| With the rapid development of science and technology, non-contact dimensional measurement based on machine vision technology research has been very widely used in industrial production, with the character of non-contact, real-time, high efficiency, and high precision. Due to manual measurement technology cannot meet the requirements of mechanical parts dimensional measurement with high accuracy and speed, therefore, this paper proposes a non-contact dimensional measurement system based on machine vision, so as to achieve accurate and rapid dimensional measurement.Firstly, to study the various hardware performance parameters, this paper chose Shenzhen Weike CCD camera, COMPUTAR lens, Linghua production RTV-24 series of multi-path online video capture card, LED light source acts as lighting and other hardware, and build a hardware system.Secondly, the image processing section includes image filtering, binarization, edge detection. Image filtering is compared to a variety of filtering methods, according to the wavelet energy entropy principle obtain the noise variance, and then set the threshold T λ ? log M at different scales. Image edge detection is compared to a variety of edge detection methods, and set the threshold T T a B, in the C x, y image to remove false edge points. When the modulus maxima of the image C x, y above the threshold T, it is true edge points. Instead, the modulus maxima below the thresholdT, it is false edge points.Thirdly, it also describes the process of measuring straight line and circle based on Hough transform, and propose the improvement algorithms with regard to the random Hough transform. This paper set up the measurement system using the GUIDE editor GUI interface provided by MATLAB, it enables a variety of image processing functions.Finally, the measurement system has been repeated experiments, analyzed the experimental data and study the source of error in the measurement system, and proposes a method to reduce the error. The experiment results show that the non-contact dimensional measurement system based on machine vision can meet the ideal effect. |
