JP6179159B2 - Appearance inspection system - Google Patents

Description

  The present invention relates to a system for inspecting the appearance of an object based on an image obtained by photographing the object with a camera.

  Conventionally, as this type of system, there is a system in which a workpiece to be inspected is photographed with a CCD camera, and the luminance data of the portion to be inspected is compared with standard data to determine the quality of the appearance (see Patent Document 1).

Japanese Utility Model Publication No. 7-23274

  By the way, when light is applied to a work by a lighting device and shooting is performed with a camera, reflected light from the work may become too strong. In that case, the appearance cannot be accurately inspected particularly in a portion where the reflected light is strong. For this reason, the worker repeatedly shoots the workpiece many times while searching for a shooting position where the reflected light from the workpiece is appropriate. Therefore, it takes time and effort to adjust the photographing position, and the inspection efficiency is reduced.

  The present invention has been made to solve these problems, and a main object of the present invention is to provide an appearance inspection system capable of improving the efficiency of inspection.

  The present invention employs the following means in order to solve the above problems.

  The first means includes an illumination device that shines light on a target, a camera that captures the target, and a determination unit that determines the quality of the appearance of the target based on an image of the target captured by the camera. The illumination device includes a light source, and an adjustment unit that adjusts a direction and an amount of light transmitted through the light source, and the determination unit is the target imaged by the camera. On the basis of the image, the adjusting unit is controlled so as to reduce the amount of light directed to a specific portion of the target where the intensity of reflected light is higher than a predetermined value.

  According to the said structure, light is irradiated to object with an illuminating device and object is image | photographed with a camera. Then, the quality of the appearance of the target is determined by the determination unit based on the target image captured by the camera.

  Here, the illumination device includes a light source and an adjusting unit that adjusts the direction and amount of light transmitted from the light source. Then, the adjustment unit is controlled by the determination unit so as to reduce the amount of light directed to a specific portion of the target in which the intensity of the reflected light is higher than a predetermined value, based on the target image captured by the camera. . For this reason, the reflected light from a specific part becomes weak and the inspection precision of a specific part can be improved. As a result, it is not necessary for the operator to adjust the photographing position and the inspection efficiency can be improved.

  In the second means, the determination unit causes the camera to shoot the target, and reduces the amount of light directed to the specific portion by a predetermined amount based on the image of the target captured by the camera. The control of the adjusting unit is repeatedly executed.

  According to the above configuration, the object is photographed by the camera, and the amount of light traveling toward the specific unit based on the photographed image is reduced by a predetermined amount. Therefore, the amount of light traveling toward the specific portion can be gradually decreased, and the amount of light traveling toward the specific portion can be easily adjusted appropriately.

  In the third means, the determination unit controls the adjustment unit so that the intensity of the reflected light from the specific portion becomes weaker than the determination value from the start of light emission by the light source to the stop thereof. .

  According to the above configuration, the adjustment unit is controlled so that the intensity of the reflected light from the specific portion becomes weaker than the determination value from when light emission is started by the light source to when it is stopped. For this reason, it is possible to avoid a plurality of times of light emission different from the operator's intention, and to suppress the operator from feeling uncomfortable.

  In the fourth means, the determination unit causes the camera to photograph the object, and reduces the amount of light directed to the specific portion by a predetermined amount based on the image of the object photographed by the camera. Control of the adjusting unit is repeatedly performed from the start of light emission by the light source to the stop thereof, so that the intensity of reflected light from the specific portion becomes weaker than a determination value.

  According to the above configuration, since both the second means and the third means are provided, the object is imaged by the camera from when the light emission is started to when it is stopped, and based on the captured image. It is repeated to reduce the amount of light directed to the specific part by a predetermined amount. Then, the intensity of the reflected light from the specific portion is made weaker than the determination value after the light emission is started and stopped by the light source. Therefore, it becomes easy to appropriately adjust the amount of light directed to the specific portion while suppressing the operator from feeling uncomfortable.

  In the fifth means, the determination unit is a non-defective image of the target photographed by the camera in a predetermined state in which the adjustment unit is controlled such that the intensity of the reflected light from the specific part is weaker than a determination value. Is used as a model image, and the control unit is controlled to the predetermined state and the target image captured by the camera is compared with the model image to determine the quality of the target appearance.

  When visual inspection is performed on a plurality of objects having the same shape, specific portions whose reflected light intensity is higher than a predetermined value in the plurality of objects substantially coincide with each other.

  In this regard, according to the above configuration, in a predetermined state in which the adjustment unit is controlled so that the intensity of the reflected light from the specific portion is weaker than the determination value, the target non-defective image captured by the camera is a model image. Is done. Then, the control unit is controlled to the predetermined state, and the target image captured by the camera is compared with the model image, and the quality of the target appearance is determined. For this reason, an adjustment part can be made into the common predetermined state in the some object of the same shape, and the image of the quality goods of the object image | photographed with the camera in the predetermined state can be made into a common model image. Therefore, it is not necessary to control the adjusting unit every time, and the inspection efficiency can be further improved.

  In a sixth means, the determination unit includes a nonvolatile storage unit that stores data, and stores the predetermined state and the model image in the storage unit.

  According to the above configuration, the predetermined state in which the adjustment unit is controlled so that the intensity of the reflected light from the specific portion is weaker than the determination value, and the model image obtained by photographing the target non-defective product in the predetermined state are nonvolatile. Stored in the storage unit. Therefore, when inspecting an object having the same shape as the object inspected before, the predetermined state and the model image can be read from the storage unit and used.

  In the seventh means, the adjusting section is constituted by a liquid crystal panel.

  According to the above configuration, since the adjustment unit is configured by the liquid crystal panel, the light transmission direction and the light transmission amount can be adjusted easily and accurately.

The schematic diagram which shows the outline | summary of an external appearance inspection system. The flowchart which shows the process sequence of liquid crystal mask pattern and model image creation. The figure which shows a liquid crystal mask pattern, a gray scale image, and a gradation. The flowchart which shows the process sequence of the quality determination of a workpiece | work. The figure which shows a gradation.

  Hereinafter, an embodiment embodying a work appearance inspection system will be described with reference to the drawings. The appearance inspection system of this embodiment is installed in an assembly line of a machine assembly factory, for example.

  FIG. 1 is a schematic diagram showing an outline of the appearance inspection system 10. The appearance inspection system 10 includes a strobe 20, a camera 30, and a control device 40.

  The strobe 20 (lighting device) is attached to a camera 30 and includes a main body 21, a lamp 22, a liquid crystal panel 23, a lens unit 24, a cover 25, and the like. The main body 21 is formed in a hollow rectangular parallelepiped shape. Inside the main body 21, a lamp 22, a liquid crystal panel 23, and a lens unit 24 are accommodated. A transparent cover 25 that transmits the light of the lamp 22 is attached to the front surface of the main body 21.

  The lamp 22 (light source) is composed of a known light emitting element, and emits light that strikes the workpiece W (target). Lighting and extinguishing of the lamp 22 is controlled by the control device 40. The liquid crystal panel 23 (adjustment unit) is formed in a plate shape and includes a polarizing plate, an alignment film, a liquid crystal layer, and the like. The liquid crystal panel 23 adjusts the direction (transmission position) in which light from the lamp 22 is transmitted and the amount of light transmitted (transmittance) by controlling the state of the liquid crystal layer. The adjustment state of light by the liquid crystal panel 23 is controlled by the control device 40. The lens unit 24 diffuses the light transmitted through the liquid crystal panel 23 and irradiates the light toward the cover 25. The focal point of the lens unit 24 is automatically adjusted so that light is irradiated to a predetermined position of the workpiece W.

  The camera 30 is a CCD camera or the like capable of color photography, and photographs the workpiece W to acquire an image. The camera 30 is attached to a hand portion (tip portion) of a robot arm (not shown). The position and direction of the arm of the robot is changed by driving each joint. Then, the position and direction of the camera 30 are changed by changing the position and direction of the arm. Thereby, the workpiece | work W conveyed to a test | inspection position by a belt conveyor etc. can be image | photographed from arbitrary viewpoints with the camera 30. FIG. The photographing operation of the camera 30 is controlled by the control device 40.

  The control device 40 (determination unit) includes a CPU, a ROM, a RAM, a storage unit 40a, a drive circuit, a position detection circuit, and the like. The ROM stores a robot system program, an operation program, and the like. The RAM temporarily stores parameter values and the like when executing these programs. The storage unit 40a is a nonvolatile memory that stores data. The position detection circuit detects a rotation angle of a motor provided at each joint based on a detection signal of an encoder provided at each joint of the arm. The CPU performs feedback control of the rotation angle of each joint of the arm to the target rotation angle based on the position information input from the position detection circuit by executing a preset operation program.

  In the present embodiment, the control device 40 determines the position and direction of the arm when the workpiece W transported to the inspection position is imaged by the camera 30 as the target position and direction of the arm in the appearance inspection of the workpiece W (shooting position). ) Is set. By driving the arm to the photographing position, the workpiece W is captured in a state suitable for inspection within the field of view of the camera 30. In this state, an image of the workpiece W is acquired by photographing the workpiece W with the camera 30. Further, the control device 40 controls the operations of the strobe 20 and the camera 30 to create a mask pattern and a model image as shown below.

  FIG. 2 is a flowchart showing a processing procedure for creating a liquid crystal mask pattern and a model image. This series of processing is executed by the control device 40 during teaching (initial setting) of the robot before the appearance inspection is started. In this series of processes, the workpiece W that has been determined as a non-defective product by the operator in advance is used.

  First, the liquid crystal mask pattern formed by the liquid crystal panel 23 is erased (S11). Specifically, prior to creating a mask pattern with the liquid crystal panel 23, the state of the liquid crystal panel 23 is controlled so that the light from the lamp 22 is not restricted by the liquid crystal panel 23.

  Subsequently, the lamp 22 of the strobe 20 is turned on, and the work W is photographed in color by the camera 30 (S12). Then, a color image of the photographed work W is input from the camera 30, and the color image is converted into a gray scale image (S13). Specifically, a color image is converted into a grayscale image by a known method such as a simple average method, an intermediate value method, or a G channel method. Here, the lamp 22 of the strobe 20 is lit continuously until the subsequent processes of S13 to S21 are completed.

  Subsequently, a pixel number i that designates a pixel of the grayscale image is set to an initial value 0 (S14). Specifically, there are 0 to n pixels in the image, and the first pixel number is 0. Then, it is determined whether or not the current pixel number i is smaller than n (S15).

  In the above determination, when it is determined that the current pixel number i is smaller than n (S15: YES), it is determined whether or not the gradation T (i) of the pixel number i is larger than the threshold value R1 (S16). Specifically, the gradation T takes a value from 0 (black) to 255 (white). The threshold value R1 (predetermined value, determination value) is set to a value obtained by subtracting the allowable error M at the time of inspection from 255.

  In the above determination, when it is determined that the gradation T (i) of the pixel number i is larger than the threshold value R1 (S16: YES), the current pixel number i is stored in a predetermined area of the RAM (S17). Then, 1 is added to the pixel number i to make it a new pixel number i (S18). Thereafter, the process is executed again from the process of S15. On the other hand, when it is determined in the above determination that the gradation T (i) of the pixel number i is not greater than the threshold value R1 (S16: NO), the process of S18 is executed.

  If it is determined in S15 that the current pixel number i is not smaller than n (S15: NO), it is determined whether there is a pixel number i stored in a predetermined area of the RAM (S19). . That is, it is determined whether or not there is a pixel number i whose gradation T (i) is larger than the threshold value R1.

  In the above determination, when it is determined that there is a pixel number i stored in a predetermined area of the RAM (S19: YES), the light transmission amount at the location corresponding to the pixel number i in the liquid crystal panel 23 is decreased by a predetermined amount. (S20). Specifically, the state of the liquid crystal panel 23 is changed so as to reduce the amount of light directed to the portion (specific portion) photographed as the pixel number i by 10%. Then, the state of the liquid crystal panel 23 is stored in a predetermined area of the RAM as a liquid crystal mask pattern (S21). In the process of S20, only the state of the part corresponding to the pixel number i in the liquid crystal panel 23 is changed, and the other part is maintained in the state stored in the RAM. Thereafter, the changed liquid crystal panel 23 is used as the liquid crystal mask pattern, and the process from S12 is executed again.

  On the other hand, if it is determined in the above determination that there is no pixel number i stored in the predetermined area of the RAM (S19: NO), all the pixel numbers i stored in the predetermined area of the RAM are erased (S22). Further, the lamp 22 of the strobe 20 is turned off. Subsequently, the liquid crystal mask pattern stored in the predetermined area of the RAM is stored in the nonvolatile storage unit 40a (S23). Specifically, after being repeatedly changed by the process of S20, the final liquid crystal mask pattern stored in the RAM by the process of S21 is stored in the nonvolatile storage unit 40a.

  Subsequently, the gray scale image when the process of S12 is executed last is stored in the nonvolatile storage unit 40a (S24). That is, the gray scale image obtained by photographing and converting the workpiece W using the final liquid crystal mask pattern is stored as a model image in the nonvolatile storage unit 40a. Then, this series of processing ends (END).

  Next, specific modes of creating the mask pattern and model image described above will be described. FIG. 3 is a diagram illustrating a liquid crystal mask pattern, a gray scale image, and gradation. Here, a work W that has been previously determined as a non-defective product by the operator is used.

  As shown in the figure, in the first shooting, the workpiece W is shot in a state where the liquid crystal mask pattern formed by the liquid crystal panel 23 is erased, that is, in a state where transmitted light is not limited by the liquid crystal panel 23. . In the grayscale image converted from the photographed color image, the reflected light from the vicinity of the center is strong. For this reason, the gradation of the pixel on the straight line C1 is 255 gradation which is the maximum value in the area A1. Therefore, even if there is a scratch or the like in the area A1 of the workpiece W, it cannot be accurately inspected.

  Therefore, in the second shooting, the state of the liquid crystal panel 23 is set so as to reduce the light transmission amount of the portion corresponding to the pixel number i determined that the gradation T (i) is larger than the threshold value R1 by 10%. has been changed. For this reason, in the grayscale image using this liquid crystal mask pattern, the reflected light from near the center is weaker than the grayscale image obtained by the first shooting. The gradation of the pixel on the straight line C2 is slightly smaller than the 255 gradation in the area A2. In this way, until the pixel number i for which the gradation T (i) is determined to be larger than the threshold value R1 disappears, the light transmission amount at the location corresponding to the pixel number i is 10% each time by the liquid crystal panel 23. It is being reduced (in steps).

  At this time, if the amount of light transmitted through the portion corresponding to the pixel number i is greatly reduced at once, the gradation graph tends to be stepped around the threshold value R1. In this case, the gradation may be significantly lower than the threshold value R1, and although the gradation can be made smaller than the threshold value R1, the appearance evaluation of the workpiece W may not be correctly performed. On the other hand, by decreasing the light transmission amount step by step (a little by little), it is possible to reduce the possibility that the gradation is significantly lower than the threshold value R1, and the appearance of the workpiece W after the gradation adjustment is evaluated. Can be performed almost correctly.

  In the n-th shooting, the liquid crystal mask pattern reduces the amount of light transmitted most near the center (lowest transmittance), and decreases the degree to which the amount of transmitted light decreases as it spreads to the periphery (transmission). Rate). In a grayscale image using this liquid crystal mask pattern, the reflected light is weakened substantially uniformly near the center. The gradation of the pixels on the straight line Cn is slightly smaller than the threshold value R1 even in the area An. As a result, in the grayscale image, the pixel number i determined that the gradation T (i) is larger than the threshold value R1 does not exist.

  Accordingly, the state (predetermined state) of the liquid crystal panel 23 at this time is stored in the nonvolatile storage unit 40a as a final liquid crystal mask pattern. Further, the grayscale image at this time is stored in the nonvolatile storage unit 40a as a model image.

  FIG. 4 is a flowchart showing a processing procedure for determining whether the workpiece W is acceptable. This series of processing is executed by the control device 40 during continuous operation of the robot during visual inspection. In this series of processes, the appearance of the work W having the same shape as the work W used in creating the mask pattern and the model image is inspected. In addition, about the process same as FIG. 2, detailed description is abbreviate | omitted by attaching | subjecting the same step number.

  First, the liquid crystal mask pattern and model image stored in the nonvolatile storage unit 40a are read (S31, S32). Then, the state of the liquid crystal panel 23 is controlled so as to become the read liquid crystal mask pattern (S33).

  Subsequently, the lamp 22 of the strobe 20 is turned on, and the work W is photographed in color by the camera 30 (S12). Then, a color image of the photographed work W is input from the camera 30, and the color image is converted into a gray scale image (S13).

  Subsequently, a pixel number i that designates a pixel of the grayscale image is set to an initial value 0 (S14). Then, it is determined whether or not the current pixel number i is smaller than n (S15).

  In the above determination, when it is determined that the current pixel number i is smaller than n (S15: YES), the gradation T (i) of the pixel number i in the grayscale image and the gradation Tm of the pixel number i in the model image It is determined whether or not the deviation from (i) is larger than the threshold value R2 (S36). The threshold value R2 is set to an allowable error M at the time of inspection.

  In the above determination, when it is determined that the deviation between the gradation T (i) and the gradation Tm (i) is larger than the threshold value R2 (S36: YES), the current pixel number i is stored in a predetermined area of the RAM. (S17). Then, 1 is added to the pixel number i to make it a new pixel number i (S18). Thereafter, the process is executed again from the process of S15. On the other hand, in the above determination, when it is determined that the deviation between the gradation T (i) and the gradation Tm (i) is not larger than the threshold value R2 (S36: NO), the process of S18 is executed.

  If it is determined in S15 that the current pixel number i is not smaller than n (S15: NO), it is determined whether there is a pixel number i stored in a predetermined area of the RAM (S19). . That is, it is determined whether or not there is a pixel number i whose deviation between the gradation T (i) and the gradation Tm (i) is larger than the threshold value R2.

  If it is determined in the above determination that there is a pixel number i stored in a predetermined area of the RAM (S19: YES), it is determined that the work W currently being inspected is a defective product (S37). On the other hand, if it is determined in the above determination that there is no pixel number i stored in the predetermined area of the RAM (S19: NO), it is determined that the work W currently being inspected is a non-defective product (S38). Then, this series of processing ends (END). In addition, when performing a visual inspection continuously about the several workpiece | work W of the same shape, it performs again from the process of S12.

  The embodiment described in detail above has the following advantages.

  The strobe 20 includes a lamp 22 and a liquid crystal panel 23 that adjusts the direction and amount of light transmitted from the lamp 22. Then, on the basis of the image of the workpiece W taken by the camera 30 by the control device 40, the liquid crystal panel is configured so as to reduce the amount of light directed to the specific portion of the workpiece W where the intensity of the reflected light is higher than the threshold value R1. 23 is controlled. For this reason, the reflected light from a specific part becomes weak and the inspection precision of a specific part can be improved. As a result, it is not necessary for the operator to adjust the photographing position and the inspection efficiency can be improved.

  -The work 30 is imaged by the camera 30 and the amount of light traveling toward the specific portion is repeatedly reduced by 10% based on the captured image. Therefore, the amount of light traveling toward the specific portion can be gradually decreased, and the amount of light traveling toward the specific portion can be easily adjusted appropriately.

  The liquid crystal panel 23 is controlled so that the intensity of the reflected light from the specific portion becomes weaker than the threshold value R1 before the light emission is started and stopped by the lamp 22. For this reason, it is possible to avoid a plurality of times of light emission different from the operator's intention, and to suppress the operator from feeling uncomfortable.

  In a predetermined state (liquid crystal mask pattern) in which the liquid crystal panel 23 is controlled so that the intensity of the reflected light from the specific portion is weaker than the threshold value R1, a good image of the work W photographed by the camera 30 is a model image. Is done. Then, the image of the workpiece W photographed by the camera 30 by controlling the liquid crystal panel 23 to a predetermined state is compared with the model image, and the quality of the appearance of the workpiece W is determined. For this reason, the liquid crystal panel 23 can be set to a common predetermined state in a plurality of workpieces W having the same shape, and a non-defective image of the workpiece W photographed by the camera 30 in a predetermined state can be used as a common model image. Therefore, it is not necessary to control the liquid crystal panel 23 every time, and the inspection efficiency can be further improved.

  A predetermined state (liquid crystal mask pattern) in which the liquid crystal panel 23 is controlled so that the intensity of reflected light from the specific portion is weaker than the threshold value R1, and a model image obtained by photographing a non-defective product of the workpiece W in the predetermined state are nonvolatile. Stored in the sex storage unit 40a. Therefore, when inspecting a workpiece W having the same shape as the workpiece W previously inspected, the liquid crystal mask pattern and the model image can be read from the storage unit 40a and used.

  A liquid crystal panel 23 is employed as an adjustment unit that adjusts the direction and amount of light transmitted from the lamp 22. For this reason, a mask pattern can be formed easily and accurately.

  In addition, the said embodiment can also be changed and implemented as follows. About the same member as the said embodiment, description is abbreviate | omitted by attaching | subjecting the same code | symbol.

  In the above embodiment, the predetermined state (liquid crystal mask pattern) and the model image of the liquid crystal panel 23 and the model image are stored in the nonvolatile storage unit 40a, but the appearance inspection of the work W is performed following the creation of the liquid crystal mask pattern and the model image. When performing, a liquid crystal mask pattern and a model image may be stored in a RAM (volatile memory).

  When the appearance inspection is performed on the workpiece W having a different shape, the processes of S11 to S24 in FIG. 2 may be executed each time.

  In the above embodiment, the liquid crystal panel 23 is controlled so that the intensity of the reflected light from the specific portion becomes weaker than the threshold value R1 before the light emission is started by the lamp 22 and stopped. The lamp 22 may emit light a plurality of times until the above control is completed.

  In the above embodiment, the work 30 is photographed by the camera 30 and the amount of light traveling toward the specific portion is repeatedly reduced by 10% based on the photographed image. However, the amount of light is 10% or more at a time. It can also be reduced. As a result, the intensity of the reflected light from the specific portion may be weaker than the threshold value R1 by a single adjustment.

  Further, as shown in FIG. 5, the adjustment of the amount of transmitted light by the liquid crystal mask pattern can be completed after three photographings. First, in the first shooting, as in the above embodiment, the workpiece W is shot in a state where the liquid crystal mask pattern formed by the liquid crystal panel 23 is erased, that is, in a state where transmitted light is not limited by the liquid crystal panel 23. ing. Note that the gradation of the output pixel does not become larger than the maximum value of 255 gradations, but the actual gradation is expected to increase toward the center as shown by the broken line in FIG.

  Therefore, in the second shooting, the state of the liquid crystal panel 23 is changed so as to reduce the total light transmission amount by, for example, 70% (predetermined amount). For this reason, the gradation of the pixels is small as a whole, and even the maximum value is smaller than 255 gradations. Here, as indicated by a two-dot chain line, an actual gradation before reducing the entire light transmission amount by 70% is estimated based on the gradation obtained by the second photographing.

  In the third shooting, the liquid crystal mask pattern adjusts the amount of light transmission at each location so that the gray level of each pixel is smaller than the threshold value R1 based on the gray level estimated as described above. doing. As a result, similar to the above embodiment, the amount of transmitted light is reduced most in the vicinity of the center (the transmittance is reduced most), and the degree of decrease in the amount of transmitted light is reduced as it spreads to the periphery (the transmittance is increased). ) According to such a process, the adjustment of the liquid crystal mask pattern can be basically completed by photographing three times, and the liquid crystal mask pattern can be appropriately adjusted while reducing the number of processes.

  A camera that captures black and white images can be used as the camera 30. In that case, conversion to a grayscale image is not necessary, and it is only necessary to compare the luminance of the pixel number i in the black and white image with the luminance of the pixel number i in the model image to make a pass / fail determination.

  Any light source can be used as the lamp 22 of the strobe 20 as long as the light source emits light whose amount of transmission can be adjusted by the liquid crystal panel 23.

  In the above embodiment, the liquid crystal panel 23 is used as the adjusting unit that adjusts the direction and amount of light transmitted from the lamp 22, but other types of adjusting units such as a thermochromic panel whose optical characteristics change depending on the temperature. Can also be adopted.

  DESCRIPTION OF SYMBOLS 10 ... Appearance inspection system, 20 ... Strobe (illuminating device), 22 ... Lamp (light source), 23 ... Liquid crystal panel (adjustment part), 30 ... Camera, 40 ... Control apparatus (determination part), 40a ... Memory | storage part.

Claims (4)

An appearance inspection system comprising: an illumination device that shines light on a target; a camera that captures the target; and a determination unit that determines the quality of the appearance of the target based on an image of the target captured by the camera. And
The lighting device includes a light source, and an adjustment unit that adjusts a direction and an amount of light transmitted from the light source.
The determination unit is configured based on the subject image captured by the camera, pre-SL as the intensity of the reflected light decreases only slightly the amount of light directed to the strong specific parts than the predetermined value of the object by controlling the regulating unit, while the intensity of the reflected light from the specific portion is less likely to significantly below the predetermined value, so that the intensity of the reflected light from the specific portion is weaker than the predetermined value A non-defective image of the target photographed by the camera in a predetermined state in which the adjustment unit is controlled is created as a model image, and the target image photographed by the camera by controlling the adjustment unit in the predetermined state; An appearance inspection system that compares the model image and determines whether the appearance of the object is good or bad . The determination unit, until stopped from to initiate the emission by the light source such that said intensity of the reflected light from the specific part weaker than the predetermined value, according to claim 1 for controlling the adjusting portion Visual inspection system. The determination unit includes a non-volatile storage unit for storing data, visual inspection system according to claim 1 or 2 and stores the predetermined condition and the model image in the storage unit. The controller may control the appearance inspection system according to any one of claims 1 to 3 which is constructed by a liquid crystal panel.