A MemoryEfricient Algorithm forRealTimeObject Counting
Suresh Subramanian · 2009
Object Counting isa challenging problem with different solutions based on theavailable computing powerandthenature ofdatatobeprocessed. Memory efficiency, simplicity andspeed areveryimportant requisites foralgorithms tobeusedinmodernday systems whichinclude distributed systems andwireless sensor networks whereitisextremely advantageous to dobasic preprocessing ofdata fromthesensors inthe nodes itself Reduced computing poweravailable inthe nodes posesachallenge andthis canbeovercome by Figure 1.Smallbinary imageanditsfirst, second theuseofalgorithms withsimple steps. Wepresent an andthird columns algorithm tomonitor continuity, countobjects and before counting [2]. In[3], thebackground image has measureparameters suchasareabyisolating patternsbeenusedasareference forcounting thenumberof andobjects fromdatabased onsimple computations. movingobjects. Eachof thesemethodsis Memoryefficiency, speed andflexibility ofthe computationally demanding andismoreapplicable for proposed algorithm havebeendiscussed. Ithasmany thespecific applications they havebeendesigned for. broadapplications that include counting objects ona Counting identical objects hasbeendoneusing High conveyer, monitoring people inatraffic signal, pattern L isolation frommulticolored images andmonitoring LevelAutoCorrelation (HLAC) wherethesimilarity continuity inrealtimeimagedatafromsensors. We between objects tobe