a hand-
