JP5273667B2 - Object identification device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To discern various objects such as persons and cars from the background with a high degree of accuracy. <P>SOLUTION: An object discerning device 100 includes: a light-receiving section 120 for, among reflected light from light with which a predetermined measurement space is irradiated, receiving the reflected light of at least three different wavelengths including an infrared region; a reflectance calculating section 130 for calculating a reflectance for each of the plurality of wavelengths; a normalization index calculating section 150 for calculating a normalization index in accordance with the reflectance calculated for each of the plurality of wavelengths; and a discerning section 170 for, in accordance with the normalization index and a preset discerning condition, discerning a plurality of different objects present in the measurement space. <P>COPYRIGHT: (C)2010,JPO&amp;INPIT

Description

  The present invention relates to an object identification device, and more particularly to an object identification device that identifies an object based on the reflectance of reflected light with respect to light irradiated on the object.

  As a method for optically detecting a specific target such as an obstacle, a method is known in which an image including the target is captured with visible light, and a person, a vehicle, or the like is discriminated from the image based on the shape. Further, as a method that can be used at night, a method of identifying a person from a background using infrared thermography is known. In addition, a method is known in which infrared light having two different wavelengths is irradiated onto a target region and the target and the background are identified using the reflectance (see, for example, Patent Document 1).

  An apparatus using these methods is expected to be applied as, for example, driving assistance for an electric wheelchair or an assisting tool for a person with low vision to recognize an obstacle while walking.

JP2008-275477

  However, the method of discriminating an object based on an image captured with visible light may cause a problem from the viewpoint of protecting privacy and protecting public morals when used in a city. Also, with this method, in order to ensure practical identification accuracy, a large load is applied to the processing system, so that it has been difficult to mount on a vehicle that moves at high speed. In addition, the method of identifying a person using infrared thermography has extremely large absorption by water in the thermal infrared region, so that the accuracy of identification is extremely reduced in rainy weather or when the environmental temperature is close to the surface temperature of the human body. There is a problem.

  On the other hand, the method using the reflectance of infrared light hardly reduces the identification accuracy due to the influence of the measurement environment as described above, and the load on the processing system is smaller than other methods. In addition, problems are less likely to occur from the viewpoint of protecting privacy and protecting public morals. However, for example, human skin can be selectively extracted from the background, and various objects cannot be identified.

  The present invention is an object identification device that solves the above-described problems, and receives reflected light of a plurality of different wavelengths including at least three of infrared light including reflected light with respect to light irradiated to a predetermined measurement space. A light receiving unit, a reflectance calculating unit that calculates the reflectance for each of the plurality of wavelengths, a normalization index calculating unit that calculates a normalization index based on the reflectance for each of the plurality of wavelengths, And an identification unit that identifies a plurality of different objects existing in the measurement space based on the normalization index and preset identification conditions.

  Further, in the object identification device of the present invention, the reflectance calculation unit calculates a plurality of reflectances for each of the plurality of wavelengths in time series, and the normalization index calculation unit calculates the plurality of the plurality of wavelengths calculated simultaneously. It is preferable that the normalization index is calculated based on the reflectance for each wavelength, and the identification unit identifies the target sequentially in time series based on the normalization index and the identification condition.

  The object identification device of the present invention further includes a secondary differential value calculation unit that calculates a secondary differential value of a function of the reflectance and the wavelength, and the identification unit includes the normalization index and the secondary index The target may be identified based on the differential value and a preset identification condition.

  In the object identification device of the present invention, the normalization index calculation unit calculates a plurality of normalization indexes based on the reflectance of each wavelength, and the secondary differential value calculation unit includes a plurality of the second derivative calculation units It is preferable to approximately calculate the secondary differential value based on a normalization index.

  In the object identification device of the present invention, it is preferable that the identification unit identifies the object by comparing the normalization index and the secondary differential value with a preset threshold value.

  In addition, the identification unit, when the identification result obtained by comparing the normalized index calculated for the previously specified known target and the threshold does not identify the known target, The average value may be a new threshold value.

  In the object identification device of the present invention, it is preferable that the plurality of wavelengths include a red region.

  Moreover, the object identification device of the present invention may further include an irradiation unit that irradiates the measurement space with light including a plurality of wavelengths.

  The above summary of the invention does not enumerate all the necessary features of the present invention, and sub-combinations of these feature groups can also be the invention.

  By providing such a feature, the present invention can identify various objects such as people and cars from the background with high accuracy. In addition, there is almost no decrease in identification accuracy due to the influence of the measurement environment, and the load on the processing system is small. Furthermore, no problem arises from the viewpoint of protecting privacy or protecting public morals when identifying objects.

The functional block diagram of the object identification apparatus 100 which concerns on embodiment of this invention is shown. The identification flow at the time of the identification part 170 identifying a several different object is shown. The identification flow at the time of the identification part 170 identifying a several different object is shown. In the object identification device 100 using the red area, a part of an identification flow when the identification unit 170 identifies a plurality of different objects is shown. The identification flow at the time of the identification part 170 identifying a several different object in the object identification apparatus 100 using the wavelength of five different infrared regions is shown. The identification flow at the time of the identification part 170 identifying a several different object in the object identification apparatus 100 using the wavelength of five different infrared regions is shown.

  Hereinafter, the present invention will be described through embodiments of the invention. However, the following embodiments do not limit the invention according to the scope of claims, and all combinations of features described in the embodiments are included. It is not necessarily essential for the solution of the invention.

  FIG. 1 shows a functional block diagram of an object identification device 100 according to an embodiment of the present invention. The object identification device 100 measures the reflectivity for at least three or more different wavelengths including the infrared region from the reflected light with respect to the light irradiated to the predetermined measurement space, and based on the reflectivity, in the measurement space , An irradiation unit 110, a light receiving unit 120, a reflectance calculation unit 130, a synchronization circuit 140, a normalization index calculation unit 150, and a secondary differential value calculation unit 160. And an identification unit 170.

  For example, when the object identification device 100 is mounted on a passenger car and is intended to identify a person or an obstacle ahead of the passenger car, the predetermined measurement space is a space on the road ahead of the passenger car. In addition, when the object identification device 100 is mounted on a wheelchair and is intended to identify a person or an obstacle approaching the wheelchair, the space around the wheelchair is present.

  The irradiation unit 110 includes a plurality of semiconductor light emitting elements 111, 112, 113 and an irradiation condition adjustment unit 118, and irradiates a predetermined measurement space with light including at least three or more different wavelengths including the infrared region. .

  In this example, the plurality of semiconductor light emitting elements 111, 112, and 113 are, for example, light emitting diodes that emit infrared light in the infrared region of 1200 nm band, 1300 nm band, and 1500 nm band by applying a voltage of a predetermined magnitude. It is. In addition, the irradiation condition adjustment unit 118 adjusts the luminance of the plurality of semiconductor light emitting elements 111, 112, and 113 to set a luminance value of a predetermined size, and outputs information related to the luminance value to the reflectance calculation unit 130. To do.

  The light receiving unit 120 includes a plurality of specific band-pass filters (band-pass filters) 121, 122, 123 and photoelectric conversion elements 124, 125, 126 provided corresponding to the specific band-pass filters 121, 122, 123, respectively. And receiving reflected light of a plurality of different wavelengths including at least three of the reflected light from the measurement space.

  When the reflected light from the measurement space is input, the specific bandpass filters 121, 122, and 123 allow only light in a predetermined wavelength band of the reflected light to pass and block other light. In this example, each of the plurality of specific bandpass filters 121, 122, 123 allows only light in different wavelength bands to pass therethrough.

  Specifically, the specific band pass filter 121 passes only the component of the 1200 nm band. Further, the specific band pass filter 122 passes only the component in the 1300 nm band, and the specific band pass filter 123 passes only the component in the 1500 nm band.

  The band of light that each of the plurality of specific bandpass filters 121, 122, 123 passes is not limited to the 1200 nm band, the 1300 nm band, and the 1500 nm band. When identifying obstacles such as, and types of road surfaces such as asphalt road surfaces and concrete road surfaces, the above three infrared bands are considered as bands with little absorption due to moisture and air and having different absorption characteristics for the identification target. Are preferably selected.

  The plurality of photoelectric conversion elements 124, 125, and 126 are elements that output a voltage having a magnitude corresponding to the intensity of received light, for example, and correspond to the intensity of light from the corresponding specific bandpass filters 121, 122, and 123. The voltage of the magnitude is output to the reflectance calculation unit 130.

  The reflectance calculation unit 130 reflects the reflected light from the measurement space based on the voltage component that has passed through each of the plurality of photoelectric conversion elements 124, 125, and 126 and information on the luminance value from the irradiation condition adjustment unit 118. The reflectances of the components in a plurality of wavelength bands are calculated.

  Specifically, the reflectance calculation unit 130 is based on the 1200 nm band voltage component from the photoelectric conversion element 124 and the information on the 1200 nm band luminance value among the information on the luminance value from the irradiation condition adjustment unit 118. Thus, the reflectance in the 1200 nm band of the light irradiated to the measurement space from the plurality of semiconductor light emitting elements 111, 112, 113 is calculated. In addition, the reflectance calculation unit 130 reflects in the 1300 nm band and the 1500 nm band based on the voltage component from the photoelectric conversion elements 125 and 126 and the information on the luminance value from the irradiation condition adjustment unit 118 in the same manner as described above. Calculate the rate.

  In this example, the reflectance calculation unit 130 calculates a plurality of reflectances corresponding to each of the wavelengths in time series. More specifically, the reflectance calculation unit 130 calculates the reflectance in the 1200 nm band, the 1300 nm band, and the 1500 nm band at a plurality of timings.

  The synchronization circuit 140 includes information regarding the timing at which the irradiation condition adjusting unit 118 adjusts the setting of the luminance values of the plurality of semiconductor light emitting elements 111, 112, and 113, and the reflectance calculating unit 130 regarding the luminance value from the irradiation condition adjusting unit 118 Synchronize with the timing to get Therefore, even when the luminance values of the plurality of semiconductor light emitting elements 111, 112, and 113 are changed, the reflectance calculation unit 130 calculates the reflectance that reflects the changed luminance value.

  The normalization index calculation unit 150 calculates a normalization index based on the reflectance calculated by the reflectance calculation unit 130. In this example, the normalization index calculation unit 150 uses a plurality of normalization indexes based on the reflectances of the 1200 nm band, the 1300 nm band, and the 1500 nm band, which are simultaneously calculated at each timing in time series by the reflectance calculation unit 130. Calculate the conversion index.

More specifically, the normalization index calculation unit 150 uses the following formula based on the reflectance (R1200, R1300, R1500) of the 1200 nm band, 1300 nm band, and 1500 nm band calculated at each timing. Three normalization indexes (ND1 to ND3) defined by the above are calculated.
ND1: (R1300-R1200) / (R1300 + R1200)
ND2: (R1500-R1200) / (R1500 + R1200)
ND3: (R1500-R1300) / (R1500 + R1300)

The secondary differential value calculation unit 160 calculates a secondary differential value of a function of the reflectance and wavelength calculated by the reflectance calculation unit 130. In this example, the secondary differential value calculation unit 160 approximately calculates the secondary differential value based on the plurality of normalization indexes calculated by the normalization index calculation unit 150. Specifically, the secondary differential value calculation unit 160 calculates a value (2 ^nd der) approximated by the following formula as a secondary differential value.
2 ^nd der: R1500-3 × R1300 + 2 × R1200

  The discriminating unit 170 discriminates a plurality of different objects existing in the measurement space based on the plurality of normalization indexes, the second derivative value, and preset identification conditions. In this example, the identification unit 170 is calculated based on the plurality of normalization indexes sequentially calculated in time series by the normalization index calculation unit 150 and the plurality of normalization indexes by the secondary differential value calculation unit 160. By comparing the second-order differential value with a preset threshold value, a target existing in a region (measurement region) in which reflected light is received by the light-receiving unit 120 in the measurement space sequentially in time series. Identify.

  In addition, the object identification apparatus 100 of this example is not restricted to the form provided with one light receiving part 120 as shown in FIG. 1, and may include a plurality of light receiving parts 120. In this case, each light receiving unit 120 preferably receives reflected light from a plurality of different measurement areas in the measurement space. In addition, each of the plurality of light receiving units 120 may constitute a line sensor that receives reflected light from adjacent measurement regions in a specific direction.

  Hereinafter, the object identification flow by the identification unit 170 will be described in more detail.

  2 and 3 show an identification flow when the identification unit 170 identifies a plurality of different objects. In this identification flow, first, the identification unit 170 determines whether or not the reflectance (R1200, R1300, R1500) of each band of the 1200 nm band, the 1300 nm band, and the 1500 nm band is approximately 0, that is, from the measurement region. It is determined whether or not these reflected light components hardly contain these band components (step S200).

  In this example, the identification unit 170 performs the above determination based on whether each value of R1200, R1300, and R1500 is smaller than 0.02. And when all of R1200, R1300, and R1500 are smaller than 0.02 (S200: YES), the identification part 170 will identify that a window glass exists in a measurement area | region.

Further, R1200, R1300, and if all of R1500 is 0.02 or more (S200: NO), further, the identification unit 170, whether or not the value of the ² nd der is smaller than a predetermined threshold value Th1 A determination is made (step S210). Here, the threshold value Th1 is set to 0.09, for example.

If the value of 2 ^nd der is equal to or greater than Th1 (S210: NO), the identification unit 170 further determines whether or not the value of ND2 is greater than a predetermined threshold value Th2 (step S220). Here, the threshold Th2 is set to 0.03, for example. The case where the value of 2 ^nd der is smaller than Th1 in step S210 (S210: YES) will be described later.

  When the value of ND2 is larger than Th2 (S220: YES), the identification unit 170 identifies that asphalt exists in the measurement region. If the value of ND2 is equal to or less than Th2 (S220: NO), the identification unit 170 further determines whether or not the value of ND2 is greater than a predetermined threshold value Th3 (step S230). Here, the threshold Th3 is set to, for example, -0.035. When the value of ND2 is equal to or less than Th3 (S230: NO), the identification unit 170 identifies that there is concrete or stone in the measurement region in the measurement space.

  If the value of ND2 is greater than Th3 (S230: YES), the identification unit 170 further determines whether or not the value of ND3 is greater than 0 (step S240). When the value of ND3 is larger than 0 (S240: YES), the identification unit 170 identifies that wool (wool) exists in the measurement region. When the value of ND2 is 0 or less (S240: NO), the identification unit 170 identifies that a metal exists in the measurement region.

In step S210, when the value of 2 ^nd der is smaller than Th1 (S210: YES), the identification unit 170 further determines whether or not the value of ND2 is larger than a predetermined threshold Th4 ( Step S250). Here, the threshold Th4 is set to −0.58, for example.

  When the value of ND2 is larger than Th4 (S250: YES), the identification unit 170 identifies that human skin or a plant exists in the measurement region. On the other hand, when the value of ND2 is equal to or less than Th4 (S250: NO), the identification unit 170 identifies that an animal or fabric (cotton) exists in the measurement region. And this identification flow by the identification part 170 is complete | finished. The predetermined threshold values Th1 to Th4 are values that can accurately identify a target when a measurement region is previously set for a known target such as metal or human skin and the identification unit 170 performs the identification flow. It is set in search.

  When the object identification device 100 of the present embodiment includes the above-described line sensor, the identification unit 170 may determine the presence of various objects as follows, for example. That is, the identification unit 170, for example, when at least one of wool (wool fabric) or animal or fabric (cotton fabric) and human skin are continuously present in one or a plurality of adjacent measurement regions, It may be determined that a person exists in the measurement space.

  In addition, for example, when asphalt exists in a plurality of continuous measurement regions, the identification unit 170 may determine that an asphalt road surface exists in the continuous measurement regions. In addition, the identification unit 170, for example, in the measurement space when it is identified that the window glass and the metal are continuously present in one or a plurality of measurement areas adjacent to the measurement area identified as the presence of the window glass. It may be determined that there is a car (for example, a passenger car).

  Thus, the object identification device 100 of the present embodiment can identify with high accuracy what kind of object is present in the predetermined measurement region. Further, since only the wavelength in the infrared region is used, there is almost no decrease in the identification accuracy due to the influence of the measurement environment, and the load on the processing system is small. Furthermore, no problem arises from the viewpoint of protecting privacy or protecting public morals when identifying objects. Moreover, by setting it as the structure provided with the line sensor containing the several light-receiving part 120, based on the identification result about a several measurement area | region, the presence or absence of obstacles, such as a person and a vehicle, can be determined more correctly. .

  By the way, when the object identification device 100 used in the daytime is used at night, or when the object identification device 100 used in the fine weather is used in the rain, the environmental conditions of the object identification device 100 are used. If the identification unit 170 identifies an object using the threshold values (Th1 to Th4) set before the environmental condition changes, the identification accuracy may decrease.

  Therefore, the object identification device 100 according to the present embodiment performs identification for the known object specified in advance after the change in the environmental condition, and when the identification result does not identify the known object, the following calibration is performed. I do.

  First, identification by the object identification device 100 is performed on a known target specified in advance, and a normalization index calculation unit 150 calculates a normalization index based on the reflectance of each wavelength from the known target. Next, the identification unit 170 calculates an average value of the normalization index and the set threshold value. Then, the identification unit 170 sets the calculated value as a new threshold value. As a result, the threshold value set in the identification unit 170 can be calibrated to a value corresponding to a change in the environmental condition, so that the target identification accuracy can be prevented from being lowered due to the influence of the change in the environmental condition.

  In addition, by repeating this calibration a plurality of times, it is possible to further improve the identification accuracy of the object after the environmental condition changes. The object identification device 100 may include a calibration that detects a change in the environmental condition, and may automatically perform the calibration according to the change in the environmental condition. In this case, the target identification device 100 may include a known target used for calibration as part of itself.

  The object identification device 100 of the present embodiment is not limited to a form that identifies an object in the measurement space using only wavelengths in the infrared region. For example, in addition to the above-described three infrared wavelengths, the object in the measurement space is further measured using a red wavelength in the 680 nm band that has specific absorption characteristics for chlorophylls present in plant leaves and the like. You may identify. In this embodiment, the irradiation unit 110 may include a semiconductor light emitting element that irradiates the measurement space with light including the red wavelength, and the light receiving unit 120 corresponds to reflected light having the red wavelength. There may be provided a specific band pass filter that allows only the component to pass.

In this embodiment, the normalization index calculation unit 150 adds each of the 1200 nm band, the 1300 nm band, the 1500 nm band, and the 680 nm band calculated at each timing in addition to the above three normalization indices (ND1 to ND3). Based on the reflectance of the band (R1200, R1300, R1500, R680), a normalization index (ND4) defined by the following formula is calculated.
ND4: (R1500-R680) / (R1500 + R680)

  In the following, the object identification flow by the identification unit 170 in the object identification device 100 of the present embodiment will be described more specifically.

  FIG. 4 shows a part of an identification flow when the identification unit 170 identifies a plurality of different objects in the object identification device 100 using the red color range. The identification unit 170 omits the description because the identification flow shown in FIG. 2 is the same when the object is identified only in the infrared region and when the red region is used. That is, the identification flow shown in FIG. 4 is an alternative to the identification flow shown in FIG.

In this example, when the value of 2 ^nd der is smaller than Th1 in Step S210 (S210: YES), the identification unit 170 determines whether or not the value of ND4 is larger than a predetermined threshold Th5 (Step S210). S260).

  When the value of ND4 is larger than Th5 (S260: YES), the identification unit 170 identifies that a plant exists in the measurement region. On the other hand, when the value of ND4 is equal to or less than Th5 (S260: NO), the identification unit 170 further determines whether or not the value of ND2 is larger than a predetermined threshold value Th4 (step S270). Here, the size of the threshold Th4 is the same as described above.

  When the value of ND2 is larger than Th4 (S250: YES), the identification unit 170 identifies that human skin (human skin) exists in the measurement region. On the other hand, when the value of ND2 is equal to or less than Th4 (S250: NO), the identification unit 170 identifies that an animal or fabric (cotton) exists in the measurement region. And this identification flow by the identification part 170 is complete | finished.

  For the threshold Th5, a value that can accurately identify a plant when a measurement region is previously set for a plant, human skin, or animal (cotton fabric) and the identification unit 170 performs the identification flow is obtained. Is set.

  In this way, by using the red region of the 680 nm band in addition to the infrared region for identification of the object, human skin or animals (cotton fabric) and plants can be identified.

  As described above, the object identification device 100 according to the present embodiment identifies a target by using more wavelengths in the infrared region, instead of using the three infrared regions and the form further using the red region. May be. For example, in order to accurately distinguish wet asphalt and wet concrete from asphalt and concrete, in addition to the above three infrared wavelengths, the measurement space is further measured using wavelengths in the 1100 nm band and 1600 nm band. You may identify objects within. In the present embodiment, the irradiation unit 110 may further include a semiconductor light emitting element that irradiates the measurement space with light having the two new infrared wavelengths, and the light receiving unit 120 includes the new infrared region. You may have a specific band pass filter which passes only the component corresponding to the reflected light of a wavelength.

Further, in this embodiment, the normalization index calculation unit 150 is configured to reflect the reflectances (R1100, R1200, R1300, R1500, 1100 nm band, 1200 nm band, 1300 nm band, 1500 nm band, and 1600 nm band) calculated at each timing. R1600), four normalization indices (ND1 to ND3) defined by the following formula are calculated.
ND1: (R1500-R1300) / (R1500 + R1300)
ND2: (R1500-R1200) / (R1500 + R1200)
ND3: (R1600-R1300) / (R1600 + R1300)
ND4: (R1300-R1100) / (R1300 + R1100)

Further, the secondary differential value calculation unit 160 calculates a secondary differential value of the function of the reflectance and the wavelength. In the present embodiment, the secondary differential value calculation unit 160 is based on the four normalization indexes calculated by the normalization index calculation unit 150 and is approximated by the following formula (2 ^nd der1, 2 ^nd der2 ) As a second derivative value.
2 ^nd der1:
[{(R1500-R1300) / (R1500 + R1300)} / 200]
-[{(R1300-R1200) / (R1300 + R1200)} / 100]
2 ^nd der2:
[{(R1500-R1200) / (R1500 + R1200)} / 300]
-[{(R1200-R1100) / (R1200 + R1100)} / 100]

  In the following, the object identification flow by the identification unit 170 in the object identification device 100 of the present embodiment will be described more specifically.

  5 and 6 show an identification flow when the identification unit 170 identifies a plurality of different objects in the object identification apparatus 100 using five different infrared wavelengths. In this identification flow, first, the identification unit 170 has approximately zero reflectance (R1100, R1200, R1300, R1500, R1600) in each band of the 1100 nm band, 1200 nm band, 1300 nm band, 1500 nm band, and 1600 nm band. It is determined whether or not there is, that is, whether or not the reflected light from the measurement region contains almost no component of these bands (step S310).

  In this embodiment, the identification unit 170 performs the above determination based on whether or not each value of R1100, R1200, R1300, R1500, and R1600 is smaller than 0.02. And when all of these reflectances are smaller than 0.02 (S310: YES), the identification part 170 will identify that a window glass exists in a measurement area | region.

Further, R1100, R1200, R1300, R1500 , and if all of R1600 is 0.02 or more (S310: NO), further, the identification unit 170, the sum of the ND1 and the ND3 to the ² nd ^der1 It is determined whether or not the multiplied value {2 ^nd der1 × (ND1 + ND3)} is smaller than a predetermined threshold value Th11 (step S320).

  If the value is equal to or greater than Th11 (S320: NO), the identification unit 170 further determines whether or not the value of ND2 is greater than a predetermined threshold value Th12 (step S330). The case where the above value is smaller than Th11 in step S320 (S320: YES) will be described later.

When the value of ND2 is larger than Th12 (S330: YES), the identification unit 170 identifies that there is an animal or fabric (cotton) in the measurement region. The value of the ND2 may be Th12 less (S330: NO), the identification unit 170, Furthermore, the value of the ² nd ^der2 it is determined whether or not smaller than a predetermined threshold Th13 (Step S340) .

When the value of 2 ^nd der2 is smaller than Th13 (S340: YES), the identification unit 170 identifies that a plant exists in the measurement region. When the value of the ² nd ^der2 is Th13 or more (S340: NO), the identification unit 170 identifies the human skin to the measurement region in the measurement space is present.

In step S320, if the value {2 ^nd der1 × (ND1 + ND3)} is smaller than Th11 (S320: YES), the identification unit 170 further determines whether the value of ND3 is smaller than a predetermined threshold Th14. It is determined whether or not (step S350).

  When the value of ND3 is smaller than Th14 (S350: YES), the identification unit 170 further determines whether or not the value of ND4 is smaller than a predetermined threshold value Th15 (step S360). If the value of ND3 is equal to or greater than Th14 (S350: NO), the identification unit 170 further determines whether or not the value of ND2 is smaller than a predetermined threshold value Th16 (step S370).

  In step S360, when the value of ND4 is smaller than Th15 (S360: YES), the identification unit 170 identifies that there is wet metal (for example, a wet car body surface) in the measurement region. When the value of ND4 is equal to or greater than Th15 (S360: NO), the identification unit 170 identifies that there is concrete or stone in the measurement region in the measurement space.

  In step S370, when the value of ND2 is smaller than Th16 (S370: YES), the identification unit 170 identifies that asphalt exists in the measurement region. Further, when the value of ND2 is equal to or greater than Th16 (S370: NO), the identification unit 170 identifies that wet concrete or stone exists in the measurement region in the measurement space. And this identification flow by the identification part 170 is complete | finished.

  The object identification device 100 according to the present embodiment uses a plurality of normalized indexes calculated based on reflectances at five wavelengths in different infrared regions, in particular, normalized indexes calculated using reflectances in the 1300 nm band and the 1600 nm band. By using it, it is possible to accurately identify metal wetted by rain or wet concrete or wet concrete.

  As mentioned above, although this invention was demonstrated using embodiment, the technical scope of this invention is not limited to the range as described in the said embodiment. It will be apparent to those skilled in the art that various modifications or improvements can be added to the above-described embodiment. It is apparent from the scope of the claims that the embodiments added with such changes or improvements can be included in the technical scope of the present invention.

DESCRIPTION OF SYMBOLS 100 ... Object identification apparatus 110 ... Irradiation part 111,112,113 ... Semiconductor light emitting element 118 ... Irradiation condition adjustment part 120 ... Light receiving part 121,122,123 ... Specific bandpass filter 124,125,126 ... Photoelectric conversion element 130 ... Reflection Rate calculation unit 140 ... synchronization circuit 150 ... normalization index calculation unit 160 ... secondary differential value calculation unit 170 ... identification unit

Claims (8)

A light receiving unit that receives the reflected light of a plurality of different wavelengths including at least three of the reflected light with respect to the light irradiated to the predetermined measurement space; and
A reflectance calculator that calculates the reflectance for each of the plurality of wavelengths;
A normalization index calculator that calculates a normalization index based on the reflectance for each of the plurality of wavelengths;
An identification unit for identifying a plurality of different objects existing in the measurement space according to the normalization index and a preset identification condition;
An object identification device comprising: The reflectance calculation unit calculates a plurality of reflectances for each of the plurality of wavelengths in time series,
The normalization index calculation unit calculates the normalization index based on the reflectance for each of the plurality of wavelengths calculated simultaneously,
The object identification device according to claim 1, wherein the identification unit sequentially identifies the object in time series based on the normalization index and the identification condition. A second derivative calculation unit for calculating a second derivative of a function of the reflectance and the wavelength;
The object identification device according to claim 1, wherein the identification unit identifies the object based on the normalization index, the second-order differential value, and a preset identification condition. The normalization index calculation unit calculates a plurality of the normalization indexes based on the reflectance of each wavelength,
The object identification device according to claim 3, wherein the secondary differential value calculation unit approximately calculates the secondary differential value based on a plurality of the normalized indexes.   5. The object identification device according to claim 3, wherein the identification unit identifies the object by comparing the normalization index and the secondary differential value with a preset threshold value.   The identification unit is an average of the normalized index and the threshold when the identification result obtained by comparing the normalized index calculated for the previously specified known target and the threshold does not identify the known target. The object identification device according to claim 5, wherein the value is a new threshold value.   The object identification device according to claim 1, wherein the plurality of wavelengths include a red region.   The object identification device according to claim 1, further comprising an irradiation unit that irradiates the measurement space with light including a plurality of wavelengths.

Images