JPS6222070A - Speed measuring instrument - Google Patents

Abstract

PURPOSE:To measure the moving speed of a moving body with high precision including a zero and a fine speed by selecting and adding simultaneously picture element information at intervals of plural pieces among pieces of picture element information obtained from photoelectric converting elements arrayed in one dimension, and shifting pieces of picture element information which are selected in every selection. CONSTITUTION:The moving body 5 is scanned by the line sensor of an image pickup part 1 and image information is inputted to an arithmetic processing part 4. The arithmetic processing part adds pieces of image information of photoelectric converting elements 1, 3, and 5 to obtain sum data D1. Then, the pieces of image information of the photoelectric converting elements 1, 3, and 5 are summed up to obtain sum data D2. Similarly, pieces of image information are shifted, one by one, to add three pieces of information. Consequently, addition sequence data equivalent to the movement of the line sensor at a speed V0 is obtained. The addition sequence data has a center frequency fc=V/2p (V: composite speed of V0 and moving speed of moving body, p: pitch of line sensor). The moving speed of the moving body is determined with high precision by deciding the fc.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Application of the Invention] The present invention relates to a non-contact speed measurement method using a spatial filter.

[Background of the invention]

For example, when a moving object consisting of a point light source is viewed through a number of slits lined up along the moving direction of the moving object, the slits through which the moving object is visible shift sequentially as the moving object moves. Assuming that the moving object is equally visible from all slits, the time Δt from when the moving object is visible through one slit to when it is visible through the next slit is given by the velocity of the moving object, v1, the pitch of the slit, and p. Then, it can be expressed as follows.

ΔT= - ■ Therefore, when the light from the moving body that passes through each slit is detected by a common sensor, a signal with a frequency fe expressed by is obtained from this sensor, and by measuring this frequency fc, the formula From (1), the speed V of the moving object can be known.

By the way, when light from a flat moving object is detected by a sensor through a moving slit with a pitch p, the frequency of the signal obtained thereby is also expressed by the above equation (1).

Generally, even if the surface of an object is flat, it is not a perfectly flat surface, but has random irregularities. However, this random unevenness can be seen as a combination of various unevenness with different spatial periods. Therefore, we fixed 1
When light from the surface of the moving object is received through two slits, the amount of received light changes depending on the spatial period of these irregularities. In other words, if the spatial period of the unevenness having the above-mentioned spatial periodicity on the surface of the moving object is l, and the moving speed of one slit is V, then a signal with a frequency of v/12 can be obtained from the sensor due to this unevenness. It turns out.

``The surface of a moving object has various periodic irregularities with different β, and the slit with the above pitch p allows only the light from among these irregularities, where l is equal to the pitch p, to pass through, formula(
It generates the frequency f shown in 1). Such a slit and sensor arrangement therefore has a filtering effect and is therefore called a spatial filter. This spatial filter has a spatial frequency that is inversely proportional to the pitch p, and on the other hand, a signal with a frequency [C] shown in equation (1) is obtained from the surface of a moving body moving at a speed of ■. .

Therefore, by using this spatial filter, the speed of the moving object can be measured.

Such a spatial filter can be placed on a moving body and the speed of this moving body can be measured by detecting the floor surface. A speedometer using this spatial filter is effective for measuring the speed of a moving object such as a self-propelled robot that needs to recognize its own position with high accuracy. Currently, speedometers for such moving objects generally measure the number of rotations of the wheels, but errors may occur if the wheels slip or the wheels wear and their diameter changes. Put it away. The speedometer using a spatial filter is a non-contact type that does not use any member that comes into contact with the floor surface, so the above problem does not occur.

By the way, a conventional speedometer using a spatial filter, as disclosed in Japanese Patent Publication No. 5B-44215, generally uses the lattice-shaped slits and sensors described above. Also, for example, Special Publication No. 55-4270
As disclosed in Publication No. 6, a speedometer is also known in which a large number of photoelectric conversion elements are arranged at regular intervals and the detection outputs of these elements are added to obtain a signal with a frequency corresponding to the speed of a moving object. ing.

On the other hand, a moving object does not always move at a constant speed, but at least when starting and stopping, it moves at a speed different from that in a steady state of movement, and along with this, the signal obtained by the spatial filter ( 1) The frequency (
(hereinafter referred to as the center frequency) rc changes. The instantaneous velocity of this moving object can be found from equation (1) by measuring this center frequency fc, but if the pitch of the slits and photoelectric conversion elements is constant and the spatial frequency of the spatial filter is constant, the starting speed of the moving object is When the moving speed is low, such as when stopping or stopping, the center frequency ``.'' also becomes low.

By the way, the accuracy of frequency measurement decreases as the frequency becomes lower, and a large error occurs when measuring frequencies close to zero. Therefore, it is not possible to accurately measure the center frequency fc when a moving object is started or stopped. For this reason, in self-propelled robots that require high self-position measurement accuracy, errors are unavoidable when starting and stopping.

In order to solve this problem, in equation (1), the center frequency f
Since c is inversely proportional to the pitch p of the slits (or photoelectric conversion elements), it is conceivable to reduce the moving speed V and also reduce the pitch p. However, in order to do this, in the prior art described above, the moving speed V is divided into several ranges, and a spatial filter with a different pit slit (or photoelectric conversion element) is installed for each range. In addition, a means for converting these is required, which leads to a complicated configuration, an increase in size, and an increase in cost.

As another method to solve the above problem,
As disclosed in No. 104, a method is known in which the grating itself corresponding to the slits is also mechanically moved. In other words, if the moving speed of the grating is V', even when the moving object on which the spatial filter is placed is stationary, the movement of the grating allows the sensor to detect the center frequency fc' according to the moving speed V'. A signal is obtained, and when the moving object is moving relative to the floor surface, the sensor detects the moving speed V of the moving object relative to the floor surface and the moving speed V of the sensor itself.
A center frequency can be obtained depending on the speed of synthesis with ′. Therefore, by calculating this composite speed from this center frequency fc and further subtracting the moving speed V', the speed V of the moving object can be obtained.
``According to this method, even if the speed V of the moving body is zero or very close to zero, the obtained center frequency fc will increase by the amount corresponding to the moving speed V' of the sensor. The speed of a moving object can be measured with high accuracy.

However, in this method, the performance of the drive device and control device for moving the sensor affects the measurement accuracy t.
When measuring the speed of a moving object where the moving speed of the sensor is not negligible, errors due to these performances will appear, and since the drive device includes mechanical elements, the reliability of the speed measuring device will be impaired, and the configuration It is also complicated and large, which poses problems from an economical perspective.

[Purpose of the invention]

SUMMARY OF THE INVENTION An object of the present invention is to provide a speed measuring device that eliminates the drawbacks of the prior art and is capable of measuring the moving speed of a moving body, including zero speed and slow speed, with high accuracy.

[Summary of the invention]

In order to achieve this object, the present invention simultaneously selects and adds every plural pixel information from pixel information obtained from one-dimensionally arranged photoelectric conversion elements, and adds the pixel information selected for each selection. By sequentially shifting the pixel information,
The feature is that the row of photoelectric conversion elements is moved in a pseudo manner.

[Embodiments of the invention]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of the speed measuring device according to the present invention, in which 1 is the ↑1 shaded area and 2 is the increased
□Width circuit, 3 is A/D converter, 4 is arithmetic processing unit, 5
It consists of a computer system, a drive circuit, etc., and an arithmetic processing section 4.
Operates by control signals from. The line sensor 1a is
As shown in FIG. 2, n photoelectric conversion elements (where n is plural) are arranged one-dimensionally at a pitch p, and an image of the moving object 5 is formed on this by an optical system (not shown). imaged. As a result, pixel information corresponding to the amount of light received is generated in each photoelectric conversion element. In FIG. 2, each photoelectric conversion element has 1.2.3. ... is numbered n.

In the main unit 1, scanning is repeatedly performed to read pixel information from each photoelectric conversion element in the order of 1, 2, 3, . If the time for this one scan is t6, one photoelectric conversion element is loaded every talnM interval. These pixel information are supplied with a conversion start pulse in synchronization with the amplified information by the amplifier circuit 2, thereby digitizing the sequentially supplied pixel information. These digitized pixel information are supplied to the arithmetic processing section 4.

Furthermore, Fig. 3 shows digitized pixel information, and each pixel information is given a code with a dash added to the code given to the photoelectric conversion element that generated it (Fig. 2). . In addition, digitized pixel information will also be simply referred to as pixel information hereinafter.

The arithmetic processing unit 4 is a RAM (random access memory)
, ROM (Read Only Memory), CPU (Central Processing Unit), etc.

The ROM stores a program for performing arithmetic processing and controlling the entire speed measuring device.The CPU reads this program from the ROM and performs calculations and various controls. The RAM stores input data and temporarily stores variables used in calculations.

Pixel information 1'+2', 3' supplied to the calculation unit 4
, . . . are sequentially located at predetermined addresses d, d2, . . . as shown in FIG. It is written to d,,... At the same time, a predetermined number of addresses are selected from this RAM at predetermined intervals, pixel information is read out, and these pixel information are added to generate addition data.

Then, these selected addresses are sequentially shifted, and pixel information read out each time is added to form addition data. Now, - as an example, every other three in Fig.
First, specify addresses d+, ds, and "d" to successively generate pixel information 1', 3', and 5', and add these to form addition data. Then, specify the address d2 + d4.d and set the pixel information 2 as Dl.
', 4', and 6' are read and added to form addition data. This is defined as addition data D2. Similarly, pixel information is read out by designating the k-th addresses a, ak*z, and dk-4, and these are added to form addition data Dk. The number of selections must satisfy +4=n at the maximum, so selection is performed (n-4) times and (n-'4) pieces of addition data are obtained.

FIG. 5 shows such addition data arranged chronologically in the order of formation, and the time required to form one addition data is equal to "t/n."

Looking at this operation for the line sensor 1a shown in FIG. 2, obtaining the addition data D1 means 1.3.
The process is to read and add the pixel information generated by the photoelectric conversion element No. 5, and to obtain the addition data D2,
, 4.6 is to read and add the pixel information generated by the photoelectric conversion elements. In the same way, the sequentially added data is equivalent to the data obtained by sequentially reading and adding every other three pixel information with a one-by-one shift, and this means that as shown in FIG. A line sensor 1a consisting of three white photoelectric conversion elements arranged at a pitch of 2p is arranged at a distance p.
This is equivalent to reading pixel information from these photoelectric conversion elements and forming addition data every time the photoelectric conversion element moves. In this case, the formation period of the addition data is t, ;/n as in Kou, and the line sensor 1a moves by this open circuit #p, so the run sensor 1a has a one-dimensional velocity V0 = The three photoelectric conversion elements arranged at equal intervals constitute a spatial filter in combination with adding the pixel information of these photoelectric conversion elements at regular intervals. Furthermore, due to the addition data forming operation of the arithmetic processing section 4, the stopped line sensor 1a shown in FIG. 2 is replaced by the line sensor 1 shown in FIG.
This is equivalent to a moving at a speed V0. In this way, the photographing section 1 in FIG. 1 can be moved in a virtual manner.

The addition data string shown in FIG. 5 includes a sine wave signal with a center frequency fc expressed by the following equation from Ko's equation (1): ■ fc=□ ......(2) p It's a signal. Therefore, processing such as fast Fourier transform is performed based on this signal, and from the obtained frequency spectrum, equation (2
) can be found. However, the formula (
The speed ■ in 2) is the above-mentioned speed ■ of the line sensor a. Composite speed (vo + VT) with the moving speed ■1 of the moving object 5
, and the center frequency fc obtained by equation (2) also corresponds to this composite speed {circle over (2)}. Therefore, the arithmetic processing unit 4
Now, the velocity V is determined from the center frequency f using the following formula: V=fc ・2p (3).
4) is calculated, and as a result, the moving speed of the moving object V
You will get a.

From this, even if the moving speed (7) of the moving body 5 is zero or near zero, the pseudo moving speed (7) of the sensor can be added to this.
. Since the center frequency ft is generated according to the added speed ■, the center frequency rC is relatively high and can be detected with high accuracy (it can be detected, therefore, the speed ■ can also be obtained with high accuracy, so the moving speed ■7 of the moving body can also be It can be measured with high precision.

The above processing procedure will be explained using the flowchart of FIG.

Processing 1: First, as a preparation before receiving and receiving data, write it to RAM.
Initialize all (to zero, for example). Then, the count variable i representing the number of times the addition data is formed is passed through the small circuit 2 to A.
/D converter 3.

The pixel information is stored at addresses d, d, as shown in FIG.

...Sequentially import.

Processing 4: The pixel information captured in Processing 3 is stored at address d,
+di*2+di+4 and sum them to form addition data #i.

Process 5: Increment the count variable i by 1.

Process 6: Determine whether i is larger than n-4 (n is the number of photoelectric conversion elements of line sensor a). If it is, proceed to process 7; otherwise, return to process 4.

Processing 7: Added data DI obtained in Processing 4 to Processing 6,
D2... Find the center frequency fc from D7-4.

Process 8: Calculate 2XpXfc (p is the pitch of the photoelectric conversion element), and set the obtained value to V.

Then, the relative speed ■T with respect to the moving body that is the true object of measurement is calculated by performing the calculation V Vo (Vo = n p/t, ).

Process 7: Output the obtained ■1.

In addition, in FIG. 6, the moving speed (■) of the photoelectric conversion element. By setting V to be greater than the possible moving speed of the moving body 5, ■, the speed V obtained from the clothing equation (3) is as follows: When the moving direction of the photoelectric conversion element and the moving body 5 are the same, vku■. In the opposite case, V>V. Therefore, equation (4) made in process 8 of FIG.
As a result of the calculation, the moving speed (7) has a positive or negative sign, and the moving direction of the moving body 5 can be determined from this. In addition, in Do's equation (3), V−V,
If =0, it means that the moving body 5 is stopped at this time.

However, if the arithmetic processing unit 4 in FIG. 1 performs arithmetic processing so that the line sensor 1a shown in FIG. When the moving speed v0 of the line sensor 1a approaches the virtual moving speed v0, the center frequency fc to be detected becomes low, and the detection accuracy of this moving speed 1 decreases.

On the other hand, the processing operation of the arithmetic processing unit 4 causes the photoelectric conversion elements 1, 3, 5, photoelectric conversion elements 2, . 4. 6. When three photoelectric conversion elements 1 are selected at a time in the order shown in FIG.
, 2.3 is equivalent to the line sensor 1a. On the other hand, in FIG. 2, photoelectric conversion elements n, n-2, n
-4, photoelectric conversion elements n-1, n-3, n 5.・
Select three pieces in order of ... and add data D+', D2'
, ..., in this case,
This line sensor 1a moves at the same speed (■) in the opposite direction to the arrow in FIG. This is equivalent to the line sensor 1a that moves at .

Therefore, through the processing of the arithmetic processing unit 4, the center frequency obtained when the line sensor 1a shown in FIG. 2 is equivalent to the line sensor 1a moving in the direction of the arrow R in FIG. If the center frequency obtained when equivalent to the line sensor 1a moving in the direction of arrow F opposite to is fR, then the speed obtained from these center frequencies f, rF is, from 2 (3), In the former case ■ □=2pi,l In the latter case, V,=2P-fr. Therefore, the moving speed v7 of the moving body 5 is VT
= IVRv, l = lv, v. It becomes 1. In this case, it goes without saying that (V, -V.) and (■, -VO) have opposite signs).

In this way, no matter which direction the line sensor 1a is equivalently moved by the processing of the arithmetic processing unit 4, the moving body 5
can measure the moving speed of.

Therefore, as another embodiment of the speed measuring device according to the present invention, the moving speed (7) of the moving body is measured by selecting the larger of the detected center frequencies fR and f.

For this purpose, the center frequency fR,f is constantly monitored, and fR
When >f, the center frequency fR is used, and when r,>r, the center frequency f is used.

The processing procedure of this embodiment will be explained using the flowchart of FIG.

Processing 1: First, prepare the data before receiving it.
[Initialize all RAM (to zero, for example). Then, the pixel information sent from the A/D converter 3 via the small circuit 2 is sequentially taken in as addresses d, d2°, . . . as shown in FIG.

Processing 4: The pixel information captured in Processing 3 is stored at address d,
, d, , , d, , and the sum thereof is taken as addition data D.

Processing 5: Increase count change i by 1.

Process 6: Determine whether i is greater than n-4 (n is the number of photoelectric conversion elements arranged in the line sensor). If it is, proceed to process 7; otherwise, return to process 4.

Process 7: Added data D+ obtained in Processes 4 to 6
, Di, . . . from Dl-4, its center frequency f
Find 8.

Process 8: Set count variable i to zero.

Processing 9: The pixel information captured in Processing 3 is transferred to address d□
, + d n-2-i + d n-4-t, and their sum is taken as addition data D,'.

Process 11: It is determined whether or not i is greater than n-4. If they are greater or equal, the process proceeds to process 12; otherwise, the process returns to process 9.

Process 12: Addition data D obtained in processes 9 to 11
1', Dz' . . . D n -4', the center frequency f is determined.

Process 13: Process 7. Center frequency f obtained in process 12
R and f are compared, and if r is greater than or equal to fR, proceed to process 14; otherwise, proceed to process 15.

Process 14: Perform the calculation 2XpXfy, set this as the speed V, and assign zero to the variable 5IGN (flag indicating direction).

Process 15: Perform the calculation 2Xl)Xf*, set this as the speed ■, and assign 1 to the variable 5IGN.

Processing I6: Processing! 4, or the speed V obtained in process 15
Therefore, the relative velocity v7 with respect to the moving object 5 which is the true measurement target is calculated by the calculation V-Vo.

Process 17: Output Vt and 5IGN.

In this way, the center frequency can be detected with high accuracy and the speed measurement accuracy can be kept high. Here, the range in which frequency analysis is possible is defined as the lower limit value fL and the upper limit value rH, and ■, =
When considering a speed value of 2xpxfL,V,=2XpXfH, when the speed is measured using the processing method shown in FIG. 9, the pseudo moving speed of the line sensor is . (=np/1g
) is set to a value close to (2), the speed measurement range can be made larger than that obtained by the processing method shown in FIG. 6 when two directions are covered.

Note that in each of the above embodiments, R of the arithmetic processing unit 4
The pixel information read from every other three addresses among the pixel information stored in A M was added to form the addition data, but it is not possible to read out the pixel data from four or more addresses and create the addition data. It may also be formed.

Further, the difference between the sum of pixel information read from a plurality of addresses every odd number of three or more and the sum of pixel information read from the same number of addresses located in the middle of these addresses may be used as addition data. Now, assuming that pixel information is to be read from three addresses every third, in FIG.
−(d = ” d = −a°+d,.s ) (
d=-z+dt. 6+d. ,. ), and in process 6, (n-10) is used instead of (n-4).

This constitutes a so-called differential spatial filter that performs subtraction processing on two sinusoidal wave signals of center frequencies fc that are in an antiphase relationship with each other, and high-frequency noise that is mixed is removed by this subtraction processing.

In this way, without changing the hardware configuration,
The configuration of the spatial filter can be changed in a pseudo manner simply by changing the software processing.

The present invention can be used to measure the speed of various moving bodies, and as an example, FIG. 10 shows a case in which it is used to measure the speed of a moving vehicle. In the same figure, a speed measuring device 8 according to the invention is attached to the bottom of a mobile vehicle 6 at . 7 is a wheel. When the moving vehicle 6 moves, the floor surface 5 is moving in the opposite direction relative to the speed measuring device 8. By measuring the moving speed and direction of the floor surface 5, the moving vehicle 6 can be moved in the opposite direction. You can know the speed and direction of movement.

Since it is not a mechanical structure, it is possible to measure the moving speed including zero with high accuracy without requiring any mechanical structure, and it is possible to achieve a simpler configuration, smaller size, and improved reliability, which is better than the conventional technology described above. This problem can be solved and a speed measuring device with excellent functions can be provided at low cost.

[Brief explanation of the drawing]

FIG. 3 is an explanatory diagram showing the chronological order of pixel information obtained during one scanning period of the line sensor shown in FIG. 2, and FIG. A schematic diagram showing the storage order of pixel information in the RAM in the arithmetic processing unit in FIG. 1, and FIG. 5 is the RA shown in FIG.
An explanatory diagram showing the chronological order of addition data formed by pixel information read out from M, FIG. 6 is an explanatory diagram showing a moving spatial filter equivalent to the spatial filter of the above embodiment, and FIG. FIG. 8 is a schematic explanatory diagram of another embodiment of the speed measuring device according to the present invention; FIG. 9 is a flowchart showing the processing procedure; 10th... A/D converter, 4.
... Arithmetic processing unit, 5... Mobile object. Demon 1 Figure 2 Figure n Child 3 Figure Child 4 Figure Child 6 Figure Ka 7 Figure Akira 8 Figure Child 10

Claims (3)

[Claims] (1) An image of a moving object is formed on a line sensor consisting of a large number of photoelectric conversion elements arranged one-dimensionally, and the pixel information generated on the photoelectric conversion elements is sequentially output by scanning the line sensor. an analog-to-digital converter that sequentially digitizes the pixel information; a memory means that stores the digitized pixel information obtained during one scanning period of the line sensor; and a plurality of each of the memory means. an arithmetic processing unit that reads and adds the pixel information to form summation data for each of a plurality of pieces of pixel information and processes a series of summation data to obtain the moving speed of the moving body; The plurality of pieces of pixel information for forming the addition data is pixel information obtained from the plurality of photoelectric conversion elements at predetermined intervals in the line sensor, and for each addition data, the pixel information obtained from the plurality of photoelectric conversion elements at predetermined intervals is The combinations of the plurality of photoelectric conversion elements are sequentially different in the arrangement direction of the photoelectric conversion elements in the line sensor, and the imaging unit is moved equivalently at a predetermined speed in the direction in which the combinations are different. A speed measuring device characterized in that it is configured to be able to. (2) In claim (1), the arithmetic processing unit obtains a center frequency of a sine wave signal formed by the series of addition data, obtains a velocity from the center frequency, and obtains an equivalent value of the imaging unit. A speed measuring device characterized in that the moving speed of the moving body is obtained by subtracting the moving speed. (3) In claim (1), the arithmetic processing section sets the equivalent moving directions of the imaging section to first and second directions that are opposite to each other, and the respective equivalent moving directions obtain the center frequencies of the sine wave signals formed by the series of addition data for each of the data, obtain the speed from the larger center frequency, and subtract the equivalent moving speed of the imaging unit to obtain the moving speed of the moving body. A speed measuring device characterized by: