Object detection is an important and practical problem in the field of machine vision, which includes predicting the position of objects in the image and classifying them based on specified categories. In order to successfully locate and identify objects in the image, it is necessary to extract appropriate features that have the power to distinguish between different objects. In recent years, convolutional neural networks (CNN) have been proposed as an effective solution for object recognition due to their ability to automatically learn visual features. This article tries to address the importance and challenges ahead in this field, and after introducing famous datasets, networks and prominent models of object detection get explained. Then, by introducing the common evaluation criteria in this area, we evaluate the mentioned models and present the latest solutions and innovations in this field.