JPH0619258B2 - Light section image position detector - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an optical cutting image position detecting device, and more particularly, to detecting a groove position of a member to be welded by optical means, and based on the detection result, a welding route. The present invention relates to a preprocessing device for a light-section image signal of a light-section image position detecting device mounted on a machine such as a visual sensor-equipped welding robot that automatically corrects the welding.

[Prior Art] A conventional light section image position detecting device is disclosed in, for example, Japanese Patent Laid-Open No. 62-2.
As described in Japanese Patent Publication No. 67607, when a rising slope and a falling slope of an image signal are detected and an absolute value of the slope is larger than a preset threshold value of the optical cutting line slope, a signal rising signal and a signal rising, respectively. The down signal is output, and the peak value of the image signal within the period of the rising signal and the falling signal is captured as the image position of the optical cutting line.

[Problems to be Solved by the Invention]

In the above-mentioned conventional technique, sputtered light or the like having a width wider than the light cutting line is planned to be incident with a constant relationship with the light cutting line, but light of irregular irradiation shape such as arc light is generated. No consideration was given to the fact that the input pixels entered as ambient light and became partially bright.

For example, when the arc light enters, the brightness gradient may exceed the threshold value set for the light section image even in this portion.
And since the interval between the rising and falling positions of the brightness gradient becomes too large, the threshold value that defines the interval between the rising and falling is exceeded, so a light-section image is generated between the rising and falling of the arc light. If present, this image will be ignored and not detected.

In order to avoid the influence of this arc light, if the gradient threshold for detecting the rising and falling gradients is increased, the reflectance of the welded member decreases and the gradient of the light section image decreases proportionally. Had the drawback that the light section image could not be detected at this position.

The object of the present invention is to prevent the position of the light cut image from being affected by arc light or the like as ambient light in the light cut image, even if the reflectance of the member to be welded changes. It is an object of the present invention to provide a light-section image position detecting device capable of surely detecting the position.

[Means for Solving the Problems]

In order to achieve the above object, the present invention provides a light source for irradiating an object to be detected with slit light in an environment where ambient light is present or having a surface whose reflectance is not constant, and a light section image generated on the object to be detected. In the light-section image position detecting device, the image-processing device includes: an image-capturing device that captures the image signal; Smoothing differential processing means for smoothing an image signal from the means and obtaining a differential value of its brightness, compression processing means for storing the image positions where the differential value is maximum and minimum together with the code, and the stored image position The present invention proposes a light-section image position detection device including means for determining a set in which the distance between the maximum position and the minimum position is within a predetermined threshold as the position of the light-section image.

Specifically, the smoothing differential processing means includes a plurality of shift registers that capture image data via the A / D converter from the imaging means and transfer the image data one after another, and two sets of pixel data before and after the shift registers. , And first and second addition means for obtaining the average value of each set, and the sign of the output of the first addition means is inverted and added to the output of the second addition means to obtain the average of the front two pixels and the rear two pixels. And a third adding means for obtaining a differential value of the value.

Further, the compression processing means, specifically, a local value storage register that stores a local maximum value or a minimum value of the output of the smoothing differentiation processing means, and an output of the smoothing differentiation processing means as an output of the local value register. First comparing means for comparing, a threshold value register for storing the threshold value of the differential value, second comparing means for comparing the output of the threshold value register with the output of the smoothing differential processing means,
A station value address storage register that stores the station value and the address of this station value, an address counter that counts the clock signal, a differential value code output of the smoothing differential processing means, an output of the first comparing means, and a second comparing means. Based on the output, an update command for updating the local value storage register with the output of the smoothing differentiation processing means and a storage command for storing the output of the address counter and the sign of the differential value of the smoothing differentiation means in the local value address storage register. And a control logic circuit for outputting the light section image, which is a means for storing a local maximum position and a local minimum position of the gradient where the brightness gradient of the light section image is equal to or more than a threshold value together with the sign of the gradient.

A welding robot is obtained by fixing the light source and the image pickup means of any one of the above optical cutting image position detecting devices together with the welding torch to the wrist of the welding arm and controlling the welding torch to follow the target position.

Further, a sealing machine can also be realized by fixing the light source and the image pickup means of the light-section image position detection device to the sealing head and controlling the sealing head to follow the target position.

By fixing the light source and the image pickup means of the light section image position detecting device in the vicinity of the grinder disk and controlling the grinder disk to follow the target position, the grinder is not affected by sparks generated during the cutting work. Is obtained.

The present invention further provides a rising slope and a falling slope of the brightness of a light-section image obtained by irradiating an object to be detected with a slit light under a boundary in the presence of ambient light or having a surface whose reflectance is not constant. The invention proposes a light-section image data compression processing circuit comprising means for storing the maximum and minimum positions, and means for storing the sign of the gradient in correspondence with the maximum and minimum positions.

[Action]

In the present invention, the smoothing differential processing unit executes the differential processing while averaging the input image signal. As a result, the output of the smoothing differentiating circuit becomes a signal in the frequency band between the fine change and the gradual change of the input image signal.
The influence of noise is removed, and the rising and falling slopes of the light section image are emphasized.

The control logic circuit outputs an update command for updating the local value storage register with the output of the smoothing differentiation processing unit according to the output and sign of the smoothing differentiation processing unit, the output of the first comparator, and the output of the second comparator. Then, a storage command for storing the station value address storage address register and the code of the smoothing differentiation processing unit in the station value address storage register is output.

The second comparator compares the value of the threshold value register with the output of the smoothing differentiation processing unit, removes a local maximum value or minimum value due to a noise component that does not completely become zero in the smoothing differentiation processing, and after compression. To reduce the amount of information in. When the output of the smoothing differentiation processing unit is larger than the maximum value by the output of the second comparator, the maximum value of the local value storage register is constantly updated and compared, and the storage position is changed based on the comparison result. Since the storage command is issued to the station value address storage register, the maximum value and the minimum value can be updated and stored in the station value address register.

Further, since the sign of the output obtained by the smoothing differential processing unit is also stored in the local value address storage register at the same time, it is possible to distinguish whether the storage position is the local maximum or the minimum. Thereby, for example, even when the arc light and the light section image overlap and the slope rises again in the part of the light section image after the rising slope appears in the part of the arc light, from the result after the compression processing,
This state can be judged without error. .

Further, since the position is stored only by the output of the smoothing differential processing unit and the brightness is not limited, even if the brightness of the light section image changes due to the reflectance of the detection target, the light section image is The rising and falling positions of can be reliably detected.

〔Example〕

Next, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a schematic view showing an example of the configuration of a welding robot system having an optical cutting position detecting device according to the present invention. This system consists of a welding robot 1 and a welding power source 2.
A robot controller 3, an optical sensor head 4,
The image processing apparatus 5 which is the object of the present invention and the welding torch 6 are included.

FIG. 2 is a view showing an example of the structure of the sensor head 4 fixed together with the welding torch 6 under the wrist of the welding robot 1. The optical system of the sensor head 4 includes a semiconductor laser 7 and
An aspherical lens 8 and a cylindrical lens 9 that convert the light beam from the semiconductor laser 7 into slit light 10.
An interference filter 12 that defines a transmitted light band when capturing a light-section image by the slit light 10 reflected on the surface of an object to be welded (not shown), an objective lens 13 that forms an image of the captured reflected light, and a diaphragm 14. And the objective lens 1
The image pickup device 15 is configured to pick up the slit light 10 and the like formed by the image forming device 3. The semiconductor laser 7, the aspherical lens 8, and the cylindrical lens 9 may be assembled together and supplied as the slit light irradiation blocks 11 and 11A. The sensor head 4 observes a light section image obtained by irradiating the welding object with the slit light 10 and converts it into an electric signal.

In order to optimally control the observation direction so that the observation area of the sensor head 4 is always in front of the welding torch,
The entire structure is attached to the bottom of the wrist of the welding robot 1 via bearings 16 and 16A, and a structure that can rotate around the welding torch 6 is adopted. The observation direction is DC motor 17
It is controlled by the gear 18 and the timing belt 19.

The image processing device 5 processes the image formed by the sensor head 4 and detects the welding position in the image. Next, the welding position is converted into three-dimensional position information of the welding object by the principle of triangulation from the data of the irradiation angle of the slit light, the observation direction of the light section image, and the magnification data. The image processing device 5 transfers the position information of the welding object to the robot control device 3.

The robot controller 3 converts the position information into robot coordinates based on the position and orientation information of the sensor head 4,
The welding target position is stored as a welding target position, the welding path taught in advance is corrected, and the welding torch 6 is scanned and controlled.

FIG. 3 is a flow chart showing the procedure of image data processing in the image processing apparatus 5, and FIG. 4 is a diagram showing an example of data at each stage of FIG.

Image data processing includes preprocessing by the compression circuit of the present invention,
The post-processing uses the result of the pre-processing. The result of the preprocessing includes the local maximum or minimum position of the differential value together with the sign of the differential value indicating whether it is the maximum value or the minimum value. From the result of the pre-treatment, the post-treatment first becomes maximum,
Next, a pair of minimum position data is searched, and further, the position data in which the distance between the maximum and the minimum is less than or equal to the threshold value is stored again as the rising position and the falling position of the brightness based on the light section image. . The data in this state is the 4th
Figure (a).

Next, the continuity of the rising and falling positions of the brightness in the horizontal direction of the screen in the vertical direction of the screen is determined, and the pairs of rising and falling elements that satisfy the continuity are determined to be in the same line area one after another. Process.
The result is shown in FIG. 4 (b).

The obtained center line forming the skeleton of each segment is approximated by a square approximation line, and the distance r with respect to the original of the image sensor 15 and the angle ψ with respect to the horizontal axis of the image sensor 15 are calculated. The distance r and the angle ψ are defined as shown in FIG. 4 (c).

Next, for each adjacent segment, the distance r and the angle ψ are compared, and if the difference between them is within a predetermined threshold value, it is determined that they are on the same straight line, and the segments are integrated. The result is shown in Fig. 4 (d).

Through these series of processes, the basic line element corresponding to the lap joint or fillet joint model is extracted.

Finally, in the case of a lap joint, the end points of the light section images obtained from the lower plate surface of the upper and lower welded members, and in the case of the fillet joint, the light section obtained from each T-shaped member to be welded. Find the intersection of the images. These points are the image sensor 15
Is the position of the welding image on the surface of. The X mark in Fig. 4 (e) is that point.

By the way, when the welding position is detected during welding by using the light cutting method, the image sensor 5 outputs an image signal in which the light cutting image, the spatter, and the welding arc light are overlapped. In addition to this, electric noise from the servo system of the robot may be mixed.

FIG. 5 is a circuit diagram showing a specific configuration of a preprocessing circuit for effectively removing the noise component. The image processing device 5 removes noise components other than the light section image from the input image containing various noise components and selectively emphasizes the light section image, and a smoothing differentiation processing unit 20 and the emphasized input image. The compression processing unit 21 extracts information necessary for specifying the position of the light section image.

Smoothing differential processing unit 20 and shift registers 23 to 27
, First adder 28, second adder 29, and inverting circuit 3
0 and a third adder 31 are included.

The compression processing unit 21 includes a register 32, a station value storage register 33, a first comparator 34, a threshold value register 35, and a second value.
A comparator 36, a register 37, a control logic circuit 38,
Station value address storage address register 39, station value address storage register 40, image horizontal direction address register 41
Includes and.

In the smoothing differentiation processing unit 20, the data obtained from the image pickup device 15 and sampled by the A / D converter 22 and converted into a digital amount in synchronization with the clock signal are taken in,
The shift registers 23 to 27 are synchronized with the clock signal.
To one after another. Next, the first adder 28 calculates the average value Avf of the previous two pixels using the data of the shift registers 23 to 27 of the adjacent five pixels, and the second adder 29 calculates the average value Avb of the next two pixels. calculate. Further, the inverting circuit 30 inverts the sign of the output of the first adder 28, and the third adder 31 adds the output of the second adder 29 and the output of the inverting circuit 30. That is, the difference value (differential value) Df between the average values of the front two pixels and the rear two pixels is obtained. Third adder 31
Is the absolute value Dfa of the differential value. Third adder 3
The presence or absence of a carry of 1 represents the sign S of the differential value.

The smoothing differential processing unit 20 has both a smoothing function for alleviating a rapid change in brightness and a differentiating function for removing a gentle change. The output after passing through the smoothing differentiation processing unit 20 has the peak frequency If the spatial frequency domain is
The range of the period from about 4.4 pixels to about 128 pixels is the bandwidth of the output signal.

In the compression processing unit 21, the differential absolute value Df of the third adder 31
a is stored in the storage register 32 in synchronization with the clock signal. At the same time, the differential absolute value Dfa and the local value storage register 33
The output of is compared by the first comparator 34, and the differential absolute value D
The second comparator 36 compares fa with the threshold value stored in the threshold value register 35. The comparison result and the code output S of the third adder 31 are stored in the storage register 37 in synchronization with the clock signal. These operations and comparisons are performed from 1 clock rising to the next clock rising.
The circuit is configured so as to end with clock synchronization. The storage register 32 and the storage register 37 store the absolute value Dfa and the sign S of the smoothed differential value at the same time point, the comparison result CM between the absolute value Dfa and the data of the local value storage register 33, and the absolute value Dfa. The comparison result CR with the data in the threshold register is stored and held until the next rising edge of the clock.

The control logic circuit 38 receives the comparison results CM, CR, the code S, and the switching signal CZ of the horizontal scanning line of the image sensor 15 stored in the storage register 37 as input signals, and enters the address data entry command LW and stores the local value. Update command LM of register 33
Then, each command signal of the counter increase command LC for shifting the storage of the station value address storage register 40 is output.

An example of a more specific circuit configuration of the control logic circuit 38 is shown in FIG. The control logic circuit 38 uses a D flip-flop that sets and holds input data as output data at the rising edge of the clock, and refers to the previous internal data state in clock cycle units.

The output command conditions of the control logic circuit 38 are as follows.

[1] Condition for outputting address data entry command LW (a) The differential value is larger than both the outputs of the threshold value register 35 and the station value storage register 33, and the sign S does not shift from + to −.

(b) The differential value is larger than the threshold value both last time and last time, and the code S
When is + and-in the previous time.

(However, the counter increase command LC is not issued two times before.) [2] Conditions for outputting the counter increase command LC (a) When the differential value is larger than the previous threshold value and is equal to or less than the threshold value this time. (However, the counter increase command LC was not issued last time.) (B) The differential value is larger than the threshold value this time, and the sign S is +
When moving from-to. (However, the previous counter increment command LC has not been issued.) [3] Conditions for outputting the station value register update command LM (a) When the differential value is greater than both the threshold value and the maximum value.

(b) When [2] (b) was established last time.

(c) When the threshold is larger than the previous threshold and equal to or smaller than the current threshold.

In addition, address data is written at the switching portion of the horizontal scanning signal, and a counter increase command is issued at the next clock.

When the differential value is larger than both the threshold value 35 and the station value storage register 33, the control logic circuit 38 always stores the image horizontal direction on the station value address storage register 40 at the same address indicated by the station value address storage address register 39. The output of the address register 41 is written while updating the station value storage register 33. Then, when the differential value becomes smaller than the threshold value,
Alternatively, when the sign changes from + to −, the station value address storage address register 39 is increased and the storage position on the station value address storage register 40 is changed. By this operation, the local address storage register 40 stores the horizontal address of the position where the differential value is the maximum in the section in which the differential value is equal to or more than the threshold value and has the same code, together with the code S of the differential value. .

In addition, the local value storage register 33 indicating the maximum value of the differential absolute value
The value of is updated when the differential value is larger than both the value of the threshold value register and the value of the local value storage register 33,
In order to detect the maximum position in the section equal to or more than the next threshold value, a differential value at this time, that is, a value equal to or less than the threshold value is set.

Furthermore, when the sign changes from + to − with the differential value being larger than the threshold value, the station value address storage address register 3
The value of 9 is increased, the output of the image horizontal direction address register 41 is written to the station value address storage register 40 and the station value storage register 33 is updated with a delay of one clock, and either the maximum position or the minimum position is reached. It prevents the state that is not remembered.

FIG. 7 is a diagram showing a light section image obtained from the image sensor 15 together with illumination light of spatter and arc as interfering light.

8 and 9 show VIII-VIII of FIG. 7, respectively.
The change in brightness on the scan line and the IX-IX scan line, the output of the smoothing differentiation processing unit 20, and the change in each input signal and output signal in the vicinity of the position where the address data is stored are shown. .

FIG. 8 shows a state in which arc light is irradiated as illumination light from the right direction on the light section image, and interference light such as halation is also present at the left end. When such a signal from the image pickup device 15 is passed through the smoothing differentiation processing unit 20, the influence of the gradual change of the illumination light is removed. Further, the local value address storage register 33 stores the position at which the rising gradient of the light-section image becomes maximum and the position at which the falling gradient becomes minimum together with the codes.

FIG. 9 shows a state in which spatter and strong arc illumination light due to welding overlap the light section image. In the output of the smoothing differential processing unit 20, the peak of the gradient due to the disturbance and then the peak of the gradient due to the light section image that is the original detection target appear. In this case, the falling slope is only one because the ambient light and the falling portion of the light section image overlap each other. The peak value of the rising and falling slopes is stored in the station value storage register 33 together with the sign of the differential value.

The results of these pre-processings are then sent to a data processing unit (not shown) in the image processing device 5, and a light section image is detected. At that time, in the conventional example, in the case as shown in FIG. 8, depending on where the threshold is set, the influence of the arc light or the spat is directly exerted, and the light section image is present in the portion where the disturbance light is present. There was a risk of recognizing it as a position.

On the other hand, in the case of FIG. 9 as well, since it overlaps with ambient light, the result is that the rising or falling portion of the light-section image itself cannot be identified, and the accurate position of the light-section image cannot be automatically detected at high speed. It was impossible.

In the present embodiment, as shown in the bottom of FIGS. 8 and 9, even if the arc light or spatter of the light-section image is overlapped, if the brightness changes in the part of the light-section image, The maximum and minimum positions of the gradient are updated and stored,
By eliminating the disturbance of ambient light, the exact position of the light section image can be detected without interruption.

In addition, since the compression processing that stores the position where the differential value becomes maximum or minimum together with the sign of the differential value at this time is executed in synchronization with the clock signal that samples the image signal, the image data is processed at high speed. It can be compressed and the speed of the detection process of welding lines and the like is increased.

Further, even when the line width of the light-section image apparently changes due to the secondary reflection that occurs when the slit light is reflected twice on the surface of the object to be detected, the change in the interval between the maximum position and the minimum position of the gradient causes The state can be determined.

Therefore, it is possible to detect the position of the welding line at a high speed by using one kind of small head over a wide range from a thin plate with a small welding current where the state of the ambient light changes to a medium-thick plate with a large welding current, and perform copy control. Welding robot etc. can be obtained.

The present invention can be applied not only to the welding robot but also to a sealing machine, a grinder, etc. in which ambient light exists.

〔The invention's effect〕

According to the present invention, even when arc light or spatter is superposed on a light-section image, if the brightness changes in the part of the light-section image, the maximum and minimum positions of the gradient are stored to prevent disturbance of ambient light. Exclusions allow adequate detection of the exact location of the light section image.

Further, since the compression processing of storing the position where the differential value is maximum or minimum together with the sign of the differential value is performed, the image data can be compressed at high speed, and the processing for detecting the welding line or the like is speeded up.

Further, even when the line width of the light-section image changes due to the secondary reflection of the slit light on the surface of the object to be detected, this state is determined by the change in the maximum position and the minimum position of the gradient, and its influence You can encourage exclusion.

Therefore, it is possible to detect the position of the welding line at a high speed by using one kind of small head over a wide range from a thin plate with a small welding current where the state of the ambient light changes to a medium-thick plate with a large welding current, and perform copy control. Welding robot etc. can be obtained.

[Brief description of drawings]

FIG. 1 is a schematic diagram showing an example of the configuration of a welding robot system equipped with a light section image position detecting device according to the present invention, and FIG. 2 is a sensor head fixed with a welding torch under the wrist of the welding robot in FIG. FIG. 3 is a diagram showing an example of the structure, FIG. 3 is a flowchart showing a procedure of image data processing in the image processing apparatus of FIG. 1, FIG. 4 is a diagram showing an example of data at each stage of FIG. 3, and FIG. FIG. 6 is a circuit diagram showing a specific configuration of a preprocessing circuit for effectively removing noise components, FIG. 6 is a circuit diagram showing more specifically an example of the configuration of the control logic circuit in the circuit of FIG. FIG. 7 is a diagram showing a light section image obtained from the image pickup device together with arc illumination light as interfering light, and FIGS. 8 and 9 are line VI in FIG. 7, respectively.
FIG. 8 is a diagram showing changes in brightness along II-VIII and IX-IX, changes in the output of the smoothing differentiation processing unit, and changes in each input signal and output signal in the vicinity where address data is stored. 1 ... Welding robot, 2 ... Welding power source, 3 ... Robot control device, 4 ... Sensor head, 5 ... Image processing device, 6 ... Welding torch, 7 ... Semiconductor laser, 8 ... Aspherical surface Lens, 9 ... Cylindrical lens, 10 ... Slit light, 11 ... Slit light irradiation block, 12 ... Interference filter, 13 ... Objective lens, 14 ... Aperture, 15 ... Imaging element, 16 ... Bearing, 17 ... DC motor, 18 ... Gear, 19 ... Timing belt, 20 ... Smoothing differentiation processing section, 21 ... Compression processing section, 22 ... A / D converter, 23 ... 27 ... Shift register, 28 ... First adder, 29 ... Second adder, 30 ... Inversion circuit, 31 ... Third adder, 32 ... Register, 33 ... Station value storage register, 34 ... First comparator , 35 ... Threshold register, 36 ... Second comparator, 37 ...... register, 38 ...... control logic circuit, 39 ...... station Value Address Address Register, 40 ...... station value address storage register, 41 ...... image horizontal address register.

Claims (7)

[Claims] 1. A light source for irradiating an object to be detected with slit light in an environment where ambient light is present or having a surface whose reflectance is not constant, and an image pickup means for capturing a light-section image generated on the object to be detected. In a light-section image position detection device including an image processing device that processes an image signal from the image-capturing means and detects a position of the light-section image on the detected object, the image-processing device includes: Smoothing differential processing means for obtaining a differential value of the brightness of the image signal while smoothing the image signal, compression processing means for storing the image positions at which the differential value is maximum and minimum together with a code, and the stored image position An optical-cross-section image position detection device, comprising means for determining a pair whose interval from the maximum position to the minimum position is within a predetermined threshold as the position of the optical-section image. 2. The light section image position detecting device according to claim 1, wherein the smoothing differential processing means captures image data from the imaging means via an A / D converter and transfers the image data one after another. A register, first and second addition means for selecting two sets of pixel data before and after these shift registers to obtain an average value of each set, and a sign of an output of the first addition means for inverting the sign of the second addition means. A light-section image position detecting device, comprising: a third addition means for adding the output and the differential value of each average value of the front two pixels and the rear two pixels. 3. The optical section image position detecting device according to claim 1, wherein the compression processing means stores a local maximum value or a minimum value of the output of the smoothing differential processing means. , First comparing means for comparing the output of the smoothing and differential processing means with the output of the local value register, a threshold value register for storing the threshold value of the differential value, an output of the threshold value register, and an output of the smoothing and differential processing means Second comparison means for comparing the station value and a station value address storage register for storing the station value and an address of the station value, an address counter for counting a clock signal, and a differential value code output of the smoothing differentiation processing means Based on the output of the first comparing means and the output of the second comparing means, an update command for updating the station value storage register with the output of the smoothing differentiation processing means and the output of the address counter. And a control logic circuit for issuing a storage command for storing the sign of the differential value of the smoothing differentiating means in the local value address storage register, and the local maximum of the gradient in which the gradient of the brightness of the light section image is equal to or more than a threshold value. An optical cutting image position detecting device, which is a means for storing the position and the local minimum position together with the sign of the gradient. 4. The light source and the image pickup means of the light section image position detecting device according to claim 1, the welding torch is fixed to a wrist of a welding arm, and the welding torch is set at a target position. A welding robot characterized by copying control. 5. The light source and the image pickup means of the light section image position detecting device according to claim 1, wherein the light source and the image pickup means are fixed to a ceiling head, and the ceiling head is controlled to a target position. Sealing machine characterized by. 6. The light source and the image pickup means of the light section image position detecting device according to claim 1, wherein the light source and the image pickup means are fixed near a grinder disk, and the grinder disk is controlled to follow a target position. A grinder characterized by that. 7. The rising slope and falling slope of the brightness of a light-section image obtained by irradiating an object to be detected with slit light under a boundary in the presence of ambient light or having a surface whose reflectance is not constant, respectively. A light-section image data compression processing circuit comprising means for storing the maximum and minimum positions and means for storing the sign of the gradient in association with the maximum and minimum positions.