java - Detection of real object using camera. Possibilities of finding real objects using camera -


i developing project in robotics using android camera. robot can move , clean surface if objects found in path. stuck in detection of objects in path, not using sensors find distance between camera , objects. using camera find objects in path. don't want find out shape, type, color of object in path. small objects in path can ignored such sand , grass etc. there should not big objects in path (in marked portion of figure below)

enter image description here

while detecting objects using camera paintings, shadows, detected objects. above figure music symbol detected object. there paid or open source library find out real object detection using camera (it more helpful if usable in java or c++). have tried opencv, javacv, jhlabs,jjil, aviary sdk, jiu, imagemagic, etc. no can find exact solution need. if working image processing please suggest me can in real object detection using android camera. if not possible please give me explanation why not possible.

the possibilities have tried:

  1. face recognition algorithm (feature detection) :- face detection algorithm based on matching pattern. not applicable in case. ie, here object may or in shape. there no pattern matching.

  2. object recognition : object recognition doing same matching patten. don't wanted recognize object, wanted find out object in marked portion in image , real object (ie, not floor paintings, shadows , etc..).

  3. edge detection : have tried different edge detection algorithms not accurate solution me. ie, paintings , shadows having edges can't assume edges found real objects.

if working image processing please suggest me can in real object detection using android camera. if not possible please give me explanation why not possible.

real time, on fly object detection area of intense research.

right now, our existing algorithms capable pattern matching, , looking pre defined images within camera frame.

the problem real world detection real world too... real. have lots of variations of kind of object or item, , impossible single library provide detection them. real world detection plausible through extensive data generation , training via machin learning, requires more processing , battery power mobile devices can provide.


Comments

Popular posts from this blog

.htaccess - First slash is removed after domain when entering a webpage in the browser -

Automatically create pages in phpfox -

c# - Farseer ContactListener is not working -