Electronic apparatus and method for selecting rank

Abstract

To provide an electronic apparatus and a method for selecting a rank which can make an analysis without deteriorating accuracy even when an amount of light is not optimum.SOLUTION: The electronic apparatus according to the present disclosure includes: a condition setting unit for setting a condition of light irradiated to an inspection object; a sensor processing unit for cutting out an output range of output data of an inspection wavelength generated by reflection light of the light set as an inspection wavelength, and generating differential data; and a characteristic amount calculation unit for calculating a characteristic amount of the inspection object from the differential data and selecting a rank of the inspection object. The differential data is cut out as the range representing the characteristic amount of a state of the inspection object.SELECTED DRAWING: Figure 4

Description

[Technical field]

[0001] The present disclosure relates to an electronic device and a rank selection method. [Background technology]

[0002] 2. Description of the Related Art Known techniques for analysis using light include techniques for analyzing an object by utilizing the fact that a difference in output at a specific wavelength is observed depending on the state of the object. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2020-176952 A Summary of the Invention [Problem to be solved by the invention]

[0004] The above-mentioned device irradiates light onto an object when performing a spectral analysis of the object. In addition, when the object is meat or the like, the above-mentioned device calculates the intensity and time of light to be irradiated onto the object by measuring the distance from the device to the object and the moving speed of the object before analysis in order to minimize the effect of heat caused by the irradiation of light, and adjusts the amount of light to be irradiated to conditions according to the object.

[0005] In the above-described device, when a multispectral light source is generated by dispersing light from a single light source, if the light intensity is adjusted to a specific wavelength, the amount of light at another wavelength may not be optimal.

[0006] In view of these problems, the present disclosure provides an electronic device and a ranking method that can perform analysis without reducing accuracy even when the amount of light is not optimal. [Means for solving the problem]

[0007] The electronic device according to the first aspect of the present disclosure includes a condition setting unit that sets the conditions of light to be irradiated onto an object to be inspected, a sensor processing unit that cuts out an output range of output data of the inspection wavelength generated by the reflected light of the light set as the inspection wavelength, and generates differential data, and a feature amount calculation unit that calculates a feature amount of the object to be inspected from pixel data of the differential data and performs rank selection of the object to be inspected, and the differential data is cut out as a range indicating a feature amount of the state of the object to be inspected. As a result, the electronic device can narrow the range of output data, thereby allowing the electronic device to obtain a sufficient output range of gradation within the range, and thus improving the accuracy of rank selection. In addition, the electronic device can reduce the amount of output data by cutting out only information necessary in the sensor as output data. In addition, even if the electronic device generates a multispectral light source by dispersing light from one light source and adjusts the amount of light to a predetermined wavelength, the electronic device can improve the analysis accuracy at other wavelengths by narrowing the output range by cutting out the output data. In addition, the electronic device can shorten the measurement time for rank selection by reducing the amount of output data.

[0008] Also, in this first aspect, the sensor processing unit includes a RAMP DAC that is provided inside a sensor that outputs pixel data of the difference data based on the reflected light of the light, and generates a reference signal whose level changes with time at a constant slope, a CLP DAC that generates an offset voltage of the reference signal, and a GAIN DAC that generates an analog gain. As a result, the electronic device can narrow the range of output data, thereby ensuring a sufficient output range of gradations within that range, and improving the accuracy of rank selection. Also, the electronic device can reduce the amount of output data by extracting only information necessary in the sensor as output data. Also, even if the electronic device generates a multispectral light source by dispersing light from one light source and adjusts the amount of light to a predetermined wavelength, the electronic device can improve the analysis accuracy at other wavelengths by narrowing the output range by extracting the output data. Also, the electronic device can shorten the measurement time for rank selection by reducing the amount of output data.

[0009] In addition, in this first aspect, the feature amount calculation unit is provided inside a sensor that outputs pixel data of the difference data based on the reflected light of the light. As a result, the electronic device can narrow the range of output data, thereby ensuring a sufficient output range of gradations within that range, and improving the accuracy of rank selection. In addition, the electronic device can reduce the amount of output data by extracting only information required in the sensor as output data. In addition, even if the electronic device generates a multispectral light source by dispersing light from one light source and adjusts the amount of light to a predetermined wavelength, the electronic device can improve the analysis accuracy at other wavelengths by narrowing the output range by extracting the output data. In addition, the electronic device can shorten the measurement time for rank selection by reducing the amount of output data.

[0010] In addition, in this first aspect, the sensor processing unit adds or subtracts the offset voltage to the comparison voltage in the reference signal, truncates the lower data of the output data of the inspection wavelength, changes the slope of the comparison voltage by the analog gain, truncates the upper data of the output data of the inspection wavelength, and cuts out the difference data by truncating the upper data and the lower data. As a result, the electronic device can narrow the range of output data, thereby ensuring a sufficient output range of gradations within that range, and improving the accuracy of rank selection. Furthermore, the electronic device can reduce the amount of output data by cutting out only information necessary in the sensor as output data. Furthermore, even if the electronic device generates a multispectral light source by dispersing light from one light source and adjusts the amount of light to a predetermined wavelength, the electronic device can improve the analysis accuracy at other wavelengths by narrowing the output range by cutting out the output data. Furthermore, the electronic device can shorten the measurement time for rank selection by reducing the amount of output data.

[0011] In the first aspect, the sensor processing unit is controlled by a control circuit provided inside the sensor, whereby feature calculation and rank selection can be performed inside the sensor, thereby speeding up the processing.

[0012] The electronic device according to the second aspect of the present disclosure includes a condition setting unit that sets the conditions of light to be irradiated onto an object to be inspected, a sensor processing unit that cuts out output ranges of output data of each of the inspection wavelengths generated by the reflected light of the light set as a plurality of inspection wavelengths and generates respective difference data, and a feature calculation unit that calculates a feature of the object to be inspected from pixel data of each of the difference data and performs rank selection of the object to be inspected, and each of the difference data is cut out as a range indicating a feature of the state of the object to be inspected. This allows the electronic device to realize feature calculation using a plurality of inspection wavelengths. Furthermore, the electronic device can perform rank selection with higher accuracy by calculating feature amounts from a plurality of inspection wavelengths and performing rank selection.

[0013] In addition, in the second aspect, the condition setting unit divides an area corresponding to one image data, and the sensor processing unit generates the difference data for each divided area using different inspection wavelengths. This allows the electronic device to reduce the amount of data when performing rank selection.

[0014] In the second aspect, the inspection wavelength is changed by control of a control circuit provided inside the sensor, whereby the sensor controls the light source to change the inspection wavelength, and thus the electronic device can acquire data at high speed.

[0015] In addition, in the first aspect, the electronic device further includes an output intensity confirmation unit that normalizes output data of the inspection wavelength based on a comparison between an actual measurement value obtained when the inspection object is irradiated with a normalization processing wavelength different from the inspection wavelength and an expected measurement value obtained when the inspection object is irradiated with the normalization processing wavelength. As a result, the electronic device normalizes the output data using the normalization processing wavelength and then performs rank selection, thereby improving the accuracy of rank selection.

[0016] In addition, in this first aspect, the output intensity confirmation unit modifies the output data of the inspection wavelength so that an actual measured value obtained when the inspection object is irradiated with the normalization processing wavelength coincides with the expected measured value obtained when the inspection object is irradiated with the normalization processing wavelength. This allows the electronic device to normalize the output data using the normalization processing wavelength and then perform rank selection, thereby improving the accuracy of rank selection.

[0017] In addition, in the first aspect, the electronic device further includes an object identification unit that identifies the type of the object to be inspected based on a first specific wavelength and a second specific wavelength, which are wavelengths at which a difference appears in output data depending on the type of the object when light is irradiated onto the object to be inspected. This allows the electronic device to identify the object and perform rank sorting even when the type of the object to be inspected has not been identified.

[0018] In addition, in this first aspect, the object identification unit calculates a ratio between the first output data and the second output data generated by the reflected light of the light set as the first specific wavelength and the second specific wavelength, and compares the calculated ratio between the first output data and the second output data with the ratio between the first output data and the second output data for each type of object to identify the type of object. This allows the electronic device to identify the object and perform rank sorting even when the type of the object to be inspected has not been identified.

[0019] In the first aspect, the object identification unit is provided inside a sensor that outputs pixel data of the difference data based on the reflected light of the light, thereby enabling the type of object to be identified inside the sensor, thereby speeding up processing.

[0020] A rank selection method according to a third aspect of the present disclosure includes a condition setting step of setting conditions for light to be irradiated onto an object to be inspected, a sensor processing step of extracting an output range of output data of the inspection wavelength generated by the reflected light of the light set as the inspection wavelength to generate difference data, and a feature calculation step of calculating a feature of the object to be inspected from pixel data of the difference data to perform rank selection of the object to be inspected, the difference data being extracted as a range indicating a feature of the state of the object to be inspected. As a result, the electronic device can narrow the range of output data, thereby allowing the output range of gradation to be sufficiently obtained within the range, and the accuracy of rank selection can be improved. In addition, the electronic device can reduce the amount of output data by extracting only information necessary in the sensor as output data. In addition, even if the electronic device generates a multispectral light source by dispersing light from one light source and adjusts the amount of light to a predetermined wavelength, the electronic device can improve the analysis accuracy at other wavelengths by narrowing the output range by extracting the output data. In addition, the electronic device can shorten the measurement time for rank selection by reducing the amount of output data.

[0021] In the third aspect, the sensor processing step generates a reference signal whose level changes with time at a constant slope, and generates an offset voltage of the reference signal; The step of generating an analog gain is included. As a result, the range of the output data is narrowed, and the electronic device can obtain a sufficient output range of gradations within the range, thereby improving the accuracy of rank selection. Furthermore, the electronic device can reduce the amount of output data by extracting only information required in the sensor as output data. Furthermore, even if the electronic device generates a multispectral light source by dispersing light from one light source and adjusts the amount of light to a predetermined wavelength, the electronic device can improve the analysis accuracy at other wavelengths by narrowing the output range by extracting the output data. Furthermore, the electronic device can shorten the measurement time for rank selection by reducing the amount of output data.

[0022] In the third aspect, the sensor processing step further includes a step of adding or subtracting the offset voltage to or from the comparison voltage in the reference signal, truncating the lower data of the output data of the inspection wavelength, changing the slope of the comparison voltage by the analog gain, truncating the upper data of the output data of the inspection wavelength, and extracting the difference data by truncating the upper data and the lower data. As a result, the electronic device can narrow the range of output data, thereby ensuring a sufficient output range of gradations within that range, and improving the accuracy of rank selection. Furthermore, the electronic device can reduce the amount of output data by extracting only information necessary in the sensor as output data. Furthermore, even if the electronic device generates a multispectral light source by dispersing light from one light source and adjusts the amount of light to a predetermined wavelength, the electronic device can improve the analysis accuracy at other wavelengths by narrowing the output range by extracting the output data. Furthermore, the electronic device can shorten the measurement time for rank selection by reducing the amount of output data.

[0023] In addition, in the third aspect, the rank selection method further includes an output intensity confirmation step of normalizing output data of the inspection wavelength based on a comparison between an expected value of a measurement value obtained when the inspection object is irradiated with a normalization processing wavelength different from the inspection wavelength and an actual value of a measurement value obtained when the inspection object is irradiated with the normalization processing wavelength. Thereby, the electronic device normalizes the output data by the normalization processing wavelength and then performs rank selection, thereby improving the accuracy of rank selection.

[0024] In the third aspect, the output intensity confirmation step modifies the output data of the inspection wavelength so that an actual measurement value obtained when the normalization processing wavelength is irradiated onto the inspection object coincides with the expected measurement value obtained when the normalization processing wavelength is irradiated onto the inspection object. This allows the electronic device to normalize the output data using the normalization processing wavelength and then perform rank selection, thereby improving the accuracy of rank selection.

[0025] In addition, in the third aspect, the rank selection method further includes an object identification step of identifying the type of the object to be inspected based on a first specific wavelength and a second specific wavelength, which are wavelengths at which a difference appears in output data depending on the type of the object when light is irradiated onto the object to be inspected. This allows the electronic device to identify the object and perform rank selection even when the type of the object to be inspected has not been identified.

[0026] In addition, in the third aspect, the object identifying step calculates a ratio between the first output data and the second output data generated by the reflected light of the light set as the first specific wavelength and the second specific wavelength, and compares the calculated ratio between the first output data and the second output data with the ratio between the first output data and the second output data for each type of object to identify the type of object. This allows the electronic device to identify the object and perform rank sorting even when the type of the object to be inspected has not been identified. [Brief description of the drawings]

[0027] [Figure 1] FIG. 1 is a schematic configuration diagram of an electronic device according to a first embodiment. [Diagram 2] FIG. 2 is a diagram illustrating multispectral data of the electronic device according to the first embodiment. [Diagram 3] 5A to 5C are diagrams illustrating clipping of output data of the electronic device in the first embodiment. [Figure 4] 2 is an example of a block diagram of a control device in the first embodiment. [Diagram 5] FIG. 2 is an example of a block diagram of a sensor according to the first embodiment. [Figure 6] 4A to 4C are diagrams illustrating a control method of a reference signal in the first embodiment. [Figure 7] 5 is an example of a flowchart for extracting an output range of a sensor in the first embodiment. [Figure 8] 5 is an example of a flowchart for classifying ranks of electronic devices in the first embodiment. [Figure 9] FIG. 11 is an example of a block diagram of a control device according to a second embodiment. [Figure 10] 13 is an example of a flowchart for classifying the ranks of electronic devices in the second embodiment. [Figure 11] FIG. 13 is a diagram illustrating feature extraction points of an electronic device according to a third embodiment. [Figure 12] 13 is an example of a flowchart for classifying the ranks of electronic devices in the third embodiment. [Figure 13] FIG. 13 is a diagram illustrating an example of a normalized processing wavelength in the fourth embodiment. [Figure 14] FIG. 13 is an example of a block diagram of a control device according to a fourth embodiment. [Figure 15] 13 is an example of a flowchart for classifying the ranks of electronic devices in the fourth embodiment. [Figure 16] FIG. 13 is an example of a block diagram of a control device according to a fifth embodiment. [Figure 17] 13 is an example of a flowchart for classifying the ranks of electronic devices in the fifth embodiment. [Figure 18] 13 shows an example of a hardware configuration of a control device 2 in the sixth embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0028] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings.

[0029] (First embodiment) FIG. 1 is a schematic configuration diagram of an electronic device according to a first embodiment.

[0030] In this embodiment, the electronic device 1 includes a control device 2, a sensor 3, and a light 4. The electronic device 1 performs a spectrum analysis of the inspection object 5 and performs rank selection. The control device 2 controls the light 4 to irradiate the inspection object 5 with light. The sensor 3 receives the light irradiated from the light 4 and reflected by the inspection object 5, performs a predetermined process within the sensor 3, and then transmits the result as pixel data to the control device 2. The control device 2 calculates a feature amount of the inspection object 5 from the pixel data transmitted from the sensor 3, and performs rank selection based on the feature amount. Details of each functional block of the electronic device 1 and a rank selection method will be described later.

[0031] FIG. 2 is a diagram illustrating multispectral data of the electronic device in the first embodiment.

[0032] The horizontal axis of FIG. 2 indicates the wavelength, and the vertical axis indicates the number of gradations. For example, the number of gradations may be 256, and the electronic device 1 can perform analysis with a precision according to the number of gradations. The electronic device 1 generates a light source of a predetermined wavelength from, for example, one light source 4 using a spectroscope or a bandpass filter. When irradiating the inspection target 5 with light, the electronic device 1 needs to reduce the output in order to reduce the effects of deterioration due to heat. However, adjusting the amount of light to one wavelength may result in the amount of light becoming too large or too small at another wavelength.

[0033] In the following, a wavelength at which a change appears in output data depending on the state of the inspection target 5 when light is irradiated onto the inspection target 5 is referred to as an inspection wavelength. In addition, in the following example, an example in which there are two inspection wavelengths will be described, and these will be referred to as a first inspection wavelength and a second inspection wavelength, respectively.

[0034] The solid line (first wavelength intensity) in Fig. 2 shows an example in which the output from the light source is small and the output of the second inspection wavelength is too small in the multispectral data. The dashed line (second wavelength intensity) in Fig. 1 shows an example in which the output from the light source is large and the output of the first inspection wavelength is too large in the multispectral data. Below, a method will be described in which the electronic device 1 extracts information required for inspection from the multispectral data in order to maintain inspection accuracy even when the output from the light source is not optimal, for example, even when the output is small.

[0035] FIG. 3 is a diagram illustrating clipping of output data of the electronic device in the first embodiment.

[0036] FIG. 3A shows an example in which the electronic device 1 does not cut out the output data, and shows multispectral data for an object to be inspected, such as meat, where the solid line shows the multispectral data when the object to be inspected 5 is in good condition, and the dashed line shows the multispectral data when the object to be inspected 5 is in poor condition. In this example, when the object to be inspected 5 is in good condition, the output is high at the first inspection wavelength and low at the second inspection wavelength. On the other hand, when the object to be inspected 5 is in poor condition, the output is low at the first inspection wavelength and low at the second inspection wavelength. In this way, the electronic device 1 is used to rank the object to be inspected 5 by utilizing the fact that the output differs at a specific wavelength depending on the condition of the object to be inspected 5.

[0037] The electronic device 1 can increase the output of the second inspection wavelength and increase the difference by increasing the intensity of the illumination 4, which is the light source. However, in the case of an electronic device 1 that generates the first inspection wavelength and 2 from the same light source, the output of the first inspection wavelength may become too large, causing excessive damage to the inspection target 5 or causing the output data to fall outside the detection range. On the other hand, when the output from the light source is small, the difference in output between a good condition and a bad condition is small relative to the number of gradations, and an error may occur in the inspection accuracy.

[0038] 3A, when rank sorting of the inspection object 5 is performed, the necessary portion of the multispectral data is a portion shown as a difference between the first inspection wavelength and the second inspection wavelength, where a difference in output is observed depending on the state of the inspection object 5. The electronic device 1 of this embodiment cuts out the output data of the inspection wavelength where a difference in output is observed.

[0039] 3B shows an example of a portion of output data cut out by the electronic device 1 in this embodiment. As described above, the electronic device 1 cuts out output data for a portion shown as a difference between the first inspection wavelength and the second inspection wavelength. If the type of object of the inspection target 5 is known in advance, the electronic device 1 may determine in advance the inspection wavelength for cutting out.

[0040] In this embodiment, the output range of the gradation before the output data is cut out is represented as (R). For the sake of explanation, the data cut out as a range indicating the feature amount is called differential data, and the upper area of ​​the differential data is called upper data. The lower area of ​​the differential data is called lower data. By cutting out the output data, the upper data and the lower data are truncated. In this example, the upper data at the first inspection wavelength is represented as U1, and the upper data at the second inspection wavelength is represented as U2. The lower data at the first inspection wavelength is represented as L1, and the lower data at the second inspection wavelength is represented as L2.

[0041] When no output data is cut out and the output range (R) of the gradation is 10 gradations, the output precision of the electronic device is R / 10. On the other hand, when output data is cut out at the first inspection wavelength, the output precision of the electronic device 1 in the differential data cut out by truncating the upper data and the lower data is (R-L1-U1) / 10, and at the second inspection wavelength, the output precision of the electronic device 1 is (R-L2-U2) / 10.

[0042] FIG. 4 is an example of a block diagram of the control device in the first embodiment.

[0043] The control device 2 can be realized, for example, by installing a program for the control device 2 in a PC (personal computer). A CPU (Central Processing Unit) in the control device 2 executes the program of the control device 2 to realize the functions of a condition setting unit 10, a light source control unit 11, a feature amount calculation unit 12, and a storage unit 13. The storage unit 13 is constructed, for example, on an auxiliary storage device on an HDD (Hard Disk Drive), and information stored during execution of the program is loaded into the memory.

[0044] The condition setting unit 10 sets the conditions of the light to be irradiated onto the inspection target 5 for the illumination 4 and the sensor 3. The condition setting unit 10 provides a user interface to the user via, for example, a display device (not shown), and the user may input the settings using an input device (not shown) such as a keyboard or a touch panel. The condition setting unit 10 performs various settings for the illumination 4 and the sensor 3 based on the settings input by the user. The condition setting unit 10 accepts input of, for example, light amount information which is information about the light amount of the illumination 4, exposure time information which is information about the exposure time, and wavelength setting information which is information about the wavelength of the illumination 4. The condition setting unit 10 also accepts input of information indicating the number of gradations of the output data, for example, bit precision, in addition to an offset voltage and an analog gain value, which will be described later.

[0045] The light source control unit 11 controls the illumination 4 based on the light amount information, exposure time information, and wavelength setting information received by the condition setting unit 10. In this way, the illumination 4 is controlled based on the conditions set by the user.

[0046] The feature amount calculation unit 12 calculates the feature amount of the inspection object 5 from the pixel data of the difference data generated based on the reflected light of the inspection object 5 acquired by the sensor 3. The feature amount calculated here is an amount necessary for rank selection of the inspection object 5, and may be a scalar amount or a vector amount. The amount acquired as the feature amount is set in advance by the user. For example, when it is desired to calculate the presence or absence of a scratch included in the inspection object 5 as a feature amount, the feature amount calculation unit 12 may calculate the number of pixels having a gradation value corresponding to the scratch included in the inspection object 5 in the pixel data.

[0047] Furthermore, the feature amount calculation unit 12 performs rank selection from the calculated feature amount. When performing rank selection, the feature amount calculation unit 12 may compare the feature amount with comparison data stored in the storage unit 13. For example, a threshold value may be used as the comparison data. The feature amount calculation unit 12 may perform rank selection from a correspondence relationship such that when a feature amount is equal to or greater than a certain threshold value, it belongs to a predetermined rank.

[0048] In this example, the feature amount calculation unit 12 is implemented as a function of the control device 2, but may be implemented inside the sensor 3. By performing feature amount calculation and rank selection inside the sensor 3, it is possible to speed up the processing.

[0049] FIG. 5 is an example of a block diagram of the sensor in the first embodiment.

[0050] The sensor 3 includes a plurality of pixels 20 (pixel array 21) arranged in a two-dimensional array. Hereinafter, a set of pixels 20 arranged in a horizontal direction will be referred to as a "row," and a set of pixels 20 arranged in a direction perpendicular to the rows will be referred to as a "column." The sensor 3 is controlled by a control circuit 40, which, for example, controls the output timing of pixel signals from each pixel 20 via a vertical scanning circuit 31.

[0051] In this example, the sensor 3 is an example of a column ADC in which an AD converter (ADC) is provided for each column of the pixel array 21, and each ADC includes a comparator 22 and a counter circuit .

[0052] In this example, each ADC is a single slope type ADC. The comparator 22 compares the reference signal, which is supplied from the RAMP DAC 25 and whose level changes at a constant slope, with the pixel signal supplied from the pixel 20, and supplies the comparison result to the counter circuit 23. Specifically, the comparison of these signals is performed by the comparator 22 by comparing the comparison voltage of the reference signal with the signal line voltage of the pixel signal. The signal output by the comparator 22 is, for example, 1-bit digital data. When the signal level of the reference signal is higher than the signal level of the pixel signal, the value output by the comparator 22 is "0", and conversely, when the signal level of the pixel signal is higher than the signal level of the reference signal, the value output by the comparator 22 is "1". Of course, the value of this signal may be taken in the opposite way. In addition, the bit length of the signal indicating the comparison result is arbitrary, and may be information consisting of multiple bits.

[0053] The counter circuit 23 has an input terminal connected to the output terminal of the comparator 22, and an output terminal connected to the latch circuit (LAT) 24. The counter circuit 23 also receives an input of a clock signal from the CLK control circuit 28. The counter circuit 23 is supplied with a comparison result from the comparator 22. The counter circuit 23 measures the time until the comparison result is inverted by counting the number of clocks of the clock signal from the start of counting until the comparison result of the comparator 22 is inverted (i.e., the signal level of the output terminal of the comparator 22 changes). Then, when the comparison result is inverted, the counter circuit 23 outputs the count value up to that point to the latch circuit 24 as the AD conversion result of the pixel signal input to one input terminal of the comparator 22 (i.e., digital data of the signal read from the pixel 20). The AD conversion result of the pixel signal is latched by the latch circuit 24, read out by the horizontal scanning circuit 30, amplified by the sense amplifier circuit 29, and then transferred to the data processing unit 50. The data processing unit 50 performs post-stage processing on the AD conversion result and outputs it to the control device 2.

[0054] The truncation of the upper data and the lower data described above is performed by the GAIN DAC26, the CLP DAC27, and the RAMP DAC25. The GAIN DAC26, the CLP DAC27, and the RAMP DAC25 may be collectively referred to as the sensor processing unit 60. The truncation of the upper data and the lower data will be described in detail below.

[0055] FIG. 6 is a diagram illustrating a method for controlling a reference signal in the first embodiment.

[0056] 6A to 6C, an example of a method for controlling a reference signal at a second inspection wavelength will be described. The description will be given assuming that the output range of the gradation before the output data is extracted is (R), the lower data is L2, and the upper data is U2.

[0057] FIG. 6A is an example for explaining the comparison between the reference signal and the pixel signal in the comparator 22 when no output data is clipped. As described above, the reference signal is input to the comparator 22, and the level of this reference signal changes with time at a constant slope. After the initial value of the comparison voltage in the reference signal is set, the comparator 22 compares the comparison voltage in the reference signal with the signal line voltage in the pixel signal. If the comparison result shows that the comparison voltage is greater than the signal line voltage, the comparator 22 reduces the voltage value from the comparison voltage by Vstep, which is the reduction in the voltage value for each comparison in the comparison voltage, and compares it with the signal line voltage. After this, the comparator 22 reduces the voltage value by Vstep until the comparison voltage becomes equal to the signal line voltage, and repeats the comparison.

[0058] FIG. 6B is an example for explaining a comparison between a reference signal and a pixel signal in the case of truncating the lower data. In this embodiment, when truncating the lower data, the sensor 3 adds an offset voltage to the comparison voltage (in this case, a negative offset voltage is added) after setting the initial value of the comparison voltage according to the control of the control circuit 40. A positive offset voltage may be subtracted from the comparison voltage. The offset voltage is determined for each inspection wavelength of the inspection object, and is determined based on a voltage value corresponding to an output range indicating a feature amount. For example, the offset voltage at a certain inspection wavelength may be determined as a voltage value corresponding to a lower limit value of output data obtained when the inspection object is in a good state. The offset voltage at a certain inspection wavelength may also be determined as a voltage value corresponding to a lower limit value of output data obtained when the inspection object is in a bad state. In addition, the sensor 3 may use a statistical method, such as using the average or variance of the lower limit value, in determining the offset voltage. FIG. 6B shows an example in which the lower data L2 is truncated from the output data because the offset voltage is added to the comparison voltage.

[0059] The CLP DAC27 is controlled by the control circuit 40 and generates an offset voltage. The generated offset voltage is added to the comparison voltage output from the RAMP DAC25, and this voltage value is supplied to the comparator 22 as the initial value of the reference signal. The CLP DAC27 may also be used in combination with a clamp function that removes dark current components in addition to truncating the lower data. In this case, the offset voltage and the clamp value are added to the comparison voltage (in this case, a negative offset voltage and a negative clamp value are added). A positive offset voltage and a positive clamp value may also be subtracted from the comparison voltage.

[0060] 6C is an example for explaining a comparison between a reference signal and a pixel signal when the upper data is truncated. In this embodiment, when truncating the upper data, the sensor 3 multiplies Vstep by an analog gain after truncating the lower data under the control of the control circuit 40, changes the slope of the comparison voltage in the reference signal, and adjusts the amount of truncation on the upper side. If the output range indicating the feature amount is D(V) and the analog gain is G (times), the range of the output data divided by the analog gain is R / G, and by setting an analog gain such that D=R / G, the sensor 3 can adjust the amount of truncation on the upper side under the control of the control circuit 40.

[0061] Furthermore, when the amount of gradation is reduced from N1 to N2 in order to reduce the amount of data, the analog gain is expressed as D=R / G×N2 / N1.

[0062] The GAIN DAC26 is controlled by the control circuit 40 to generate an analog gain. When the comparator 22 compares the comparison voltage in the reference signal with the signal line voltage in the pixel signal, if the comparison signal is greater than the signal line voltage, the comparator 22 reduces the voltage value by Vstep×G from the value of the comparison voltage and compares it with the signal line voltage. After this, the comparator 22 reduces the voltage value by Vstep×G until the value of the signal line voltage becomes equal to the comparison voltage, and repeats the comparison. The analog gain generated by the GAIN DAC26 is input to the RAMP DAC25, and the value obtained by subtracting the value of Vstep×G from the comparison voltage in the RAMP DAC25 is supplied to the comparator 22 as a reference signal. The example of FIG. 6C shows an example in which the amount of truncation of the upper data U2 is adjusted after the lower data is truncated by the operation described in FIG. 6B.

[0063] In addition, if the output range of the original gradation of the extracted output data is too wide, the number of gradations may be reduced to reduce the amount of data. For example, if the original gradation is 10 gradations, the electronic device 1 can reduce the amount of data by generating pixel data with 5 gradations for the extracted output data.

[0064] In addition, the sensor 3 cuts out the output data by controlling the input signal to the ADC. This allows the sensor 3 to redo sampling of the inspection target 5 in the form of an analog signal rather than a digital signal, making it possible to process the output data without reducing accuracy compared to cutting out the output data outside the sensor 3, for example, by the control device 2.

[0065] FIG. 7 is an example of a flowchart for extracting the output range of the sensor in the first embodiment.

[0066] For the sake of simplicity, this flowchart will be described taking as an example a case in which the comparison result is inverted at the timing when the comparison between the signal line voltage and the comparison voltage in the comparator 22 is performed twice.

[0067] In step S1, the RAMP DAC 25 generates a comparison voltage to be compared with a pixel signal under the control of the control circuit 40. In step S2, the control circuit 40 determines whether the comparison between the reference signal and the comparison signal is the first comparison.

[0068] If the control circuit 40 determines in step S2 that the comparison between the reference signal and the comparison signal is the first comparison, then in step S3, the RAMP DAC25 sets the initial value of the comparison voltage generated in step S1. This voltage may be a voltage that allows the maximum value of the signal line voltage to be measured. In step S4, the sensor 3 adds an offset voltage generated by the CLP DAC27 to the initial value of the comparison voltage output from the RAMP DAC25 under the control of the control circuit 40. This voltage value is also input to the comparator 22 as a reference signal. The sensor 3 can truncate the lower data by adding the offset voltage to the initial value of the comparison voltage.

[0069] In step S5, comparator 22 compares the signal line voltage generated by the charge photoelectrically converted in pixel 20 with the comparison voltage input as a reference signal, and determines whether or not the comparison voltage is greater than the signal line voltage. If comparator 22 determines in step S5 that the value of the comparison voltage is greater than the value of the signal line voltage (Yes in step S5), in step S6, counter circuit 23 performs counting processing under the control of control circuit 40, and counts up the counter value.

[0070] After step S6, the flow of steps S1 and S2 is executed again. Next, as the second comparison, the description will be given again from step S1. In step S1, the RAMP DAC25 generates a comparison voltage to be compared with the pixel signal under the control of the control circuit 40. The comparison voltage at this time is generated as the same value as the previous comparison voltage. Next, in step S2, the control circuit 40 determines that this is not the first comparison, and proceeds to step S7, where the GAIN DAC26 generates a voltage value of Vstep×G under the control of the control circuit 40 and supplies it to the RAMP DAC25. Also, the RAMP DAC25 generates a new comparison voltage by lowering the voltage value by Vstep×G from the generated comparison voltage under the control of the control circuit 40. Thereafter, each time step S7 is repeated, the comparison voltage is lowered by Vstep×G. Also, this voltage value is input to the comparator 22 as a reference signal.

[0071] Furthermore, in step S5, if comparator 22 determines that the comparison voltage value is equal to the signal line voltage value or that the signal line voltage value is greater than the comparison voltage value (No in step S5), in step S8, counter circuit 23 stops the counting process under the control of control circuit 40 and outputs the count value at that time to latch circuit 24. Furthermore, latch circuit 24 completes the AD conversion by latching the count value at that time as pixel data under the control of control circuit 40.

[0072] In step S9, the data processing unit 50 performs subsequent processing under the control of the control circuit 40. For example, in the case of the column ADC shown in Fig. 5, the pixel data is processed sequentially row by row. In step S10, the data processing unit 50 outputs the pixel data to the control device 2 row by row under the control of the control circuit 40.

[0073] FIG. 8 is an example of a flowchart for classifying the ranks of electronic devices in the first embodiment.

[0074] In this figure, the electronic device 1 will be described with respect to a rank selection flow in a stable environment where a fixed amount of reflected light is obtained from the inspection target 5. In other words, the description will be given assuming that light intensity measurement using an illuminometer or the like is not necessary. Also, for the sake of simplicity, the description will be given with respect to an example where rank selection is performed using an inspection wavelength of one wavelength.

[0075] In step S21, the condition setting unit 10 sets the measurement conditions for the inspection object 5. For example, the setting of the conditions includes, as the settings for the illumination 4, light amount information, which is information on the amount of light irradiated from the illumination 4, exposure time information, which is information on the exposure time of the light irradiated from the illumination 4 to the inspection object 5, and wavelength setting information, which is information on the setting of the inspection wavelength of the light irradiated from the illumination 4. As the settings for the sensor 3, the condition setting unit 10 accepts input of information indicating the number of gradations of the output data in addition to the offset voltage and analog gain value. For example, the user inputs the setting information to the condition setting unit via an input device such as a keyboard or a touch panel. In addition, since the offset voltage and analog gain value are determined to some extent depending on the type of the inspection object 5, when the type of the object of the inspection object 5 is known to the electronic device, the condition setting unit 10 may automatically set these values. In step S22, the light source control unit 11 controls the illumination 4 based on the conditions set by the condition setting unit 10 to irradiate the inspection object 5 with light. In this example, the illumination 4 irradiates the inspection object 5 with light having an inspection wavelength. Furthermore, the light reflected by the inspection object 5 is received by the pixel 20 of the sensor 3 .

[0076] In step S23, the sensor 3 adjusts the output range according to the control of the control circuit 40, and outputs pixel data to the control device 2. The upper and lower data of this pixel data are truncated, and a gradation value corresponding to the feature of the inspection object 5 is allocated from the entire output range of gradation in the range indicating the feature of the inspection object 5. For example, when the pixel data of the inspection object 5 is expressed as image data, the part in good condition or the part in bad condition indicating the feature is displayed in more detail according to the gradation range, compared to the case where the output data is not cropped. In step S24, the feature calculation unit 12 calculates the feature of the inspection object 5 from the obtained pixel data. The feature is calculated under conditions predetermined by the user, such as the number of pixels corresponding to a predetermined gradation value in the pixel data. The feature may also be calculated by AI (Artificial Intelligence). For example, it is considered that the user constructs a learning model so as to output the feature of the inspection object 5 by inputting pixel data of the inspection object 5, and realizes it as a functional block of the control device 2.

[0077] In step S25, the feature amount calculation unit 12 performs rank selection based on the feature amount obtained in step S24. The feature amount calculation unit 12 may perform rank selection, for example, by comparing the feature amount with information stored in the storage unit 13. It is considered that the feature amount calculation unit 12 performs rank selection based on a correspondence relationship such that when the feature amount is equal to or greater than a certain threshold, it belongs to a predetermined rank.

[0078] According to this embodiment, the electronic device 1 extracts an output range showing the characteristics of the inspection target 5 from within the sensor 3 and outputs the output range to the outside. By narrowing the range of the output data, a sufficient output range of gradations can be obtained within the range, and the accuracy of rank selection can be improved. Furthermore, the electronic device 1 can reduce the amount of output data by extracting only the necessary information from within the sensor 3 as output data.

[0079] Furthermore, according to this embodiment, even when the electronic device 1 generates a multispectral light source by dispersing light from one light source and adjusts the amount of light to a specific wavelength, the analysis accuracy at other wavelengths can be improved by narrowing the output range through clipping of the output data.

[0080] According to the present embodiment, the electronic device 1 can reduce the amount of output data, thereby shortening the measurement time for rank selection. When an AI is used for calculating features, the electronic device 1 can improve the accuracy of input data to the AI ​​and reduce the amount of data, thereby reducing the load when generating output data.

[0081] Second embodiment FIG. 9 is an example of a block diagram of a control device in the second embodiment.

[0082] 9, an example in which the electronic device 1 performs rank sorting using one inspection wavelength is shown, but in this embodiment, an example in which the electronic device 1 performs rank sorting using multiple wavelengths is described. The block diagram and the flow chart of the sensor 3 are the same as those in FIGS. 5 and 7, and therefore the description will be omitted.

[0083] In this embodiment, the electronic device 1 calculates the feature amount using one inspection wavelength, and then calculates the feature amount using a different inspection wavelength. The feature amount calculation is performed as many times as the number of inspection wavelengths. Therefore, after the feature amount calculation using one inspection wavelength is completed, the feature amount calculation unit 12 requests the condition setting unit 10 to change the setting.

[0084] FIG. 10 is an example of a flowchart for classifying the ranks of electronic devices in the second embodiment.

[0085] The flow from step S31 to S34 is the same as step S21 to S24, so the description will be omitted. In step S35, the feature calculation unit 12 judges whether the calculation of the feature has been completed with the settings of all the inspection wavelengths. In step S35, when the feature calculation unit 12 judges that the calculation of the feature has not been completed with the settings of all the inspection wavelengths (No in step S35), the feature calculation unit 12 requests the condition setting unit 10 to change the setting. The condition setting unit 10 that has received the setting change request in step S35 sets the condition again in step S31. At this time, the condition setting unit 10 sets the wavelength setting information different from the wavelength setting information set last time. In addition, the condition setting unit 10 may receive new input from the user for the condition setting, or may automatically set the condition based on the setting initially received from the user. For example, if the inspection target 5 has been determined, the setting conditions such as the inspection wavelength are determined to some extent, so that the condition setting unit 10 may set the measurement conditions according to the type of the object of the inspection target 5. The setting conditions may change the setting of the inspection wavelength and the number of inspections based on the previously obtained feature amount. The condition setting unit 10 may automatically set conditions for other settings such as the offset voltage and analog gain value in addition to the wavelength setting information.

[0086] In this manner, the electronic device 1 repeats the processes of steps S31 to S35 until the calculation of the feature amounts is completed using the settings for all the test wavelengths.

[0087] In step S35, when the feature calculation unit 12 determines that feature calculation is completed for all the inspection wavelengths (Yes in step S35), the feature calculation unit 12 performs rank selection of the inspection object 5 based on all the calculated feature amounts. In the rank selection, the feature calculation unit 12 may compare each feature amount by, for example, comparing with a plurality of threshold values ​​stored in the storage unit 13.

[0088] According to the present embodiment, the electronic device 1 can realize feature calculation using a plurality of inspection wavelengths. Moreover, the electronic device 1 can perform rank selection with higher accuracy by calculating feature amounts from a plurality of inspection wavelengths and performing rank selection.

[0089] Third embodiment FIG. 11 is a diagram for explaining feature extraction points of an electronic device in the third embodiment.

[0090] In this embodiment, the electronic device 1 divides the inspection object 5 into regions corresponding to one image data, and calculates features for each region using different inspection wavelengths. In this example, the data processing unit 50 switches the inspection wavelength of the lighting 4 under the control of the control circuit 40. The output from the sensor 3 is differential data generated by irradiating each region with light of different inspection wavelengths for pixel data of a region corresponding to one image data. In this example, the electronic device 1 divides a single inspection object 5 moving on a conveyor into regions using three different types of inspection wavelengths over time, and performs feature extraction. In this example, for the sake of explanation, the light sources having different inspection wavelengths are represented as lighting 4, 4', and 4'', from the left.

[0091] The block diagram of the electronic device 1 is the same as that in Fig. 9, and therefore the description thereof will be omitted. Also, the block diagram of the sensor 3 is the same as that in Fig. 5, and the flowchart is the same as that in Fig. 7, and therefore the description thereof will be omitted.

[0092] When the electronic device 1 divides the area, it is possible that the condition setting unit 10 sets conditions for the area. For example, the condition setting unit 10 may specify the area by coordinates. The condition setting unit 10 may also change the range detected by the sensor 3 for each divided area.

[0093] FIG. 12 is an example of a flowchart for classifying the ranks of electronic devices in the third embodiment.

[0094] In step S41, the condition setting unit 10 sets conditions related to region division as condition setting. The condition setting unit 10 may divide the region corresponding to one image data into as many regions as the number of inspection wavelengths. The condition setting related to this region division may be set by user input. The other steps from S41 to S42 are the same as steps S21 to S22, so the description will be omitted. In step S43, the data processing unit 50 changes the inspection wavelength of the illumination 4 under the control of the control circuit 40. The data processing unit 50 also generates pixel data based on the reflected light of the light irradiated by the illumination 4 to the inspection target 5. At this time, the sensor processing unit 60 also adjusts the output range under the control of the control circuit 40, and outputs it to the control device 2 as pixel data.

[0095] In step S44, the data processing unit 50 judges whether the output of pixel data is completed with the settings of all the inspection wavelengths. In step S44, if the data processing unit 50 judges that the output of pixel data is not completed with the settings of all the inspection wavelengths (No in step S44), the data processing unit 50 requests the condition setting unit 10 to change the setting. The condition setting unit 10 that has received the setting change request in step S44 sets conditions again in step S41. At this time, the condition setting unit 10 sets wavelength setting information different from the wavelength setting information that was previously set. For example, when the user sets six inspection wavelengths, the condition setting unit 10 may set the first to third inspection wavelengths as the first wavelength setting information, and set the fourth to sixth inspection wavelengths as the second wavelength setting information upon receiving the setting change request. In this way, the electronic device 1 repeats the processes of steps S41 to S44 until the input of pixel data to the feature calculation unit 12 is completed with the settings of all the inspection wavelengths.

[0096] In step S44, if the feature calculation unit 12 determines that the input of pixel data has been completed for all the inspection wavelengths (Yes in step S44), in step S45, the feature calculation unit 12 calculates features based on all the input pixel data. In step S46, the feature calculation unit 12 ranks the inspection object 5 based on the calculated features. In ranking, the feature calculation unit 12 may compare each feature with a plurality of thresholds stored in the storage unit 13, for example.

[0097] According to this embodiment, the electronic device 1 divides an area corresponding to one image data of one inspection target 5 into as many areas as the number of inspection wavelengths, and calculates the feature amount for each area. This allows the electronic device 1 to reduce the amount of data when performing rank selection.

[0098] Furthermore, according to this embodiment, the electronic device 1 can acquire data at high speed because the sensor 3 controls the change of the inspection wavelength of the light source.

[0099] (Fourth embodiment) FIG. 13 is a diagram showing an example of a normalization process wavelength in the fourth embodiment.

[0100] Since the feature calculation unit 12 wants to compare the expected value of the measurement value obtained when the inspection wavelength is irradiated onto the inspection object 5 with the actual value of the measurement value obtained when the inspection wavelength is irradiated onto the inspection object 5, if the amount of light, etc. varies and the measurement value fluctuates, the analysis cannot be performed correctly. For example, if a fixed amount of reflected light cannot be obtained from the inspection object 5 or the height of the object varies, and rank selection cannot be performed in a stable environment, the amount of light, etc. will vary, and therefore normalization of the output intensity of the illumination 4 will be required as a preprocessing step for calculating the feature amount.

[0101] The normalization of the output intensity of the illumination 4 is performed using a normalization processing wavelength. The normalization processing wavelength is a wavelength different from the inspection wavelength. For example, the normalization processing wavelength is a wavelength that causes little change in output data depending on the state of the inspection object 5. In the example of FIG. 13, the first inspection wavelength and the second inspection wavelength indicate wavelengths that cause changes in output data depending on the state of the inspection object 5, and the normalization processing wavelength uses a value that causes little change in output data that is a wavelength other than these. In this example, the normalization processing wavelength is selected from wavelengths that have a lower frequency than the first inspection wavelength and the second inspection wavelength. Since the normalization processing wavelength differs depending on the type of the inspection object 5, it is considered that a user or the like sets this value in advance and stores it in the storage unit 13.

[0102] In this embodiment, the normalization of the output intensity is performed by modifying the output data using the normalization processing wavelength so that the actual measured value obtained when the normalization processing wavelength is irradiated onto the inspection target 5 approaches or matches the expected measured value obtained when the normalization processing wavelength is irradiated onto the inspection target 5. For example, this modification can be performed by adjusting the exposure time or light amount of the illumination 4. Similarly, for the inspection wavelength, the output data can be normalized by adjusting the exposure time or light amount before irradiation with the illumination 4.

[0103] FIG. 14 is an example of a block diagram of a control device in the fourth embodiment.

[0104] In this embodiment, the control device 2 normalizes the output intensity of the light source as a normalization process before rank selection. Unlike the first embodiment, the control device 2 in this embodiment includes an output intensity confirmation unit .

[0105] The output intensity confirmation unit 14 receives as input the normalization processing output information obtained from the control device 2 and the sensor 3, performs normalization processing, and calculates normalization information. This normalization processing output information is, for example, an actual measurement value of the measurement value obtained when the normalization processing wavelength is irradiated onto the inspection target 5. For example, this actual measurement value may be a gradation value of the output data.

[0106] In the normalization process, the condition setting unit 10 sets a condition for a normalization process wavelength for performing the normalization process, and irradiates the inspection target 5 with light at this wavelength. The output intensity confirmation unit 14 confirms the output data generated based on the reflected light of the inspection target 5 acquired by the sensor 3. The output intensity confirmation unit 14 compares the expected value of the measurement value obtained when the inspection target 5 is irradiated with the normalization process wavelength with the actual measured value obtained when the inspection target 5 is actually irradiated with the normalization process wavelength. This expected value is, for example, stored in the storage unit 13 in advance, and loaded into the memory when the program of the control device 2 is executed. When there is a difference between the expected value and the measured value as a result of comparing these values, the output intensity confirmation unit 14 transforms and normalizes the output data so that the measured value approaches or matches the expected value. The output intensity confirmation unit 14 outputs information regarding the exposure time or light amount used in the normalization to the condition setting unit 10 as normalization information. This normalization information is used as the condition setting of the condition setting unit 10.

[0107] After the normalization process, the condition setting unit 10 calculates the feature amount and performs rank selection according to the inspection wavelength. The feature amount calculation unit 12 calculates the feature amount based on the normalized output data. That is, the electronic device 1 performs the normalization process when acquiring pixel data for the first time, and calculates the feature amount when acquiring pixel data for the second time. The normalization process may be performed by acquiring pixel data multiple times. The accuracy of normalization can be improved by changing the conditions for each pixel data to be acquired. The electronic device 1 can perform rank selection without reducing the throughput by operating the sensor 3 at high speed.

[0108] Furthermore, the comparison between the expected value and the actually measured value is not performed using one normalization wavelength, but using a plurality of normalization wavelengths, thereby enabling normalization with higher accuracy.

[0109] Furthermore, the normalization process can be realized not by adjusting the exposure time or the amount of light, but by adjusting the cut-out range of the output data or the threshold value of the rank selection. For example, when the output data is output as a value 1.1 times the expected value of the measured value, the normalization process can be performed by cutting out the lower limit or upper limit of the cut-out range by 1.1 times. Similarly, when the output data is output as a value 1.1 times the expected value of the measured value, the normalization process can be performed by setting the threshold value in the rank selection to 1.1 times. In this way, when the normalization process is performed by adjusting the cut-out range or the threshold value of the rank selection, the electronic device 1 can complete the normalization process and the rank selection in one measurement, and the throughput can be improved.

[0110] FIG. 15 is an example of a flowchart for classifying the ranks of electronic devices in the fourth embodiment.

[0111] 8, 10 or 12, the rank selection flow in step S56 will be mainly described with reference to steps S51 to S55, which are the normalization process. In this flow, the type of the object 5 to be inspected will be specified by the user.

[0112] In step S51, the condition setting unit 10 sets a normalized processing wavelength as the wavelength setting information. The condition setting unit 10 may, for example, accept an input of the type of the inspection object 5 by a user and set a normalized processing wavelength corresponding to the type of the inspection object 5. In step S52, the light source control unit 11 controls the illumination 4 based on the conditions set by the condition setting unit 10 to irradiate the inspection object 5 with light. In this example, the illumination 4 irradiates the inspection object 5 with light having a normalized processing wavelength. In addition, the reflected light reflected by the inspection object 5 is received by the pixel 20 of the sensor 3.

[0113] In step S53, the sensor 3 outputs pixel data to the control device 2 under the control of the control circuit 40. In the normalization process, the sensor 3 does not adjust the output range. In step S54, the output intensity confirmation unit 14 confirms the obtained output data and calculates normalization information of the inspection target 5. The normalization information is calculated, for example, based on a comparison between an expected value of a measurement value obtained when the inspection target 5 is irradiated with the normalization processing wavelength and an actual value of a measurement value obtained when the inspection target 5 is irradiated with the normalization processing wavelength. For example, the normalization information is calculated as an exposure time or a light amount value required for normalization. In addition, the output intensity confirmation unit 14 outputs the calculated normalization information to the condition setting unit 10.

[0114] In step S55, the condition setting unit 10 sets conditions based on the normalization information received as input from the output intensity confirmation unit 14. In this example, the normalization information is calculated as the exposure time and light amount values ​​required for normalization, and these conditions are set during the rank selection flow. In step S56, the condition setting unit 10 sets the inspection wavelength as the wavelength setting information, and performs feature calculation and rank selection. Although step S56 is shown as one flow, for example, rank selection is performed by the same flow as steps S21 to S25. The flow of step S56 may be performed inside the sensor 3, not in the electronic device 1 as a whole.

[0115] According to this embodiment, the electronic device 1 performs normalization processing of the output data using the normalization processing wavelength and then performs rank selection, so that the accuracy of rank selection can be improved.

[0116] Fifth embodiment FIG. 16 is an example of a block diagram of a control device in the fifth embodiment.

[0117] In this embodiment, the electronic device 1 identifies the object type of the inspection target 5 before rank sorting. Unlike the fourth embodiment, the control device 2 in this embodiment includes an object identification unit 15. In identifying the object type of the target, the object type of the inspection target 5 is identified, for example, as meat or vegetables.

[0118] In this example, the object identification unit 15 is implemented as a function of the control device 2, but may be implemented inside the sensor 3. Identifying the type of object inside the sensor 3 can speed up the processing.

[0119] The object identification unit 15 receives the identification information obtained from the control device 2 and the sensor 3 as an input, identifies the type of the object, and calculates object information that is information about the type of object of the inspection object 5.

[0120] In the identification process, the condition setting unit 10 sets conditions for a specific wavelength for performing the identification process, and irradiates the inspection object 5 with light at this wavelength. The specific wavelength is selected from a wavelength at which a difference appears in the output data depending on the type of object when the inspection object 5 is irradiated with light. This specific wavelength may be multiple wavelengths. The object identification unit 15 identifies the object from pixel data generated based on the reflected light of the inspection object 5 acquired by the sensor 3 as identification information. The identification process may be performed using one type of wavelength, or multiple wavelengths as necessary.

[0121] Since the specific wavelength is a wavelength that has not been subjected to normalization processing, when output data obtained when the inspection target 5 is irradiated with light of the specific wavelength is compared with output data obtained when light of the specific wavelength is irradiated for each type of object, the accuracy of identifying the type of object of the inspection target may decrease due to variation in output intensity. Therefore, the type of object of the inspection target may be identified using output data obtained by one type of specific wavelength, or may be identified using, for example, a ratio of output data obtained when light of multiple specific wavelengths is irradiated to the inspection target 5. By identifying the type of object of the inspection target using the ratio of multiple output data, it is possible to suppress the decrease in accuracy associated with variation in output intensity.

[0122] For example, a case will be described in which the object identification unit 15 uses three wavelengths, specific wavelength A, specific wavelength B, and specific wavelength C (hereinafter simply referred to as A, B, and C), as the specific wavelengths. When the ratio between the output data of A and the output data of B is A / B, and further, the ratio between the output data of B and the output data of C is B / C, the object identification unit 15 can identify an object having a similar ratio as the type of object of the inspection target 5. Data on the ratios of these objects may be stored in advance in the storage unit 13 for each type of object, for example, and the object identification unit 15 may identify the object by referring to this data.

[0123] Moreover, the number of specific wavelengths is not limited to three. The object identification unit 15 may be configured to calculate a ratio (first value) between the first output data and the second output data generated by the first specific wavelength and the second specific wavelength that are different from each other. The object identification unit 15 may be configured to compare the calculated ratio with a ratio (second value) between the first output data and the second output data for each type of object, and identify the type of object.

[0124] The type of the object may be identified by AI. For example, the AI ​​may construct a learning model to identify the type of the object 5 by receiving pixel data of the object 5 and output the identification result, and may be realized as a functional block of the control device 2.

[0125] After the object identification unit 15 identifies the type of the inspection object 5, the condition setting unit 10 selects conditions related to the output intensity of the illumination 4 required for the inspection object 5, and the exposure time and light amount corresponding to the output intensity, based on the object information. This information may be stored in advance as a database in the storage unit 13, for example. The database may be configured so that the required output intensity, exposure time, and light amount of the illumination 4 are stored in association with each type of inspection object 5.

[0126] Furthermore, after identifying the inspection target object 5, the output intensity checking unit 14 may normalize the output intensity of the illumination 4. As in the fourth embodiment, the output intensity checking unit 14 receives as input normalization process output information obtained by the normalization process performed by the control device 2 and the sensor 3, checks the output intensity, and calculates normalization information. When performing rank selection, the condition setting unit 10 normalizes the output data of the illumination 4 based on the normalization information.

[0127] After the identification process and normalization process, the condition setting unit 10 performs feature calculation and rank selection according to the inspection wavelength.

[0128] FIG. 17 is an example of a flowchart for classifying the ranks of electronic devices in the fifth embodiment.

[0129] Steps S61 to S64 show an example of a flowchart for the identification process, and step S65 shows an example of a flowchart for the normalization process.

[0130] The rank selection flow in step S65 is similar to the flow described in FIG. 8, 10 or 12, and the normalization information calculation flow in step S65 is similar to the flow described in S51 to S54. Therefore, the following description will focus mainly on steps S61 to S64, which are the specific processing part.

[0131] In step S61, the condition setting unit 10 sets a condition for a specific wavelength as wavelength setting information. If there are multiple specific wavelengths, one of them is set as the specific wavelength. In step S62, the light source control unit 11 controls the illumination 4 based on the condition set by the condition setting unit 10 to irradiate the inspection object 5 with light. In this example, the illumination 4 irradiates the inspection object 5 with light having a specific processing wavelength. Also, the reflected light reflected by the inspection object 5 is received by the pixel 20 of the sensor 3.

[0132] In step S63, the sensor 3 outputs the pixel data to the control device 2 under the control of the control circuit 40. In this example, the sensor 3 does not adjust the output range in the identification process. In step S64, the object identification unit 15 identifies the object from the obtained pixel data. The type of the object to be inspected is identified, for example, by irradiating the object to be inspected 5 with light of multiple specific wavelengths and calculating a ratio from the multiple output data obtained. The object identification unit 15 uses the ratio of the output data obtained by the multiple specific wavelengths to identify objects having a similar ratio as the object type of the object to be inspected 5.

[0133] In step S65, the control device 2 and the sensor 3 calculate normalization information based on the normalization processing wavelength. The calculation of the normalization information is performed in the same flow as the flow described in S51 to S54. In step S66, the condition setting unit 10 sets conditions related to the inspection wavelength based on the normalization information received from the output intensity confirmation unit 14. In this example, the normalization information is calculated as the exposure time and light amount values ​​required for normalization, and the condition setting unit 10 sets these conditions during the rank selection flow. In addition, the condition setting unit 10 sets an inspection wavelength according to the type of the specified inspection object 5 as wavelength setting information. The condition setting unit 10 sets the inspection wavelength as the wavelength setting information, and performs feature calculation and rank selection.

[0134] Furthermore, the electronic device 1 may perform rank selection using a plurality of wavelengths. By calculating the feature amount using a plurality of wavelengths, the electronic device 1 can perform rank selection with higher accuracy.

[0135] Furthermore, even when different objects to be inspected 5 are mixed, the electronic device 1 may use a plurality of specific wavelengths to identify the object type of the object to be inspected 5, respectively. Furthermore, for example, the condition setting unit 10 may automatically set specific wavelengths and perform identification processing for each specific wavelength, so that the electronic device 1 may automatically identify the object type of the object to be inspected 5, normalize the output intensity of the illumination 4, and select the rank.

[0136] According to this embodiment, even if the type of object of the inspection target 5 is not specified, the electronic device 1 can specify the object and perform rank sorting.

[0137] Furthermore, according to this embodiment, the electronic device 1 automatically sets conditions for multiple specific wavelengths and automatically performs specific processing using each specific wavelength, thereby enabling automatic rank selection without human intervention.

[0138] Sixth embodiment FIG. 18 shows an example of a hardware configuration of the control device 2 in the sixth embodiment.

[0139] The control device 2 according to this embodiment includes a CPU (Central Processing Unit) 101, a main memory device 102, an auxiliary memory device 103, a communication interface 104, and an input / output interface 105, which are interconnected by a bus .

[0140] The CPU 101 executes a computer program that realizes the above-mentioned functional configuration of the control device 2 on the main memory device 102. The CPU 101 executes the computer program to realize the functions of the condition setting unit 10, the light source control unit 11, the feature calculation unit 12, and the storage unit 13 in FIG.

[0141] The main memory device 102 stores the programs that realize the processing of this embodiment and data required for executing the programs, as well as data generated by executing the programs. The programs are deployed and executed on the main memory device 102. The main memory device 102 is, for example, a RAM (Random Access Memory), but is not limited to this.

[0142] The auxiliary storage device 103 stores the above programs, data required for executing the programs, and data generated by executing the programs. These programs and data are read into the main storage device 102 during processing of this embodiment. The auxiliary storage device 103 is, for example, a hard disk, an optical disk, a flash memory, or a magnetic tape, but is not limited to these. The storage unit 13 in FIG. 4 may be constructed on the auxiliary storage device 103.

[0143] The communication interface 104 is a circuit for communicating with the sensor 3 wirelessly or via wires.

[0144] The input / output interface 105 is a circuit for connecting to an input device such as a keyboard or a touch panel, which are examples of an input device, and an output device such as a display device.

[0145] The bus 106 is a circuit for connecting the CPU 101, the main memory device 102, the auxiliary memory device 103, the communication interface 104, and the input / output interface 105 to one another.

[0146] The above-mentioned program may be pre-installed in the control device 2, or may be stored in a storage medium such as a CD-ROM. Also, the program may be uploaded onto the Internet.

[0147] The control device 2 may include one or more of each of the CPU 101, the main storage device 102, the auxiliary storage device 103, the communication interface 104, and the input / output interface 105.

[0148] Furthermore, the control device 2 may be configured as a single computer device, or may be configured as a system made up of a plurality of computer devices connected to each other.

[0149] According to this configuration, the functions of the control device 2 in the first embodiment can be realized by software. Also, while Fig. 18 shows an example of the hardware configuration of the control device 2 in Fig. 4, the functions of the control device 2 can be realized by software in other embodiments with a similar configuration.

[0150] Although the embodiments of the present disclosure have been described above, these embodiments may be implemented with various modifications without departing from the scope of the present disclosure. For example, two or more embodiments may be implemented in combination.

[0151] The present disclosure may also be configured as follows.

[0152] (1) a condition setting unit that sets conditions for irradiating the object with light; a sensor processing unit that extracts an output range of output data of the inspection wavelength generated by reflected light of the light set as the inspection wavelength, and generates difference data; a feature amount calculation unit that calculates a feature amount of the inspection object from pixel data of the difference data and ranks the inspection object, the difference data is extracted as a range indicating a feature amount of a state of the inspection object; electronic equipment.

[0153] (2) The sensor processing unit includes: a sensor that outputs pixel data of the difference data based on the reflected light of the light; A RAMP DAC generates a reference signal whose level changes at a constant rate over time. A CLP DAC that generates an offset voltage of the reference signal; and a GAIN DAC for generating analog gain. (1) An electronic device as described in (1).

[0154] (3) The electronic device according to (1), wherein the feature calculation unit is provided inside a sensor that outputs pixel data of the difference data based on reflected light of the light.

[0155] (4) The sensor processing unit includes: adding or subtracting the offset voltage to or from a comparison voltage in the reference signal, and truncating lower data of the output data of the inspection wavelength; The analog gain is used to change the slope of the comparison voltage, and the upper data of the output data of the inspection wavelength is discarded. extracting the difference data by truncating the upper data and the lower data; (2) An electronic device as described in (1).

[0156] (5) The electronic device according to (2), wherein the sensor processing unit is controlled by a control circuit provided inside the sensor.

[0157] (6) a condition setting unit that sets conditions for irradiating the object with light; a sensor processing unit that extracts an output range of output data of each of the inspection wavelengths generated by reflected light of the light set as a plurality of inspection wavelengths, and generates respective difference data; a feature amount calculation unit that calculates a feature amount of the inspection object from pixel data of each of the difference data and ranks the inspection object, Each of the difference data is extracted as a range indicating a feature amount of the state of the inspection object. electronic equipment.

[0158] (7) The condition setting unit performs area division for an area corresponding to one piece of image data, The sensor processing unit generates the difference data by the different inspection wavelengths for each of the divided regions. (6) An electronic device as described in (6).

[0159] (8) The electronic device according to (7), wherein the inspection wavelength is changed under the control of a control circuit provided inside the sensor.

[0160] (9) The electronic device described in (1), further comprising an output intensity confirmation unit that normalizes output data of the inspection wavelength based on a comparison between an actual measurement value obtained when the inspection object is irradiated with a normalization processing wavelength different from the inspection wavelength and an expected measurement value obtained when the inspection object is irradiated with the normalization processing wavelength.

[0161] (10) The electronic device described in (9), wherein the output intensity confirmation unit modifies the output data of the inspection wavelength so that an actual measured value of a measurement value obtained when the normalization processing wavelength is irradiated onto the inspection object matches the expected value of a measurement value obtained when the normalization processing wavelength is irradiated onto the inspection object.

[0162] (11) The electronic device described in (1), further comprising an object identification unit that identifies the type of object of the inspection object based on a first specific wavelength and a second specific wavelength, which are wavelengths at which a difference in output data appears depending on the type of object when light is irradiated onto the inspection object.

[0163] (12) The object identification unit is Calculating a ratio between first output data and second output data generated by reflected light of the light set as the first specific wavelength and the second specific wavelength; comparing the calculated ratio between the first output data and the second output data with a ratio between the first output data and the second output data for each type of object, thereby identifying the type of object; (11) An electronic device as described in (11).

[0164] (13) The electronic device according to (11), wherein the object identification unit is provided inside a sensor that outputs pixel data of the difference data based on the reflected light of the light.

[0165] (14) a condition setting step for setting conditions for light to be irradiated onto the inspection object; a sensor processing step of extracting an output range of output data of the inspection wavelength generated by reflected light of the light set as the inspection wavelength, and generating difference data; a feature amount calculation step of calculating a feature amount of the inspection object from pixel data of the difference data and ranking the inspection object; the difference data is extracted as a range indicating a feature amount of a state of the inspection object; Rank selection method.

[0166] (15) The sensor processing step includes: A reference signal is generated whose level changes at a constant rate over time. generating an offset voltage of the reference signal; generating an analog gain; (14) A rank selection method according to (14).

[0167] (16) The sensor processing step includes: adding or subtracting the offset voltage to or from a comparison voltage in the reference signal, and truncating lower data of the output data of the inspection wavelength; The analog gain is used to change the slope of the comparison voltage, and the upper data of the output data of the inspection wavelength is discarded. further comprising the step of extracting the difference data by truncating the upper data and the lower data. (15) A rank selection method according to (15).

[0168] (17) The rank selection method according to (15), further comprising an output intensity confirmation step of normalizing output data of the inspection wavelength based on a comparison between an expected value of a measurement value obtained when the inspection object is irradiated with a normalization processing wavelength different from the inspection wavelength and an actual value of a measurement value obtained when the inspection object is irradiated with the normalization processing wavelength.

[0169] (18) The rank selection method according to (17), wherein the output intensity confirmation step modifies output data of the inspection wavelength so that an actual measured value of a measurement value obtained when the normalization processing wavelength is irradiated onto the inspection object coincides with the expected value of a measurement value obtained when the normalization processing wavelength is irradiated onto the inspection object.

[0170] (19) The ranking method according to (15), further comprising an object identification step of identifying the type of the object to be inspected based on a first specific wavelength and a second specific wavelength, which are wavelengths at which a difference appears in output data depending on the type of object when light is irradiated onto the object to be inspected.

[0171] (20) The object identification step includes: Calculating a ratio between first output data and second output data generated by reflected light of the light set as the first specific wavelength and the second specific wavelength; comparing the calculated ratio between the first output data and the second output data with a ratio between the first output data and the second output data for each type of object, thereby identifying the type of object; (19) A rank selection method according to (19). [Explanation of symbols]

[0172] 1: Electronic device, 2: Control device, 3: Sensor, 4: Lighting, 4': Lighting, 4'': Lighting, 5: inspection object, 10: condition setting unit, 11: light source control unit, 12: feature amount calculation unit, 13: memory unit, 14: output intensity confirmation unit, 15: object identification unit, 20: pixel, 21: pixel array, 22: comparator, 23: counter circuit, 24: latch circuit, 25: RAMP DAC, 26: GAIN DAC, 27: CLP DAC, 28: CLK control circuit, 29: sense amplifier circuit, 30: horizontal scanning circuit, 31: vertical scanning circuit, 40: control circuit, 50: data processing unit, 60: sensor processing unit, 101: CPU, 102: Main storage, 103: Auxiliary storage, 104: communication interface, 105: input / output interface

Claims

1. a condition setting unit that sets conditions for irradiating the object with light; a sensor processing unit that extracts an output range of output data of the inspection wavelength generated by reflected light of the light set as the inspection wavelength, and generates difference data; a feature amount calculation unit that calculates a feature amount of the inspection object from pixel data of the difference data and ranks the inspection object, the difference data is extracted as a range indicating a feature amount of a state of the inspection object; electronic equipment.

2. The sensor processing unit includes: a sensor that outputs pixel data of the difference data based on the reflected light of the light; A RAMP DAC that generates a reference signal whose level changes with a constant slope over time; a CLP DAC for generating an offset voltage of the reference signal; and a GAIN DAC for generating analog gain.

2. The electronic device according to claim 1.

3. The electronic device according to claim 1 , wherein the feature amount calculation section is provided inside a sensor that outputs pixel data of the difference data based on the reflected light of the light.

4. The sensor processing unit includes: adding or subtracting the offset voltage to or from a comparison voltage in the reference signal, and truncating lower data of the output data of the inspection wavelength; The analog gain is used to change the slope of the comparison voltage, and the upper data of the output data of the inspection wavelength is discarded. extracting the difference data by truncating the upper data and the lower data; 3. The electronic device according to claim 2.

5. The electronic device according to claim 2 , wherein the sensor processing unit is controlled by a control circuit provided inside the sensor.

6. a condition setting unit that sets conditions for irradiating the object with light; a sensor processing unit that extracts an output range of output data of each of the inspection wavelengths generated by reflected light of the light set as a plurality of inspection wavelengths, and generates respective difference data; a feature amount calculation unit that calculates a feature amount of the inspection object from pixel data of each of the difference data and ranks the inspection object, Each of the difference data is extracted as a range indicating a feature amount of the state of the inspection object. electronic equipment.

7. The condition setting unit performs area division for an area corresponding to one piece of image data, The sensor processing unit generates the difference data by the different inspection wavelengths for each of the divided regions.

7. The electronic device according to claim 6.

8. The electronic device according to claim 7 , wherein the inspection wavelength is changed under the control of a control circuit provided inside the sensor.

9. 2. The electronic device according to claim 1, further comprising an output intensity confirmation unit that normalizes output data of the inspection wavelength based on a comparison between an actual measurement value obtained when the inspection object is irradiated with a normalization processing wavelength different from the inspection wavelength and an expected measurement value obtained when the inspection object is irradiated with the normalization processing wavelength.

10. 10. The electronic device according to claim 9, wherein the output intensity confirmation unit modifies the output data of the inspection wavelength so that an actual measurement value obtained when the normalization processing wavelength is irradiated onto the inspection object matches the expected measurement value obtained when the normalization processing wavelength is irradiated onto the inspection object.

11. 2. The electronic device according to claim 1, further comprising an object identification unit that identifies a type of the object to be inspected based on a first specific wavelength and a second specific wavelength, which are wavelengths at which a difference appears in output data depending on the type of the object when light is irradiated onto the object to be inspected.

12. The object identification unit is Calculating a ratio between first output data and second output data generated by reflected light of the light set as the first specific wavelength and the second specific wavelength; comparing the calculated ratio between the first output data and the second output data with a ratio between the first output data and the second output data for each type of object, thereby identifying the type of object; The electronic device according to claim 11.

13. The electronic device according to claim 11 , wherein the object identification unit is provided inside a sensor that outputs pixel data of the difference data based on the reflected light of the light.

14. a condition setting step for setting conditions for light to be irradiated onto the inspection object; a sensor processing step of extracting an output range of output data of the inspection wavelength generated by reflected light of the light set as the inspection wavelength, and generating difference data; a feature amount calculation step of calculating a feature amount of the inspection object from pixel data of the difference data and ranking the inspection object; the difference data is extracted as a range indicating a feature amount of a state of the inspection object; Rank selection method.

15. The sensor processing step includes: A reference signal is generated whose level changes at a constant rate over time. generating an offset voltage of the reference signal; generating an analog gain; The rank selection method according to claim 14.

16. The sensor processing step includes: adding or subtracting the offset voltage to or from a comparison voltage in the reference signal, and truncating lower data of the output data of the inspection wavelength; The analog gain is used to change the slope of the comparison voltage, and the upper data of the output data of the inspection wavelength is discarded. further comprising the step of extracting the difference data by truncating the upper data and the lower data. The rank selection method according to claim 15.

17. 16. The rank selection method according to claim 15, further comprising an output intensity confirmation step of normalizing output data of the inspection wavelength based on a comparison between an expected value of a measurement value obtained when the inspection object is irradiated with a normalization processing wavelength different from the inspection wavelength and an actual value of a measurement value obtained when the inspection object is irradiated with the normalization processing wavelength.

18. 18. The rank selection method according to claim 17, wherein the output intensity confirmation step modifies output data of the inspection wavelength so that an actual measured value of a measurement value obtained when the inspection object is irradiated with the normalization processing wavelength coincides with the expected value of a measurement value obtained when the inspection object is irradiated with the normalization processing wavelength.

19. 16. The ranking method according to claim 15, further comprising an object identification step of identifying the type of the object to be inspected based on a first specific wavelength and a second specific wavelength, which are wavelengths at which a difference in output data appears depending on the type of object when light is irradiated onto the object to be inspected.

20. The object identification step includes: Calculating a ratio between first output data and second output data generated by reflected light of the light set as the first specific wavelength and the second specific wavelength; comparing the calculated ratio between the first output data and the second output data with a ratio between the first output data and the second output data for each type of object, thereby identifying the type of object; 20. The rank selection method according to claim 19.

Images