Research On Consideration Of Intelligent Control And Image Recognition Based On Machine Vision

Based on the theory of machine vision, the movement control of intelligent car, and the real-time characteristics identification and spatial coordinates determination of obstacles images deeply investigated by theoretically and technically with the computer, embedded control system and image recognition technologies in this thesis. The main work and achievements are(1) The development of machine vision, the hardware and software of embedded-controlling systems are summarized, respectively. The important scientific significance and practical value of thesis are introduced too.(2) One mathematical model of intelligent car is proposed based on the kinematic theory. Then, using ATmage16 unit, the optimized motion controlling hardware platform is developed and tesed.(3) After the real-time operating systemμC/OS-Ⅱis completed transplantated to ATmage16 control platform, all of hardwares such as PWM, SPI, and UART are drived by our coded programe, the applications programe of the motor control, signal sample and communications are coded too, finally, the syetem's characteristics of software are tested.(4) Calibration of the camera is realized by 2D plane target method, the low-pass and Sobel convolution Core are adopted to detect the feature point of images by the low-pass and convolution, respectively.(5) Using the known epipolar feature to matching feature points, the spatial coordinates are determined based on the principle of parallax.The thesis belongs to the basic and frontier researchs for inter-discipline intelligent control and image identify, the results have impotment practical value.