JPH08286132A - Optical scanner - Google Patents

Abstract

PURPOSE: To obtain such a clock signal as picture elements are uniformly aligned on a scanned plane with a light beam deflected with a constant angular velocity or a specified angular velocity characteristic by dividing a reference clock and using it as a picture element clock. CONSTITUTION: A time interval between adjacent picture elements is constituted of a first time interval of an integer multiple of the reference clock pulse separation and a second time interval of another integer multiple of the reference clock pulse separation different from the first one. Divided frequency ratio value of each picture element interval to the reference clock pulse interval is written beforehand in a picture element correction data memory 130, and is loaded into a frequency divider 131 at everytime one picture element is transmitted, and a picture element clock is generated by a line buffer 111 every time the reference clock generated by a reference clock pulse oscillator is divided. Although interval errors between individual adjacent picture elements are large, the errors come lower than a certain value in the case that an interval is evaluated at every several picture elements or more by properly locating the divided ratio deciding each picture element interval in the scanning direction.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an optical scanning device used in a laser beam printer or the like, and more particularly to a pixel clock generating method.

[0002]

2. Description of the Related Art In image recording devices such as laser beam printers and optical scanning devices used in various image reading and measuring devices, rotary polygon mirrors have been widely used as deflectors for deflecting and scanning light.

FIG. 7 shows a typical structure of an optical scanning device using a rotary polygon mirror. In FIG. 7, the light beam emitted from the light source 1 is shaped into parallel rays by the collimator lens 2 and is deflected by the rotating polygon mirror 3. The deflected light beam is imaged as a predetermined spot on the surface to be scanned 5 by the scanning lens 4.

The rotating polygonal mirror 3 is a regular polygonal column around which a plane mirror surface parallel to the rotation axis is arranged, and rotates at a constant angular velocity. Therefore, the light beam incident on the rotary polygon mirror 3 is deflected at a constant angular velocity. When the light beam thus deflected reaches the surface 5 to be scanned as it is, the moving speed of the spot is higher in the peripheral portion than in the central portion where the distance to the rotary polygon mirror 3 is short on the surface 5 to be scanned. It gets faster.

As a first technique for solving this problem, the scanning lens 4 is provided with negative distortion so that the image of the spot contracts around the surface 5 to be scanned, that is, the deflected beam is curved toward the center. Many methods have been used to give characteristics. That is, when the focal length of the scanning lens is f and the angle formed by the light beam deflected from the optical axis is θ, the distance h from the optical axis to the spot on the surface to be scanned is f.
・ It can be represented by θ. A scanning lens having such characteristics is generally called an "fθ lens", and many structures and designs have been proposed.

On the other hand, as a second technique for correcting the linear velocity of the spot on the surface 5 to be scanned, the linear velocity is kept as it is, and the time interval of the pulse corresponding to the pixel is decreased toward the peripheral portion, that is, the pixel There is a method of obtaining pixels at equal intervals by increasing the frequency of the clock toward the periphery.

Therefore, there is a method of providing a variable frequency oscillator with a signal modulated according to the scanning period to obtain a predetermined clock. Further, in the technique disclosed in Japanese Patent Laid-Open No. 62-32769, a signal obtained by dividing a high-frequency original vibration by a selected division ratio N and a pixel clock are also selected by a selected division ratio M. The pixel clock is controlled by phase comparison with the signal divided by. By sequentially switching between N and M during scanning, the pixel interval on the surface to be scanned is made substantially constant. That is, the frequency of the pixel clock is N / M times the reference clock.

In both the first and second techniques described above, the pixel clock serves as a writing clock when recording an image, and the sampling clock serves as a sampling clock when reading an image.

[0009]

However, according to the first conventional technique, the scanning lens must simultaneously satisfy the two optical characteristics of the field curvature (astigmatism) characteristic and the fθ (distortion aberration) characteristic. Therefore, the number of lenses is increased, the structure is complicated, and the cost is increased. Also, in order to avoid this, when trying to reduce the number of lenses, the shape of the surface becomes complicated,
There was a problem that it was difficult to obtain accuracy.

Further, according to the second technique, restrictions on the design and manufacture of the lens are reduced. However, the oscillation accuracy of the variable frequency oscillator is lower than that of a general crystal oscillator, and the quality of the image to be recorded is deteriorated due to the lateral shift (jitter) of pixels for each scanning line. In addition, JP-A-62-327
In the technique disclosed in Japanese Patent Publication No. 69, the circuit configuration is complicated and expensive.

Further, in the second technique, it is easy to make a monotonous change in the clock frequency, but there is a problem that it is difficult to make a fine change in the scanning width.

Further, relatively local fluctuations in the scanning speed caused by processing errors and assembling errors of the scanning optical system differ from scanning device to scanning device, and complete correction is difficult with the above-mentioned prior art.

In view of the above-mentioned problems of the prior art, the present invention provides a clock signal that causes a light beam deflected with a constant angular velocity or a predetermined angular velocity characteristic to have pixels at equal intervals on the surface to be scanned. Is obtained with a simple circuit configuration.

[0014]

In order to solve the above problems, an optical scanning device of the present invention comprises a light source for emitting a light beam,
A deflector for deflecting a light beam emitted from the light source,
A scanning optical system for forming an image of the light beam from the light source as a spot on the surface to be scanned, the linear velocity at which the spot scans the surface to be scanned is non-constant, and the surface to be scanned is In an optical scanning device that scans an image placed on a pixel unit of pixels arranged on a scanning line based on a reference clock having a constant frequency, the time interval between the pixel and an adjacent pixel is an integer multiple of the pulse interval of the reference clock. And a second time interval which is different from the first time interval and is an integer multiple of the pulse interval of the reference clock.

Further, in the optical scanning device of the present invention, in each of the plurality of scanning lines forming the image located on the surface to be scanned, adjacent pixel intervals at the same position in the scanning direction on the scanning line are The feature is that they are the same regardless of the scanning line.

Alternatively, in the optical scanning device of the present invention, in each of the plurality of scanning lines forming the image located on the surface to be scanned, adjacent pixel intervals at the same position in the scanning direction on the scanning line are Not the same for each scan line,
It is also characterized by having no periodicity.

[0017]

In the optical scanning device of the present invention, the linear velocity of the scanning speed of the spot scanned on the surface to be scanned is not constant, and the reference clock from the single reference clock oscillator of constant frequency is divided. Used as a pixel clock. Although there is a large error in the spacing between pixels in each adjacent pixel, it is fixed when the spacing between pixels is evaluated with several pixels or more by properly arranging the frequency division ratio that determines the spacing between each pixel in the scanning direction. The error is less than or equal to.

Further, by making the arrangement of the frequency division ratio different randomly for each scanning line, it is possible to prevent the nonuniform pixel intervals from having periodicity.

[0019]

FIG. 1 is a block diagram of a first embodiment of an optical scanning device according to the present invention. The light source 1 is a semiconductor laser and is controlled by the laser driving circuit 101 to have a predetermined light amount. The light beam emitted from the light source 1 is shaped into a substantially parallel beam by the collimator lens 2. The shaped light beam is deflected by the rotating polygon mirror 3.
The deflected light beam is imaged on the surface 5 to be scanned by the scanning lens 4. At this time, the speed of the spot moving on the surface to be scanned 5 (scanning speed) is not a constant speed, and as shown in FIG. 1.10 times 465.66 mm /
It is sec. Before the scanning range, the light beam is incident on the horizontal sync detector 6 to generate a horizontal sync pulse.

This optical scanning device is, for example, 600 dpi (1
When used in a recording apparatus having a resolution of pixels per inch, dot per inch), the interval of one pixel is 42.
Since it is 33 μm, the time required for the spot to pass through one pixel is 100 nsec at the center of the scan. The image data to be recorded is stored in the image data memory 110, and the data of a single or a plurality of scanning lines is transferred to the line buffer 111 at a predetermined timing.
The reference clock generated by the reference clock oscillator 120 is counted by a predetermined value held in the write start position memory 132 starting from the horizontal sync pulse described above, and then the first pixel data of the scanning line is transferred to the laser drive circuit 101. The light source 1 is modulated. Thereafter, the pixel data forming the scanning line is sequentially transferred from the line buffer 111 to the laser drive circuit 101.

At this time, if the moving speed of the spot on the surface 5 to be scanned is constant, the time interval for sending out the pixel data from the line buffer 111 is such that the data is transferred in synchronization with the pixel clock of a constant frequency. Good. For example, assuming that the scanning speed in the central portion (423.33 mm / sec) is constant, a pixel clock of 10 MHz may be used. In practice, an integral multiple of the reference clock is divided and used to improve the synchronization accuracy of the writing position. Eg 8
If the frequency is divided, the reference clock is 80 MHz.

In this embodiment also, a reference clock having a constant frequency of 80 MHz is used. Since the scanning speed at both ends A increases by 10%, the pixel clock may be increased by 10% to 11 MHz to compensate for this. However, if the division ratio of the reference clock is changed from 8 to 7, the pixel clock becomes 11.43 MHz, which is too fast. Further, the scanning speed in the middle portion B is 444.5 mm.
/ Sec increases by 5% in the central part, and if the pixel clock is written at 10.5 MHz, the pixel pitch will be an appropriate value, but this is a value closer to 8 divisions than 7 divisions of the reference clock. is there.

If the division ratio is further increased to divide by 16,
It is obvious that the pixel spacing can be set more accurately,
The frequency of the original vibration becomes 160 MHz. Therefore, it is necessary to use a high-speed element for the frequency dividing circuit, which is not preferable in terms of cost and heat generation. Further, such a high frequency oscillator itself becomes expensive.

The difference of 1 in the number of frequency divisions in the case of dividing by 8 is a length of more than 5 μm in terms of the pixel interval, which is a human interval as an interval of 1 pixel or a size difference of 1 pixel. Invisible to the eye. Further, the size is equal to or slightly smaller than one size of the developer particles when the electrophotographic process is used for the recording apparatus.

Therefore, microscopically, there is almost no substantial difference in the recorded images regardless of whether the frequency division is 7 or 8, and the positional error of the pixel at a slightly longer distance (that is, macro) is caused. It becomes a problem. From such a point of view, in the present invention, macroscopic scanning position errors are almost eliminated by appropriately distributing pixel intervals having different frequency division numbers without changing the oscillation frequency of the reference clock at all.

For example, a pixel interval corresponding to a pixel clock of 10.5 MHz may be obtained on average near the position B in FIG. Near this position B, the pixel interval is 3 when the above reference clock is divided by 7 to form a pixel clock.
It is 8.89 μm, which is 44.45 μm when the frequency is divided by 8. First, focusing on 10 pixels near the point B, if the pixel interval of 8 division is dispersed by 6 pixels and the pixel interval of 7 division is dispersed by 4 pixels, the pixel interval at 10 pixels is 44.45 × 6.
+ 38.89 × 4 = 422.26 (μm), the original 1
0.2 for the interval of 423.33 (μm) for 0 pixel
Only 5% shorter. Next, for 100 pixels near the point B, when the pixel interval of 8 division is dispersed by 62 pixels and the pixel interval of 7 division is dispersed by 38 pixels, the pixel interval at 100 pixels is 44.45 × 62 + 38.89 × 38. = 4233.32
(Μm), the original interval of 100 pixels is 4233.33.
It is only 0.009% longer than (μm). This is the normal f as mentioned in the description of the prior art.
This is much smaller than the error (0.1 to 2%) in the scanning speed of the θ lens. In this way, the error becomes smaller as the distance between the pixels becomes longer.

When it is attempted to allocate each pixel near the position of B in FIG. 2 to the position represented by the frequency division by 7 or 8 of the reference clock, whichever is closer, the array is, for example, 8.
7,8,7,8,8,7,8,8,7,8,7,8,
The distribution is 8, 7, 8 ... Further, as shown in B of the figure, at the position where the scanning speed from the scanning end is + 6.67%, exactly the frequency division of 7 and the frequency division of 8 appear alternately.

The value of the frequency division ratio with respect to the reference clock at each pixel interval is written in the pixel correction data memory 130 in advance, and the frequency divider 131 is transferred every time one pixel is transferred.
The pixel clock is generated in the line buffer 111 every time the reference clock generated by the reference clock oscillator 120 is divided. Pixel data is sent to the laser drive circuit 101 in synchronization with this pixel clock.

As described above, according to the first embodiment of the present invention, the interval between arbitrary adjacent pixels has a large error with respect to the pixel interval determined by a predetermined resolution, but a plurality of pixels are separated. In terms of pixels, the spacing error is small, and in reality, the spacing error does not pose a problem for arbitrary pixels separated by 100 pixels.

As described above, even in the first embodiment of the present invention, when it is used in a normal image recording apparatus or the like, a sufficient effect is exhibited, but as a method for further mitigating a microscopic pixel interval error. A second embodiment of the present invention will be described below.

FIG. 3 is a block diagram of a second embodiment of the optical scanning device of the present invention. Since many parts are the same as those of the first embodiment, the different parts will be mainly described. . Also in this optical scanning device, the change of the scanning speed in the scanning range is as shown in FIG. 2 as in the first embodiment.

In the first embodiment, the value of the frequency division ratio that determines the interval between the pixels is stored in the pixel correction data memory 130, and when a certain pixel on the scanning line is focused, the value of the scanning line varies. The pixel spacing is constant.

For example, as described above, at the position where the scanning speed is + 6.67% with respect to the center, the division ratios 7 and 8 appear alternately. Naturally, the ratio at which the frequency division ratio appears is determined by the scanning speed, but unless the scanning speed changes abruptly, the frequency division ratios are arranged in such a regular manner from a local point of view. Depending on the relationship between the period in which this regularity occurs and the regularity of the image to be recorded or read by the optical scanning device, so-called moire fringes may occur and the image may be very unsightly.

Further, in the dither method or the area pattern method which expresses the image density as a group of a plurality of pixels in the image recording expressing the gradation (halftone), the size of the matrix expressing the density and the rule of the frequency division ratio. Density unevenness may occur depending on the relationship with the sex cycle.

In order to avoid this, in the second embodiment of the present invention, as shown in FIG. 3, the pixel position correction table 133 holds a plurality of division ratio arrangements corresponding to the respective pixel intervals, and performs scanning. One of them is randomly selected for each time and loaded into the pixel correction data memory 130.
The random numbers are uniform random numbers, and the probability that each placement data is used is equal.

The contents of the pixel position correction table 133 will be described in more detail. Now, the table holds data of 10 frequency division ratios, and the interval between adjacent pixels is not a constant frequency division ratio. For example, at the position of B in FIG. 2, the scanning speed is +5 as compared with the central part as described above.
%, The frequency division ratio is 7 (pixel spacing is 38.89 μm)
Is 4/10 and the frequency division ratio is 8 (pixel interval 44.45
If the data is arranged so that there are 6/10 cases of (μm), the pixel spacing at this position has a probability of 40% of 38.8.
With 9 μm, the remaining 60% has a probability of 44.45 μm, so the average is 42.26 μm, and the correct pixel spacing is 4
The pixel interval is reduced by 0.3% with respect to 2.33 μm.

In practice, if the division ratio distribution is determined only on the condition that the average value of the pixel intervals is as described above, an error may occur in the cumulative pixel position, so that the absolute pixel It is better to assign them probabilistically based on the coordinate position. FIG.
Shows a part of the array of pixel positions of the pixel position correction table according to the second embodiment of the present invention, in which pixels i to i + 3 are arranged in the scanning direction, and No.
0-No. 9 sets of 10 data are shown. Then, as in the above description, the scanning speed near the position of B in FIG. 2 is assumed.

If the position of the pixel i happens to be almost coincident with the position obtained by dividing the reference clock by an integer ratio, the next i +
The first pixel is located between the position where the reference clock is divided by 7 and the position where it is divided by 8, counting from the i-th pixel, 3.44 μm from the position divided by 7, and 2. 1 from the position divided by 8.
The position is 12 μm. Therefore, 10 in the correction table
Of the data in the set, the data may be determined so that 4 are at the 7-divided position and 6 are at the 8-divided position.

Similarly, the i + 2nd pixel is located between the position divided by 15 and the position divided by 16 when counted from the i-th pixel, and 1.33 μm, 16 from the position divided by 15
The position is 2.23 μm from the divided position. Therefore,
Of the data in the correction table, two may be at the 16-divided position and eight at the 15-divided position.

Further, similarly, nine i + 3th pixels may be arranged at the 23 frequency division position, and one may be arranged at the 22 frequency division position.

FIG. 4 shows the 10 sets of data created in this way. The average interval between pixel i and pixel i + 1 is 4
2.26 μm, the average interval between pixel i + 1 and pixel i + 2 is 4
2.26 μm, the average interval between pixel i + 2 and pixel i + 3 is 4
It becomes 2.78 μm, and a good pixel interval can be obtained.

One of the 10 sets of data thus created is randomly selected for each scan, and the frequency division ratio of each pixel is determined. When an actual image is scanned using this table, the (vertical) line in the sub-scanning direction becomes jagged by one pulse of the reference clock. But as I said, the jagged size is the pixel size,
Or, since it is sufficiently smaller than the line width, there is almost no problem. For example, in FIG. 4, when only the i + 1th pixel is used and the size of the pixel is a circle of 60 μm and a vertical line is drawn, it becomes as shown in FIG. 5, and it can be seen that there is no problem visually.

According to the second embodiment of the present invention described above, the error in the value obtained by averaging the intervals of the pixels lined up on the scanning line over the entire image can be made very small as compared with the first embodiment. . Further, since the frequency division ratio with respect to the reference clock is consequently randomly distributed to each pixel and has no periodicity, it is suitable for scanning an image having periodicity. In particular, if the number of pieces of correction data included in the pixel position correction table 133 is not increased to 10, it is possible to perform more precise correction.

Next, a third embodiment of the present invention will be described. Also in the optical scanning device in the third embodiment, the second
Similar to the first embodiment, most of the components are the same as those in the first embodiment, and the description of the overlapping portions will be omitted.

Similar to the second embodiment, the third embodiment prevents the frequency division ratio of the reference clock from being regularly repeated at the pixel intervals and has regularity as compared with the first embodiment. The object is to avoid the problem when scanning an image and to keep the average pixel interval constant without error.

FIG. 6 is a block diagram showing a third embodiment of the present invention. Here, the pixel position data memory 150 holds where each pixel on the scanning line is located with respect to a position that can be created by dividing the reference clock. Each time one pixel of data is transferred from the line buffer 111, the random number generator 140 generates a random number, and the frequency division ratio generator 150 calculates the frequency division ratio based on the pixel position information of each pixel. That is, in the second embodiment, the frequency division ratio for each pixel, which is determined by using the pixel position correction table created in advance, is generated in real time.

For example, in the (i + 1) th pixel in the second embodiment, the probability of generating a pixel clock by dividing the reference clock by 8 from the ith pixel is 62% (actually, a numerical value giving a binary probability). Is stored in the pixel position data memory 150.

According to this method, the table as in the second embodiment is not necessary, and only one set of pixel position information needs to be held in the pixel position data memory 150, and the amount of memory can be saved.

Further, in the third embodiment, no matter which frequency division ratio the pixel adjacent to the pixel of interest is in, it is completely independent.
The division ratio was decided randomly. On the other hand, a method in which a distance error generated at a certain pixel is dispersed to surrounding pixels and processed is also conceivable. When the i + 1-th pixel is located at the frequency divided by 8, the error is −2.12 μm. This error is corrected by surrounding pixels, i.e., i + 2nd on the same scan line,
A method may be considered in which the division ratio is determined in consideration of the distributed error when dividing the i-th, i + 1-th, and i + 2-th pixel on the next scanning line and determining the division ratio of those pixels.

In this method, the concept of dispersion of density error of the "error diffusion method" which is often used when binarizing a halftone image is applied to the pixel position, and the average pixel position error is reduced. can do.

However, a memory for holding the distributed error is required, and the periodicity may appear at the pixel position depending on the error distribution algorithm. is necessary.

The scanning speed characteristic of the optical scanning device of the present invention is shown in FIG.
As described above, not only the one in which the scanning speed monotonously increases from the scanning center to the end, but the one in which the scanning speed decreases conversely or the one in which the scanning speed fluctuates in a complicated manner, If pixel correction data or a correction table suitable for it is created, the correction can be easily performed.

In each of the above embodiments, the oscillation frequency of the reference clock is divided by 7 or 8 to obtain a pixel clock. However, depending on the setting of the fluctuation range of the scanning speed and the division ratio, for example, 7 minutes are used. There are cases where three division ratios are combined, such as division, division by 8, division by 9, and so on.

The division ratio for dividing the reference clock is not limited to 7 to 9 as described above, but in the combination of the division by 3 and the division by 4, the error in the adjacent pixel interval becomes too large, and the application May be limited. If the frequency division ratios are set to a combination of approximately 6 to 20, it is possible to meet all uses, and the effects of the present invention are sufficiently exhibited. Further, if the frequency division ratio is made larger than this, the interval precision becomes acceptable even if only the adjacent pixels are viewed without arranging frequency division ratios having different values in the scanning direction, so that the present invention is not necessary. However, needless to say, it is not desirable to unnecessarily increase the frequency division ratio, as described above, because it will increase the circuit cost.

The block configurations shown in the first to third embodiments of the present invention show an example for realizing the present invention.
Each component is not essential. Furthermore, each component may be realized by an electric circuit or software.

Further, the scanning speed of each manufactured optical scanning device is precisely measured individually, and the frequency division value corresponding thereto is measured by the pixel correction data memory 130 (or the pixel position correction table 1).
33 or the pixel position data memory 150), it is possible to accurately correct the disparity in the pixel interval due to the manufacturing error or the assembly error of the lens for each device.

Further, the optical deflector used in the optical scanning device of the present invention is not limited to the one having a constant deflection angular velocity of the deflected light beam such as a rotating polygon mirror, but the angular velocity fluctuates in a sinusoidal manner like a galvanometer mirror. It is clear that the same effect can be obtained by using the deflector.

Alternatively, in the optical scanning device of the present invention, not only is the deflector located on the front side of the scanning lens as shown in the above embodiment, but the deflector is located on the rear side of the scanning optical system. A so-called post-objective optical system can be easily used. That is, any type of deflector or scanning optical system can be used in the optical scanning device of the present invention, ignoring the uniform speed of the scanning speed, and optical scanning of very high imaging performance and low cost. The device can be obtained.

In the above-mentioned embodiment, the case where the optical scanning device according to the present invention is applied to the recording device has been described. However, when the pixel clock is replaced with the sampling clock and the flow of the image data is reversed, the reading device is read. Can be applied in the same way.

[0060]

As described above, in the optical scanning device of the present invention, it is possible to obtain a constant pixel interval macroscopically regardless of the characteristics of the scanning speed of the spot on the surface to be scanned. it can. In particular, the surface shape of the lenses forming the scanning optical system is simpler and the number of lenses is smaller than in the case of using an optical system in which the scanning speed is constant.

In addition, the pixel clock generation circuit for obtaining this pixel interval is composed of only a constant frequency oscillator and a simple frequency dividing circuit, so it is much simpler and more accurate than the conventional circuit configuration. At the same time, stability and reliability are greatly improved.

Further, according to the optical scanning device of the present invention, the distribution of the division ratio of the reference clock is finely set at each position in the scanning range, so that the scanning speed can be controlled depending on the accuracy of processing and assembling the optical system. Local non-uniformity can also be sufficiently corrected.

Further, by using the optical scanning device as shown in the second and third embodiments of the present invention, the intervals between adjacent pixels can be made substantially equal, and high precision or gradation expression can be realized. Suitable for applications that require

Furthermore, according to the third embodiment, it is possible to make the adjacent pixels seemingly equidistant as described above without having a correction table which is disadvantageous in terms of circuit cost.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an optical scanning device according to a first embodiment of the present invention.

FIG. 2 is a scanning speed characteristic diagram in a scanning region of the optical scanning device according to the first embodiment of the present invention.

FIG. 3 is a block diagram showing a configuration of an optical scanning device according to a second embodiment of the present invention.

FIG. 4 is an array diagram of pixel positions in a pixel position correction table of the optical scanning device according to the second embodiment of the present invention.

5 is a diagram showing a state of a line formed by the (i + 1) th pixel in the pixel position correction table shown in FIG.

FIG. 6 is a block diagram showing a configuration of an optical scanning device according to a third embodiment of the present invention.

FIG. 7 is a schematic view of a conventional optical scanning device.

[Explanation of symbols]

 1 light source 3 rotating polygon mirror 4 scanning lens 6 horizontal synchronization detector 101 laser drive circuit 110 image data memory 111 line buffer 120 reference clock oscillator 130 pixel correction data memory 131 frequency divider 133 pixel position correction table 140 random number generator 150 pixel position Data memory

Claims (3)

[Claims] 1. A light source for emitting a light beam, a deflector for deflecting the light beam emitted from the light source, and a scanning optical system for forming an image of the light beam from the light source as a spot on a surface to be scanned. The linear velocity at which the spot scans the surface to be scanned is non-constant, and the image placed on the surface to be scanned is based on a reference clock having a constant frequency in units of pixels arranged on the scanning line. In an optical scanning device that scans by scanning, a time interval between the pixel and an adjacent pixel is different from a first time interval that is an integral multiple of a pulse interval of the reference clock, and the reference clock of the reference clock is different from the first time interval. An optical scanning device comprising a second time interval which is an integral multiple of the pulse interval. 2. In each of a plurality of scanning lines forming the image located on the surface to be scanned, adjacent pixels at the same position in the scanning direction on the scanning line have the same time interval regardless of the scanning line. The optical scanning device according to claim 1, wherein 3. In each of a plurality of scanning lines forming the image located on the surface to be scanned, the time intervals of adjacent pixels located at the same position in the scanning direction on the scanning line are not the same for each scanning line. The optical scanning device according to claim 1, wherein the optical scanning device is absent and has no periodicity.