Moving object detection and classification using background subtraction algorithm and mixture of Gaussian algorithm

Abstract

Moving object detection and classification is an important aspect in the country border where the
security is essential. Many methods are available for object detection, tracking and identification.
However identifying the objects accurately in real time using image processing is very difficult task.
Many research work have been used the Background subtraction and histogram algorithm for moving
object detection and classification respectively. The histogram algorithm cannot classify the object
accurately because it is more difficult to compare data sets.
Hence the histogram algorithm is replaced with mixture of Gaussian algorithm for classification of
moving object. Mixtures of Gaussian algorithm can easily compare the data sets. Background
subtraction algorithm subtracts each frame with previous frame which is called as background image
for reference. Data set is compared using mixture of Gaussian algorithm which classifies the moving
object and it is giving the information with time stamp. Here Raspberry pi with camera is used for
capturing real time image and open cv is used for moving object detection and classification. This
approach is accurate and efficient especially in country boarder where the security is significant even
in absence of humans.