The Method Based on Arm - Linux Machine Vision for Rapid Detection of Image Feature Points

For the traditional detection methods, there is a serious problem of noise, which leads to inaccurate results. The rapid detection method of ARM-Linux image feature points based on machine vision, the first shooting target image by using machine vision technology, through the ARM-Linux system to carry on the analysis of the feature points extraction, wavelet denoising and further processing of the image information; then according to the results of the image processing to establish the Facet model, calculation of feature points for image two order directional derivative, image acquisition the feature points of two order directional derivative minimum value is less than zero pixels as image feature extraction, based on the maximum value of the normalized image and local non maxima suppression to complete rapid detection of image feature points. The experimental results show that the proposed method can reduce the detection cost of image feature points, reduce the noise influence and has strong stability, wide applicability and high security.