Recognition of Locust Based on Improved ResNet

Abstract: A dozen species of locusts (Orthoptera: Acrididae) are a major threat to food security worldwide. Their outbreaks occur on every continent except Antarctica, threatening the livelihood of 10% of the world’s population. The locusts are infamous for their voracity, polyphagy, and capacity for long-distance migrations. For effective control, the insects need to be detected on the ground before they start to develop air borne swarms. Detection systems need to determine pest density and location with high speed and accuracy. Location of the swarms on the ground then enables their control by the application of pesticides and bio-pesticides. This work proposes a locust species recognition method based on Resnet50 -convolutional neural network (CNN). We experimentally compared the proposed method with other the state-of-the-art methods on the established dataset. Experimental results showed that accuracy of this method reached higher than the state-of-the-art methods. This method has a good detection effect on the fly species recognition.