The Study Of Articles’ Fast And Efficient Image Recognition Algorithms Based On Template Matching Technique

Template matching algorithm has broad application prospects in the field of image recognition with its simple logic, wide application, and good noise immunity. However, there are some defections such as large amount of computation, the accuracy and speed of template matching couldn’t reach the actual application requirements in dealing with image recognition in complex background or multi-target image recognition. Therefore, this paper mainly study and raise the application of kinds of template matching algorithms and other image processing algorithms in AGV visual navigation system and cigarette recognition system, in the purpose to improve the instantaneity and robustness of template matching algorithm.In AGV visual navigation system, this article uses the deflection angle as characteristic quantity in path template matching to reduce the complexity of control system by partition fitting, which enhances the system’s reliability. It also designs a set of simple and reliable AGV workstation flags to complete the basic operation, and puts forwards flags identify methods based on the Hu invariant moment feature of outer contour to meet the needs of AGV’s real-time operation.A new cigarette recognition method has been proposed in this paper by using the uniqueness of cigarette bar codes. Firstly, it extracts the one-dimensional bar codes from cigarette pictures with adaptive morphological filters. Then, it used projection method to locate readable characters and split figures. Finally, after recognizing Barcode figures based on feature weight fuzzy template matching, the paper completes the cigarette recognition. Experiments show the method has higher efficiency, better accuracy and stronger robustness, so it is able to meet the demand for practical use in the society.