Target detection method using context awareness

The invention provides a target detection method utilizing context awareness, which comprises the following steps: acquiring image data to be detected, and setting a training set and a test set; constructing a convolutional neural network based on context awareness; training the convolutional neural network to obtain a training model; and inputting the test set image into the training model to obtain the position and category information of the target in the image. According to the method, position positioning and category detection are performed on the target in the image based on a deep learning method, object information is provided for target detection to be applied to the fields of agriculture, military, security and protection and the like, and development and further research of multiple applications such as information monitoring and the like are promoted. 本发明提供了一种利用上下文感知的目标检测方法，包括获取待检测的图像数据，并设置训练集和测试集；构造基于上下文感知的卷积神经网络；训练卷积神经网络，得到训练模型；将测试集图像输入训练模型中，得出图像中目标的位置和类别信息。本发明基于深度学习的方法对图像中的目标进行位置定位和类别检测，为目标检测应