Design And Practice Of Smart Camera Based On OpenCV

With the continuous development of microelectronics and integrated circuit technology,machine vision field began to change from the traditional separation machine vision system which has complicated structure,high cost and big volume,to a smart camera system which is higher integration,higher reliability,less energy consumption,lower cost and more convenient on application development.At present,the smart camera is relatively expensive,and relies on imports.This article presents a design solution of low-cost smart camera system based on open source,and design the prototype of smart camera under laboratory conditions.This smart camera is based on embedded Linux open-source operating systems and OPENCV open-source computer vision library project.In hardware design,the Raspberry Pi based on BCM2835 was selected.The author designed the management and configuration software of smart camera.In this software,user can chose and configurate the image processing,input condition and output of IO,According to the characteristics of target image.The smart camera would do the functions like image acquisition,image processing,image analysis and statistics,IO output and so on.This software is developed by C++,using QT embedded GUI(graphic user interface)development kit and OPENCV(Open Source Computer Vision Library)which was transplanted.In addition,this article studies the foundational image processing algorithms which was used in smart camera,include image filter,edge detection,Hough transfom and so on.And based on traditional Canny edge detection operator,an improved adaptive Canny operator is proposed in this article.Simulation results shows that edges detected by the immproved adaptive Canny operator in this article has better continuity and single edge response than that on traditional Canny operator.Especially in the noise environment,it has better inhibitory effect of noise and better effect of edge detection.At last,under laboratory conditions,the experiments of coins detection and angle detection of servos rocker was used to introduce the functions and operation of configuration software of smart camera.In coins detection experiment,the number of coins was counted and the center coordinates was returned.Meanwhile,smart camera also can measure the radius of the coin relatively accurately,its average error was between 0.1 mm to 0.2 mm.In the experiment of angle detection of servos rocker,smart camera can measure angle of servos rocker with standard error of 0.27 degrees.Experiments shows that rhis smart camera has the advantages of high integration,flexible function,low cost and so on,it also has a certain measurement accuracy.And the feasibility of the design solution proposed by this article has been verified through experiments.The design solution of smart camera can be used in industrial automation line field,and instead of artificial,it can test workpiece like edge detection,geometry size and Angle detection,object identification,counting statistics and soo on.the IO interface of smart camera can communicate with lighting of machine vision or sorting system.Open source software interface helps smart camera to develop visual project which is more complex.