Design And Implementation Of Embedded Image System Based On Texture Features

Texture represents the arrangement of the image grayscale element, whose local arrangement is regular while in the mass, irregular. Texture recognition has been widely used in numerous computer vision and pattern recognition applications such as medical imaging, satellite remote sensing, and industrial inspection. Texture recognition is a hot topic of image recognition research. However, it is difficult to analyze the texture due to the diversity of textures, the changes of rotation and scale.Embedded system is a special system setting for the application with the strict request for the function, the reliability, the cost, the volume, the power loss, which is designed for practical application by utilizing software and hardware. It is generally composed of the embedded micro controller, the periphery hardware devices, embedded operating system, and user’s application, which is used to control, supervise or manage other devices. It has been used in the field of control, consumer electronic product and etc, and one of the most important research aspects of the electronic field.This paper uses co-occurrence matrix parameters of wavelet transformed sub-band images as texture features and the extracted features are optimized using genetic algorithm,Multi-class support vector machines is arranged to complete the classification.The results of the experiments prove that the algorithm in this paper is able to guarantee the precision rate of texture recognition compared with traditional Bayesian method.This paper is composed of the following parts.Firstly, this paper summarizes the technique of ARM system and embedded image systems; Secondly,it presented an image recognition algorithm based on texture features; Thirdly, based on the algorithm, we design and implement an embedded image recognition system, including designs of peripheral circuit, Boot loader program, and features extraction and classifier software. At last,tecture test database are selected to evaluate the designed system.Experiment results demonstrate the superiority of the proposed approach compared with some existing methods and the expected design goal.The designed system has been used in practical products.