Abstract: Aiming at the problems of the random spatial distribution of real raindrop streaks and the loss of background details during the extraction of raindrop streak features, an algorithm for reading the indications of pointer instruments in rainy environments based on SPCANet (Spatial and Channel Attention Network) is proposed. SPCANet introduces a channel attention mechanism on the basis of the SPANet (Spatial Attentive Network) network to construct an enhanced Bottleneck module. It adopts a multi-scale feature extraction and fusion strategy, combines a hybrid attention mechanism, and optimizes the spatial attention module. A composite loss function is designed to constrain the model training from multiple dimensions. These improvement measures work synergistically, aiming to significantly enhance the overall performance of the image deraining model in complex scenarios. Experimental results show that the proposed method can effectively remove raindrop streaks and better restore the detailed information of the image background.