In recent years, image processing methods based on convolutional neural networks (CNNs) have achieved very good results. At the same time, many branch techniques have been proposed to improve accuracy. Aiming at the change detection task of remote sensing images, we propose a new network based on U-Net in this paper. The attention mechanism is cleverly applied in the change detection task, and the data-dependent upsampling (DUpsampling) method is used at the same time, so that the network shows improvement in accuracy, and the calculation amount is greatly reduced. The experimental results show that, in the two-phase images of Yinchuan City, the proposed network has a better antinoise ability and can avoid false detection to a certain extent.

In 2016, the IEEE task group ah (TGah) published a new standard IEEE 802.11ah, aimed at providing network connectivity among a large number of Internet of Things (IoT) devices. Restricted access window (RAW) is one of the fundamental MAC mechanisms of IEEE 802.11ah. It reduces the contention overhead in the dense wireless environment by dividing stations into different RAW groups. However, how to optimize the RAW parameters is still an open issue, especially in the run-time environment. In this paper, we propose a run-time RAW optimization scheme, namely RO-RAW, to improve the performance of RAW in the IEEE 802.11ah networks. RO-RAW adopts the Extended Kalman Filter method to estimate the channel status and adjusts the RAW parameters according to the number of competing stations in real-time. The evaluation via NS-3 simulations shows that, by tuning the RAW parameters appropriately, RO-RAW substantially improves throughput, latency, and packet loss performance compared with another RAW optimization scheme in different simulation scenarios. The results further show that, when the channel is relatively congested, RO-RAW improves the RAW performance more significantly.