ELEPHANT AND LEOPARD IMAGE DETECTION SYSTEM

Abstract

Elephant and Leopard survive in a forest area may cause a major problem to Human habitat resulting in death of livestock, crop damage, human fatality and injuries caused by elephants and leopards. And some times elephants being killed by human. For this reason, an automatic unsupervised Elephant and Leopard Image Detection System (ELIDS) is proposed as a solution to human conflict with elephants and leopards and bringing context to elephant and leopard conservation. The elephants and leopards image captured within the forest border areas are extracted using wavelet to get multilevel wavelet coefficients, in which image features and similarity of the elephant and leopard query images was compared with data base images by using image vision algorithms. And the warning message is shipped to the forest range officials intimating that an elephant or leopard has been detected and is approaching human habitat while an elephant or leopard touches the forest border areas. In order to retrieve the image from the database on accurate time and to speed up the image retrieval time, the system make use of the very familiar Euclidean method and Manhattan distance methods. On making use of these two methods the proposed method process the image faster. So this system retrieves more number of images in very less amount of time period and makes the projected optimized distance method more efficient and reliable