Target positioning and tracking device and image identification method

Abstract

The invention relates to a target positioning and tracking device and an image identification method and relates to the image identification technology field. The target positioning and tracking device comprises a paster device, an image identification terminal (3) having a camera and a computer (4), wherein the paster device comprises an inner circle paster (1) and an outer circle paster (2), the image identification terminal (3) is an intelligent handset, an inner side of the outer circle paster (1) and an outer side of the inner circle paster (2) are connected, the paster device is pasted on a surface of the target, the image identification terminal (3) is used for shooting and identifying the target having the paster device, and the computer (4) is in communication connection with the image identification terminal (3). The target positioning and tracking device utilizes the paster device with an excellent structure, a detection algorithm including threshold limit and contour detection is employed, time complexity of the algorithm is only n<2>, the speed is about 100 times as fast as a same type recognition pattern algorithm, and operation timeliness and accuracy on the handset or an ARM board are guaranteed through the algorithm.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a device for locating and tracking a target and an image recognition method. Background technique [0002] Currently commonly used image detection algorithms include pattern recognition, matching pursuit and threshold limit. Among them, the most important of the pattern recognition algorithm and the matching pursuit algorithm is the classification algorithm. Generally, there are K-NN classification method, Bayesian classification method, PCA classification method, etc. The disadvantage is that the calculation amount is huge, and a 32*32 image is not optimized. The time operation dimension is also 1024, the amount of operation reaches one million, and the time complexity is n to the 4th power, where n is the image side length. Regardless of whether it is optimized by PCA dimension reduction or kd-tree optimization, it will Secondly, for objects that are constantly chang...

AI Technical Summary

Problems solved by technology

Among them, the most important of the pattern recognition algorithm and the matching pursuit algorithm is the classification algorithm. Generally, there are K-NN classification method, Bayesian classification method, PCA classification method, etc. The disadvantage is that the calculation amount is huge, and a 32*32 image is not optimized. The time operation dimension is also 1024, the amount of operation reaches

Images