Multi-sensor information fusion is a new type of information fusion technology developed in recent years. In machine vision, edge information is a very important type of information, and edge extraction is the foundation and prerequisite for subsequent processing. This method also has certain applications in medical diagnosis, human simulation, and control of certain complex industrial processes. The idea of multi-sensor data fusion may seem simple and straightforward, but it is a very difficult task in practical applications. Image registration is a branch of image pattern recognition and multi image fusion, which has great application value in remote sensing, military and other aspects. On this basis, a new multi-sensor information fusion method is proposed. Firstly, the basic principles of information fusion technology were elaborated and corresponding processing modes were provided. On this basis, this article explores the current development status and existing problems of information fusion from three aspects: the architecture of information fusion, the application scope of information fusion, and the algorithms of information fusion. On this basis, the development direction of information fusion technology is proposed.