Ground object change detection method, ground object change detection system and terminal

Abstract

The invention is suitable for the technical field of image processing, and provides a ground object change detection method, a ground change detection system and a terminal, and the method comprises the steps: obtaining a first image before a ground object changes and a second image after the ground object changes in a target region; inputting the first image and the second image into a feature differential convolutional neural network model to obtain an output first change intensity graph; and based on the first change intensity map, generating a binary image of ground object change. The method, the system and the terminal improve the robustness and stability of remote sensing image change detection, and improve the precision of change detection. 本申请适用于图像处理技术领域，提供一种地物变化的检测方法、检测系统及终端，其中方法包括：获取目标区域中地物发生变化前的第一影像及地物发生变化后的第二影像；将所述第一影像及所述第二影像输入至特征差分卷积神经网络模型中，得到输出的第一变化强度图；基于所述第一变化强度图，生成地物变化的二值图像，提升遥感影像变化检测的鲁棒性和稳定性，提升变化检测的精度。