·Table of Contents
·Methods and Instrumentation
On-Line Measurement for the Depth of the Oil Pump Jet Pressure Vessel
Zhao Jiyin, Gai Shuang, Zhao Jingrong, Huang Ting
School of Information Science & Engineering, Jilin University of Technology,Changchun, China 130025
In this paper, we introduce a new method based on video image processing for measuring the pressure vessel depth of the oil pump jet by indirectly measuring the two-dimensional size of the punch needle. Using the CCD camera with high resolution ratio, the system can compelete the collecting of two-dimensional size datas which are needed to be detected and the storing and processing of the two-valued image datas is performed with two-port of RAM technique. The research for two-valued image processing and image subdividing make it possible that the horizontal resolution ratio of the system is 4mm, and the vertical resolution ratio of it is 2.5mm.
Key Words: image processing, two-dimensional size, on-line measurement system
The oil pump jet is one of the important parts of the automobile engine, and the pressure vessel depth is an important parameter of the oil pump jet, which influence the oil injection quality of the engine and the performance of the automobile directly. The high accuracy on-line measurement for the pressure vessel depth is an important part in manufacturing the oil pump jet. In this paper, we introduce a new method for measuring the pressure vessel depth of the oil pump jet by indirectly measuring the dimension of the punch needle.
The pressure vessel depth of the oil pump is punched by the punch, which is key technology, the pressure vessel depth depends on the dimension of the punch needle. The shape and the dimension of the punch needle are shown in Fig. 1.
Fig 1: shape of punch needle|
The pressure vessel depth of the oil pump is the length from F2 to the tip of the punch needle. So, if the positions, the F2 and the punch needle tip are measured, the accurate positions of F2 and the punch needle tip in the scanning line are known, and then the length of the punch L is gotten. So we introduce a new indirect method for measuring the dimension of the punch needle. The system is a type of photo-electronic-mechanic device, which includes area matrix CCD image sensor, optical system, microprocessor unit, and numerical control machine tools. The two-dimension information of the punch needle dimension is acquired by CCD image sensor in disconnect mode, after processed the two dimension information using video image technique, The system can then obtain the desired diameter value of the key position of the punch in level direct, and the calculate the length of the punch needle finally, the on-line measurement for the pressure vessel depth of the oil pump jet is relived indirectly. The measuring system is shown in Fig. 2.
Fig 2: block diagram of measuring needle|
II Measuring Principle
When the punch needle is under the exposure to the parallel light beam, an image signal of it will be formed on the CCD image sensor. The light sensitivity of the image element array is different. The part of image which is exposed to light corresponds to the low level of the video signal. While the other part which is shaded by the punch needle corresponds to the high level of the video signal. Thus the detecting pulses, which correspond to the width of the part shaded by the detected object on a base line, are formed. If the scanning speed is VL, the pulse width of image signal ( i. e. the time that the scan line scanning the same line of the image need to take) is TL, then the image width (or diameter) of the object on the same line d is that the pulse width times the scan speed, i.e. d= VL´TL. For the CCD image sensor, its scan speed is constant and known. Thus, if the pulse width of the image signal can be worked out. Then the image width on the line of the detected object can be calculated. The ratio of the real width D to the image width of the detected object d is the amplification of the imaging system. If the amplification of the system is K', then
where KLT = K'VL is called the equivalent factor on the direction of the diameter of the punch needle.
The equivalent pulse width KL can be obtained by measuring and scaling for the standard test piece. Then we can measure the positions of the needle's diameter F2 and the end of the punch pin. The pressure vessel depth of the oil pump jet is the length from F2 to the end of punch pin. For this system, it corresponds to the accurate position on the scan line of F2 and the end of the punch needle. If the scanning time of scan line from F2 to the end of the punch pin is TH, the vertical scanning speed is VH, then the length of the punch needle (the pressure vessel depth of the oil pump jet) is:
The vertical scanning speed VH is constant for some CCD image sensor and is known before. So the length of the punch needle is in proportion to the vertical scanning speed TH. Processing the video signal by the sigle-chip computer, we can get the length of the punch needle.
III Two-Valued Processing of the Video Signal
Here the so called data of size refer to the pulse width value corresponding to the dark image on a TV line scanning line. In order to measure the pulse width accurately and quickly, we add the high frequency clock signal into the pulse width of the pulse signal and count this signal with binary by the hardware circuit to obtain the dimensional data of the image. The cancellation signal of the horizontal blanking signal can be formed by delaying the horizontal sync signal and adjusting the duty ratio. Meanwhile the cancellation signal and the video image signal after two-valued processing (including the horizontal blanking signal and the image signal of the detected object) are put into the "AND" circuit.
Then the horizontal sync signal and the horizontal blanking signal are removed by the time sequence matching. So the output only contains the two-valued image signal of the horizontal projecting image of the detected object. The block diagram of it is shown in Fig. 3. The waveforms of every point are shown in Fig. 4.
Fig 3: Block diagram for two-valued processing of image signal|
Fig 4: Waveforms for every point of two-valued processing circuit of the image signal|
In order to improve the accuracy of counting, the frequency of the 20MHz clock signal need to be as high as possible and must be stable. Due to the limitation of electrical parameter of the counter, 20MHZ integrated oscillator is used in the system.
Then the formula (1) becomes:
where NL is the pulse counting number of the 20MHZ oscillator in the pulse width of the dimension of punch needle.
By scaling processing of the F2 standard test piece, KL=4mm/pulse, the diameter dimension of the punch needle is in proportion to the counting number of pulse of the 20MHZ oscillator within the pulse width. That is
IV Microprocessor Unit
It consists of MCS-51 computer of slice and through circuit, the computer of slice is key part. This unit can perform the data of the image dimension, saving, using, digital filtration, display of measuring value and output of controlling signal and so on. It's primary feature is that, it can complete the high-speed data automatic acquisition by two-port RAM technique, and then writes the data to outer RAM directly. After written, the CPU will used the total line of microprocessor unit to draw the dimension data of the image, process data and perform the output of the processed results.
The functions of this part are as follows:
- performing dimension data of the object measured.
- writing the dimension data in the corresponding position of outer RAM under the action of the address signal together with the writing signal.
Data acquisition use picture saving method, it consists of four parts, first part is controlling circuit of gathering time sequence, the second part is sequence generator of data, address and writing logical signal, the third part is the interface of the micro-computer, the forth part is out RAM. The structure block diagram is shown in Fig.5.
Fig 5: structure block diagram for CCD video signal of picture saving method|
After the data are gathered, the computer on slice will perform the transfer of the bus, data acquisition and data processing, Its hardware principle block diagram is shown in Fig. 6.
Fig 6: block diagram of data processing|
The outer bus is transferred to inner bus by the output control signal of computer on slice, thus, the computer on slice can process the data in outer RAM through the bus. The high-speed acquisition and saving for measuring signals will be completed by two-ports RAM.
V The Image Subdividing Method for Improving Vertical Resolution Ratio
Because of the system is used to measure the length from the diameterF2 to the tip of the needle, we must find the accurate position of F2 first. But, generally speaking, F2 can't be just in the position of a scanning line, in most conditions, the accurate of F2is in the position between two scanning lines, meanwhile the position of F2is on the conical surface of the needle. So the accurate position of F2 can be found with linear interpolation method, the tip of the needle can't be also just in the position of a scanning line, its true position is in the position between the line of the last non-zero in the line data sequence and the next neighbor line. Because of the tip of the needle is spherical surface, the interpolation method can't be used to calculate the true position of the tip, we should use approximate processing. These would affect the measuring accuracy.
The solving method for these problems is that the interpolation accuracy in the position between F2 and the tip, and take the number of maximum interpolation line as the corresponding data, amplify the all lines from next neighbour line of F2 to the last non-zero datum (the end mark is that there are five zero at the end of a datum), the vertical measuring accuracy will be increased. The amplifying coefficient of the system is 6, and we carry out a reduction under the condition of unchanged accuracy, namely, we think the position of F2 is in 1/3, 1/2, 2/3 position between two scanning lines, thus, the position of F2 can be determined only carrying out one time of minus, the calculating time is shorter and the measuring accuracy is not be lowered. The processing for the tip of the needle is that determine if its value of the minimum minus 3 is , if it is greater than 3, one should be plus circularly. But its value can be greater than 5, the error will not be greater than 1/4 interval of TV line. Because of the length of punch needle of this system occupied more than one hundred TV lines.
So according to the formula (2), the pressure vessel depth of oil pump jet is
where NH is the line number of TV line scanning after six-point interpolation from the diameter F2 of the punch needle to the end of punch needle. Scaling to the 1.950mm standard test piece, KH=2.5mm/line number of line scanning. Then the dimension of the punch needle is
A on-line measurement system of two-dimensional size of punch needle is given in this paper. By two-valued image processing and image subdividing, the horizontal resolution ratio of the system is 4mm and the vertical resolution ratio of the system is 2.5mm. This system is fixed and operated in the work site of Beijing Tian Wei oil pump group company. It has better working stability and anti-interference ability, it also provides a new method for solving other questions of on-line measurement of two-dimensional size.
- Melen. R. D. and J. D. Meindl: One-phase CCD: A new approach to charge-coupled device clocking IEE J. Solid Stante Circuits, SC-7 92-93. (1973)
- Caldwell. L.V. et al.: Low light Level CCD-TV. Proc. 5th Int. Conf. on CCD. 45.(1979)
- Gradl, D.A: 250MHz CCD Driver, IEEE Jour. of Solid-State Circuits, SC-16. 100. (1981)