The Study Of Positioning Technique At Small Area In A Large Scale Based On Image Recognition

The positioning technique at small area in a large scale has made significant development in the field of aerospace, biomedical, marine, meteorology, home appliances, such as GPS positioning technology, Wireless Sensor Networks (WSN), Radar positioning technology, Grating orientation technology, Nanometer positioning technology, and so on. However, study of the positioning technique at small area in a large scale used in the mechanical field has almost still remains in the initial positioning mechanism design. Its self scale and positioning algorithm are the insuperable barrier of the positioning technique at small area in a large scale. Therefore, there are great difficulties to be used in machinery.Image recognition technology is the product of the development of the computer. Its rapid development of in few decades has more and more widely used in military, agriculture, industry and other fields, such as sailing aviation monitoring stored grain pest identification, crop positioning, machine vision, and so on. The application of image recognition technology in industry most is machine vision in industrial robots, welding, etc. There are still lots of space in the positioning research.Based on the analysis above, this paper presents a method based on the positioning of the image geometric features, and the pixel center of the characteristics of the region will be regarded as a reference point. According to the principle of the relative position between the reference point and target point will not change, we can locate the target point.Firstly, the article describes the application of the positioning technique at small area in a large stale, and image processing technology in the positioning technology in the status of application and research. Secondly, proposed the positioning technique of small area in a large scale based on image recognition, demonstrated its feasibility in theoretical basis. and deduced the positioning mathematical formula according to the geometric relationship between points and lines. Then the reference point and target point to determine the mathematical calculations, and image processing technology needed in this process were discussed. In order to simplify image recognition process, this article carried out a detailed analysis of the straight-line reconstruction by using of the Hough transform. This article had made some corresponding improvement to the traditional image rotation matching technology, so as to make the matching speed largely improved. And the image rotation matching experiments of different shape had done to obtain its positional deviation. Finally, the article discussed the determination of the target location, including its specific positioning steps, and did the instance of the verification, and draw the conclusions of the whole system positioning system error of not more than3units of pixels.