微型电脑应用

在复杂巡检光照环境下,输电线路无人机巡检目标定位误差较大,多无人机的定位时间较长。因此,提出基于改进自注意力机制的输电线路多无人机固定机巢巡检目标定位算法,以解决巡检效率低下、定位误差较高等问题。在PT(pyramid transfO rm er)网络中引入自注意力机制模块,生成通道注意力权重矩阵。采用多头自注意力机制处理特征图,进一步提高PT网络对输电线路特征图像素间内在关系的捕捉能力;计算输电线路特征图的每个空间位置之间的相似度分数,生成一个空间注意力权重矩阵,实现输电线路的巡检目标定位。实验结果表明,所提算法能够准确定位输电线路所有传感器的位置,且在不同时间段的自然光照条件下,目标定位误差低于0.6m。无人机在升阔地带的线路故障定位时间控制在8s以内,缓解了巡检效率低下和定位误差较高的问题。

In complex inspection lighting environments,the positioning error of unmanned aerial vehicles inspection targets for transmission lines is relatively large,and the positioning time of multiple unmanned aerial vehicles is long.Thus,a multiple fixed drone nest inspection target positioning algorithm for transmission lines based on an improved self-attention mechanism is proposed to solve the problems of low inspection efficiency and high positioning errors.A self-attention mechanism module in the pyramid transformer(PT) network is introduced to generate a channel attention weight matrix.A multi-head self-attention mechanism is adopted to process feature maps,which improves the PT network' s ability to capture the intrinsic relationships between transmission line feature map elements.The similarity score between each spatial position in the transmission line feature map is calculated to generate a spatial attention weight matrix,and the target positioning of transmission lines inspection is achieved.The experimental results show that the proposed algorithm can accurately locate the positions of all sensors on the transmission lines,and the target positioning error is below 0.6 m under natural lighting conditions at different time periods.The fault positioning time of unmanned aerial vehicles in open areas is within 8 s,which alleviates the problems of low inspection efficiency and high positioning errors.