Locusts are the most dangerous pests that cause the devastation of crops and plants around the world. The consequences of locust attacks are disastrous for both food security and livelihoods of the rural populations. Currently, effective identification of these insects requires large numbers of trained experts and may cost millions of dollars. On top of that, the large amount of chemical insecticides used can have serious side effects on the environment. This means precise automated locust identification systems could minimize total amount of chemical usage to combat the locusts. rnTo identify the desired locust this research mainly focusses on analyzing and extracting unique features of the locust from the acquired training dataset and then train the proposed system accordingly. However, to include all locust types the collected 412 total training image datasets have distinctive ranges of sizes, shapes, backgrounds and views. Having such distinctions makes them complex for identification. Thus, to solve this kind of complexity and to increase identification accuracy, we use deep neural network model such as ResNet50. rnTherefore, the design of the proposed system model includes various stages starting from acquiring locust images up to identifying locust. Finally, the performance of the proposed system is assessed by its ability to identify locusts from a given test dataset. According to the experiment results, this research work has shown that the performance of the proposed system is 95.2% when it comes to classification of locusts from non-locusts using ResNet50.