The aim of this study is to explore neural network-based image recognition techniques and their applicability and performance in different application scenarios. With the rapid development of artificial intelligence technology, image recognition technology is more and more widely used in various fields, especially the introduction of convolutional neural network (CNN), which makes image recognition technology show excellent performance in processing complex data. Traditional machine learning algorithms often show low recognition accuracy and generalisation ability when dealing with complex image data due to their shallow structure. Deep learning, on the other hand, has made significant progress in image recognition tasks by constructing a multi-layered neural network structure that can better extract and characterise complex features in images. In the study of this paper, an improved convolutional neural network model is proposed, which consists of a convolutional layer, a filter bank, a residual learning module, and a classifier. In order to validate the effectiveness of the model, three datasets (D1, D2, and D3) with different sizes and features are used for experiments in this paper. The experimental results show that the proposed CNN model exhibits high recognition accuracy when dealing with different datasets, especially in the classification task of specific targets and realistic scenes, the model is able to extract and recognise key features effectively, showing good convergence and stability. The experimental results not only demonstrate the excellent performance of the deep learning-based target recognition algorithm on different datasets, but also reflect the stability and gradual optimisation ability of the model during the training process. Although a slight overfitting phenomenon occurs on the validation set, these problems can be effectively mitigated and improved by means of appropriate regularisation and hyperparameter tuning. The research in this paper not only lays the foundation for the further development of deep learning in the field of image processing, but also provides technical support and innovative impetus for intelligent applications in various industries. The successful application of deep learning in the field of image processing not only provides strong support for the development of computer vision technology, but also brings new opportunities and challenges for the intelligent development of various industries.