2P1-B05 画像処理を用いた自動害獣捕獲システムに関する研究 : 顔画像による獣種判別と全身画像による成・幼獣判別

抄録

In recent years, there has been widespread damage caused by wild animals. Medium-sized harmful animals, such as masked palm civets and raccoons, cause crop and house damages. Large-sized harmful animals, such as wild boars and deer, cause crop and forest damages. The most effective methods to catch harmful animals are a medium cage trap for medium-sized harmful animals and a wild boar cage trap for large-sized harmful animals. This study improves automatic capture systems in order to prevent mistaken trapping pet cats by a cage trap, and to reliably capture adult wild boars. For a medium cage trap, developed system uses image processing to discriminate between harmful animals, masked palm civets and raccoons, and others by using features based on the Histogram of Oriented Gradients (HOG). For a wild boar cage trap, the automatic capture system uses image processing for the discrimination of adult and infant wild boars by whole-body images.