| 1. Introduction The image measurement technique apply the digital image processing technique to accurate measurement realm. This technique obtains the geometry parameter of the object through detecting the edge of the object. The image measurement technology developed very quickly in the domestic and international. It has been used in geometry measurement extensively. Such as the aerial remote sensing measure, the tiny size of the accurate and complicated accessory measures and the external appearance detection etc. The automobile pipe has the levity curvature and complicated characteristics of shape. It very difficult detect by the manual method. With its high sensitivity, high resolution, high linearity, high velocity, large dynamic range, broad spectrum responsibility, non-contact, and low cost compared with other measurement instruments of the same grade of accuracy, image measurement system is very suitable to real-time precise measurement of geometric parameters of automobile pipe. 2. Research Content The main task of this dissertation is the design of an image measurement system for detecting size of automobile pipe. Image measurement system is a colligate technology that gathers the optics, optoelectronics, accurate machinery and computer technology. Digital image processing is its nucleus. The image measuring system is made up of image gathering system, image processing system, in addition, corresponding lighting system, interface hardware and PC. The mainly work and contents of this dissertation are assigned as follow: (1) Whole conceptual design of the image measuring system for detecting size of automobile pipe fittings This text design the mechanical structure and electric system of the measuring system. And analyzing measurement principle of whole system. Because of request high parallel degree and avoid dust of production spot, must design the optical system. This paper adopt tungsten halogen lamp makes the light source. Though constant current source supplying power to light to repress the influence of the nonlinear upon the CCD video signal. (2) Hardware design for image gather system. Hardware design for image gathering system made up of image gathering and buffering memory module, image process module, serial communication module and power module. Image gathering and buffering memory module is composing of CCD camera, video picture processor SAA7113H, frame memory and programmable controller FPGA. It finished the collection and memory of the information of the picture. The video signal that CCD image sensor outputs change to YUV422 form through SAA7113H chip. This form is 720*625 pels of every frame. FPGA produce address and enable signal intercepting 512*512 pels of that SAA7113H gather writing into the frame memory. DSP processor reads the digital picture from frame memory, and processing image by compressing, filtering etc, then communication to PC through the serial interface. The digital image deposit structure adopts two frames alternation. When SAA7113H writes one frame picture to frame memory 1, DSP reads another frame picture from frame memory 2. After frame memory 1 overwrite and frame memory 2 finish reading, FPGA switch read and write bus. SAA7113H continues writing the next frame to frame memory 2, and DSP reads the first frame from frame memory 1. The image process module is responsible for the realization of the pretreatment algorithm of the image, controlling the collection and memory of the image, and communication. The image process module use DSP as the core. This text select TMS320VC5409 chip of TI company. In this system, after the electricity on DSP, dispose SAA7113H and FPGA at first, then DSP controlled the collection and memory of the image. After this DSP process image and communicate to PC. The communication module is responsible for sending the pretreatment image data to PC. This text transmits the image data with RS-232 communication interface. Realize the communication with PC by the synchronous serial interface McBSP of DSP is connected with serial asynchronous transceiver MAX3111E directly. Design the power module. Because of the demand for the power of each chip is different. It have simulation circuit 5V,3V power and have digital part 3.3V, 2.5V, 1.8V power. So at the time of the design, the simulation electricity is separated form digital electricity. (3) System software design System software including FPGA control module, image gather program, serial communication module and image pretreatment, image binarization, , edge detectionof Canny operator, B spline profile fitting. FPGA control module realized image ping-pong memory. This text realize address produce and bus switching two functions by combining VHDL language and principle draw input. Serial communication module is composed of MCBSP serial initialization, MAX3111 work pattern and baud rate set up, interrupt service program. The purpose of the pretreatment of image is to reduce stain, noise and various kinds of interference, improve the quality of the picture. This text adopted the method that combining mean value filtering and many image average to improve image quality. In this paper image binarization adopts region-based recursive thresholding. The idea region-based recursive thresholding come from the hierarchical structure model. Carry on binarization of all part layer by layer according to the imagination of the hierarchical structure model. On every layer, carry on binarization segment again in the inside area according to the segment results of previous layer. Therefore the result received is used in the segmentation of the next layer. Like this, every idiographic area segment carry on under the condition that every area is relatively independent. So prevent the same layers of other grey level influence the segmentation result In this paper chose the canny operator to edge detect because it can not only detect the clear boundary but also provide the grads information. Parametric curve is selected to descript the shape of objects. In our test, cubic B-spline is used to approximate the object contour. And finally, we give a method to choose optimal description. (4) The experiment of the automobile pipe fittings measurement based on image measuring system and data analysis. The automobile pipe has the levity curvature and complicated characteristics of shape. In this paper did the experiment after choosing a representative part which is a switching bend pipe. In many times measurement random choose the data to analyze. The result of analyzed and calculation indicates that the precision of the system superior to 0.15%, satisfy the systematic precision demand. 3. Conclusion This paper design size measurement system of the automobile pipe fittings based on image measurement. It realizes the image collection and image transmit in hardware. And in software realize image binarization, edge detect, B-spline curvefitting, detect dimension of automobile pipe. Through the experiment, the system addressed the request of the system. It can detect dimension of auto pipe quickly and accurately.
|
PDF Full Text Request