JP2013137225A - Object detection device and object detection method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an object detection device and an object detection method capable of quickly detecting an object over a wider range.SOLUTION: An object detection device 101 includes an imaging unit 11 for imaging a one-dimensional image for a plurality of wavelength ranges selected in advance in accordance with physical property of a detection object, and a detection unit 12 for detecting a substance in pixels of the one-dimensional image.

Description

  The present invention relates to an object detection apparatus and an object detection method, and more particularly to an object detection apparatus and an object detection method for detecting an object using a one-dimensional image.

  “Imaging spectrometer”, [online], [searched April 27, 2011], Internet <URL: http://www.ryoden.co.jp/fa/system/fa28_4.html> (Non-patent document 1 ) Discloses instruments used in the fields of industrial and research spectroscopic analysis. That is, in the imaging spectrometer, the light at each point in the line-shaped area is dispersed by a special prism and a grating structure. Then, by combining an ordinary lens, a black and white two-dimensional camera, and an imaging spectrometer, an imaging spectrometer capable of detecting the wavelength distribution of the line area is realized. Light intensity data is obtained for 256 to 1024 wavelengths at each point in the line area.

“Imaging Spectrometer”, [online], [Search April 27, 2011], Internet <URL: http://www.ryoden.co.jp/fa/system/fa28_4.html>

  However, in the imaging spectrometer described in Non-Patent Document 1, since it is necessary to obtain light intensity data for multiple wavelengths, the processing time is increased, and the sensor portion for detecting the light intensity of each wavelength is very It becomes expensive.

  The present invention has been made to solve the above-described problems, and an object thereof is to provide an object detection apparatus and an object detection method that can detect an object at a higher speed with an inexpensive configuration.

  In order to solve the above-described problem, an object detection apparatus according to an aspect of the present invention includes an imaging unit for imaging a one-dimensional image with respect to a plurality of wavelength bands selected in advance according to physical properties of a detection target, A detection unit for detecting a substance in each pixel of the one-dimensional image.

  When a detection target such as a person is recognized by its shape, for example, data of several tens of pixels is required to express the shape. When detecting a substance in each pixel of a one-dimensional image, if there is a detection result of the substance of about one pixel or several pixels by appropriately selecting the wavelength band used for the detection, the detection target from the background of the one-dimensional image The inventors of the present application have found that and the position can be extracted. In other words, even if the imaging range is wide, it is possible to ensure the accuracy of specifying the detection target and its position. Since it does not depend on the shape, it is possible to specify the target even when the entire target cannot be detected. There is little need to change the focal length in order to increase the number of pixels used for recognizing the detection target, and the movement of the detection target can be easily tracked. In addition, by selecting the wavelength band according to the physical property of the detection target, it becomes unnecessary to analyze the images of each of the many wavelength bands, so that the planar area can be analyzed at high speed. In addition, since object detection processing can be performed with a small number of image sensors, it is possible to perform object detection at a higher speed over a wide range. In addition, the object detection process can be simplified by the configuration in which the substance detection is performed using the one-dimensional image. In addition, since it is not necessary to detect the intensity of light having a large number of wavelengths, the sensor portion can be configured at low cost.

  Preferably, the imaging unit includes a plurality of line sensors that are provided corresponding to the wavelength band and detect the intensity of the corresponding wavelength band.

  With such a configuration, object detection can be performed satisfactorily by using a combination of inexpensive line sensors without using an expensive two-dimensional sensor.

  More preferably, the imaging unit further includes a branching unit for branching received light and irradiating the plurality of line sensors.

  With such a configuration, an optical system can be constructed with a simple configuration in a configuration using a plurality of line sensors.

  Preferably, the detection unit detects a substance in the pixel based on an evaluation value calculated by combining a part or all of the intensity values for the plurality of wavelength bands in one pixel.

  With such a configuration, it is possible to appropriately evaluate the characteristics of the detection target and the like with respect to the wavelength without obtaining images in a large number of wavelength bands, and the target detection process can be performed more accurately. In addition to the evaluation value of the target pixel itself, the intensity value, the evaluation value, and the substance discrimination result of the surrounding pixels can be used for detecting the substance in the target pixel.

  Preferably, at least some of the plurality of wavelength bands are included in the infrared region.

  With such a configuration, it is possible to more appropriately evaluate the substance contained in the detection target or its background.

  As one embodiment, the pixel is a combination of a plurality of pixels respectively corresponding to the plurality of wavelength bands, and each pixel includes a wavelength selection filter for selecting a corresponding wavelength band, and the wavelength selection filter. And an imaging device for receiving the transmitted light.

  In order to solve the above-described problems, an object detection method according to an aspect of the present invention includes a step of capturing a one-dimensional image for a plurality of wavelength bands selected in advance according to physical properties of a detection target, and the one-dimensional image Detecting a substance in each pixel.

  According to the present invention, an object can be detected at a higher speed with an inexpensive configuration.

It is a functional block diagram of the object detection apparatus which concerns on embodiment of this invention. It is a figure which shows in detail the structure of the optical system and the light-receiving part in the object detection apparatus which concerns on embodiment of this invention. It is a figure which shows the structure of the branching device in the object detection apparatus which concerns on embodiment of this invention in detail. It is a figure which shows the structure of the light-receiving part in the object detection apparatus which concerns on embodiment of this invention in detail. It is a figure which shows notionally the image processing by the light-receiving part in the target detection apparatus which concerns on embodiment of this invention. It is a figure which shows the transmission wavelength of the wavelength filter part in the light-receiving part of the target detection apparatus which concerns on embodiment of this invention. It is a functional block diagram of the detection part in the object detection apparatus which concerns on embodiment of this invention. It is the flowchart which defined an example of the operation | movement procedure when the determination part in the detection part of the target detection apparatus which concerns on embodiment of this invention performs a substance determination process.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the drawings, the same or corresponding parts are denoted by the same reference numerals and description thereof will not be repeated.

  FIG. 1 is a functional block diagram of an object detection apparatus according to an embodiment of the present invention. The target detection device 101 can be used for a monitoring device, and further for detecting an intruder in a monitoring area and capturing an image of the intruder.

  With reference to FIG. 1, the target detection apparatus 101 includes an imaging unit 11 and a detection unit 12. The imaging unit 11 includes an optical system 21 and a light receiving unit 22.

  The imaging unit 11 captures a one-dimensional image for a plurality of wavelength bands selected in advance according to the physical properties of the detection target. A hyperspectral (HS) camera can be used for this imaging. In the present embodiment, the HS camera captures electromagnetic waves in a target wavelength range including at least the near infrared range. The target wavelength range is, for example, 900 nm to 1800 nm, and includes the plurality of wavelength bands selected in advance. These wavelength bands are selected, for example, up to 10 wavelength bands.

  Here, the reason why the number of wavelength bands to be selected is 10 or less is that if it is larger than that, it becomes difficult to discriminate substances at a high speed (almost real time) with a constant resolution, which is not suitable for monitoring applications. In addition, even if the number of wavelength bands to be selected is 10 or less, the number of substances to be identified and the accuracy can be sufficiently secured. In consideration of practicality of a monitoring device or the like, it is preferable to select the 2-6 wavelength band, and it is more preferable to select the 3-5 wavelength band.

  The HS camera is installed in a diagonally downward direction at a place where the target area can be photographed, for example, at the top of a pillar near the target area. When the target area cannot cover the entire monitoring area or when a blind spot occurs, a plurality of HS cameras can be installed.

  In the imaging unit 11, the light receiving unit 22 includes a plurality of pixels P. The plurality of pixels P each output an electric signal indicating the reception intensity of the electromagnetic wave in each selected wavelength band.

  The light receiving unit 22 captures a one-dimensional image of the target area using an electrical signal output from the plurality of pixels P.

  In the present embodiment, the detection unit 12 determines a substance for each pixel P based on the one-dimensional image captured by the light receiving unit 22. The detection unit 12 only needs to be able to determine a substance according to the detection target of the target detection device 101. When the target detection device 101 is used as a monitoring device, for example, it is only necessary to be able to determine human skin and other substances. The substance determination is not limited to each pixel P but may be performed for each pixel group, and may be performed for each pixel group in a partial region and for each pixel group in the remaining region.

  FIG. 2 is a diagram showing in detail the configuration of the optical system and the light receiving unit in the object detection device according to the embodiment of the present invention.

  Referring to FIG. 2, the optical system 21 includes an objective lens 61, a slit mechanism 62, and a branching device 63.

  The optical system 21 receives light from the monitoring area and irradiates the light receiving unit 22. In the optical system 21, the objective lens 61 collects light from the target area LN and forms an image on the slit of the slit mechanism 62. The target area LN has a height of 25 cm and a width of 40 m, for example, and the distance between the optical system 21 and the target area LN is 40 m.

  The light receiving unit 22 creates a one-dimensional image based on the light (here, infrared rays) received from the optical system 21. A target TG, here a person, is detected based on this one-dimensional image, and an image of the target TG in the target area LN is taken.

  The light receiving unit 22 includes a plurality of sets including the wavelength filter unit 31 and the imaging element unit 32, and here, three sets as an example. The wavelength filter unit 31 and the image sensor unit 32 constitute a line sensor. A plurality of line sensors are provided corresponding to the wavelength band for substance discrimination, and detect the intensity of the corresponding wavelength band.

  The branching device 63 in the optical system 21 branches the light received from the slit of the slit mechanism 62 and irradiates each line sensor.

  FIG. 3 is a diagram showing in detail the configuration of the branching device in the object detection device according to the embodiment of the present invention.

  Referring to FIG. 3, branching device 63 includes a beam splitter 71, a half mirror 72, and mirrors 73 and 74.

  The beam splitter 71 divides the light of a certain optical axis received from the slit of the slit mechanism 62 at an intensity ratio of 2: 1, irradiates the half mirror 72 with light of intensity “2”, and emits light of intensity “1”. Irradiate the mirror 73.

  The half mirror 72 divides the light received from the beam splitter 71 at an intensity ratio of 1: 1, irradiates one light to the wavelength filter unit 31 in the line sensor 1, and irradiates the other light to the mirror 74.

  The mirror 74 reflects the light received from the half mirror 72 and irradiates the wavelength filter unit 31 in the line sensor 2.

  The mirror 73 reflects the light received from the beam splitter 71 and irradiates the wavelength filter unit 31 in the line sensor 3.

  FIG. 4 is a diagram showing in detail the configuration of the light receiving unit in the object detection device according to the embodiment of the present invention.

  With reference to FIG. 4, in the line sensors 1 to 3, the image sensor section 32 includes a plurality of image sensors R arranged in a line. The imaging element R is a light receiving element using, for example, InGaAs. The image sensor R outputs an electrical signal indicating the intensity of the received light.

  The wavelength filter unit 31 includes a wavelength selection filter F. The wavelength selection filter F transmits light in a corresponding wavelength band among a plurality of wavelength bands for substance discrimination and irradiates each image sensor R in the corresponding image sensor section 32.

  Specifically, the wavelength filter unit 31 that is aligned with each image sensor R in the image sensor unit 32 is disposed immediately above the image sensor unit 32. The wavelength filter unit 31 is integrated with the image sensor unit 32, for example. The vertical and horizontal sizes of the light receiving surface of the wavelength filter unit 31 substantially coincide with the light receiving surface of the imaging element unit 32. Each image sensor R outputs an electrical signal indicating the intensity of light in the wavelength band transmitted by the corresponding wavelength selection filter F.

  In the light receiving unit 22, for example, two sets of a plurality of image pickup devices R arranged in a row are provided, and each set is shifted within the interval of one image pickup device R, and the difference in the output of the image pickup device R between the sets is set. It is also possible to form a large number of pixels by calculating. For example, instead of a set of 32 image sensors R arranged in a row, two sets of 16 image sensors R arranged in a row are prepared and arranged as described above, so that 32 Pixels can be formed at low cost.

  FIG. 5 is a diagram conceptually showing image processing by the light receiving unit in the object detection device according to the embodiment of the present invention.

  FIG. 6 is a diagram illustrating a transmission wavelength of the wavelength filter unit in the light receiving unit of the target detection device according to the embodiment of the present invention.

  Referring to FIG. 6, line sensors 1 to 3 are provided with wavelength selection filters F that respectively transmit light in a plurality of wavelength bands for substance discrimination.

  It is preferable that at least some of the plurality of wavelength bands are included in the infrared region, and further in the near infrared region. More specifically, the line sensor 1 has a wavelength selection filter that transmits light having a wavelength in the 1600 nm band, for example. The line sensor 2 has a wavelength selection filter that transmits light having a wavelength in the 1200 nm band, for example. The line sensor 3 has a wavelength selection filter that transmits light having a wavelength of 940 nm band, for example. Each wavelength band may have a width of ± several tens of nm with respect to the center wavelength, for example.

  The pixel P is a combination of a plurality of pixels SP respectively corresponding to a plurality of wavelengths. In the example shown in FIG. 5 and FIG. 6, one pixel P is constituted by the three pixels SP of the line sensors 1 to 3 shown in FIG. 4, and calculation is performed for each pixel P to determine the substance of the image in the pixel P. Do. Then, by displaying the pixels P determined to be human skin separately from the other pixels P, for example, a human-type spectrum image can be obtained.

  The detection unit 12 detects a substance in the pixel P based on an evaluation value calculated by combining some or all of intensity values for a plurality of wavelength bands in one pixel P. More specifically, the detection unit 12 compares the intensities of the plurality of selection lights in the pixel P in the one-dimensional image, and determines the substance in the pixel P based on the comparison result. The detection unit 12 determines a substance for each pixel P based on the electrical signals output from the plurality of light receiving elements R in the pixel P.

  In the target detection apparatus according to the embodiment of the present invention, the imaging unit 11 captures a one-dimensional image for a plurality of bands of wavelengths selected in advance according to the physical property of the detection target, as described above. Since it is not necessary to perform analysis processing on each one-dimensional image having a large number of wavelengths, it is possible to perform object detection at a higher speed over a wide range. In addition, since it is not necessary to detect the intensity of light having a large number of wavelengths, the sensor portion can be configured at low cost.

  Furthermore, in the object detection apparatus according to the embodiment of the present invention, in the imaging unit 11, a plurality of line sensors are provided corresponding to the wavelength band, and detect the intensity of the corresponding wavelength band.

  With such a configuration, object detection can be performed satisfactorily by using a combination of inexpensive line sensors without using an expensive two-dimensional sensor.

  Furthermore, in the target detection apparatus according to the embodiment of the present invention, in the imaging unit 11, the branching device 63 branches the light received from the target area and irradiates the plurality of line sensors.

  With such a configuration, an optical system can be constructed with a simple configuration in a configuration using a plurality of line sensors.

  Furthermore, in the target detection device according to the embodiment of the present invention, the detection unit 12 uses the pixel based on the evaluation value calculated by combining some or all of the intensity values for a plurality of wavelengths in one pixel P. The substance in P is detected.

  With such a configuration, it is possible to more appropriately evaluate the detection target and the substances contained in the background. In addition to the evaluation value of the pixel itself, the intensity value, the evaluation value, and the substance discrimination result of the surrounding pixels can be used for detecting the substance in one pixel. For example, the intensity values of a plurality of pixels may be averaged, and the evaluation value may be calculated from the average value.

  In the target detection device according to the embodiment of the present invention, at least a part of the plurality of wavelength bands is included in the infrared region.

  With such a configuration, an appropriate wavelength band can be selected as the selection light, and substance discrimination can be performed more accurately.

  In the object detection device according to the embodiment of the present invention, the pixel P is a combination of a plurality of pixels SP respectively corresponding to a plurality of wavelength bands. Each pixel SP includes a wavelength selection filter F for selecting a corresponding wavelength, and an image sensor R for receiving transmitted light from the wavelength selection filter F. Note that the arrangement of the filters is not limited to the example of FIG.

  FIG. 7 is a functional block diagram of a detection unit in the target detection device according to the embodiment of the present invention.

  Referring to FIG. 7, in this example, detection unit 12 includes a reflectance calculation unit 91, a normalization index calculation unit 92, a secondary differential value calculation unit 93, and a determination unit 94.

  The detection unit 12 can calculate the reflectance of the plurality of selection lights in the pixel P of the one-dimensional image, and can determine the substance in the pixel P of the one-dimensional image based on the calculated reflectance of the plurality of selection lights. .

  More specifically, the reflectance calculation unit 91, based on the one-dimensional image received from the light receiving unit 22, has wavelength components of 1100 nm band, 1200 nm band, 1300 nm band, 1500 nm band, and 1600 nm band in the pixel P of the one-dimensional image. The reflectances R1100, R1200, R1300, R1500, and R1600 are calculated. When irradiating infrared light, the reflectance can be calculated from the emission luminance of the infrared light source and the intensity value of each pixel. When using natural light, the reflectance may be calculated separately by measuring the illuminance.

The normalization index calculation unit 92 uses the reflectances R1100, R1200, R1300, R1500, and R1600 calculated by the reflectance calculation unit 91 to calculate normalization indexes ND1 to ND5 defined by the following expressions.
ND1 = (R1500-R1200) / (R1500 + R1200)
ND2 = (R1300-R1200) / (R1300 + R1200)
ND3 = (R1600-R1300) / (R1600 + R1300)
ND4 = (R1300-R1100) / (R1300 + R1100)
ND5 = (R1500-R1300) / (R1500 + R1300)

The secondary differential value calculation unit 93 calculates the secondary differential value of the function of the reflectance and wavelength. Specifically, the secondary differential value calculation unit 93 uses the reflectances R1100, R1200, R1300, R1500, and R1600 calculated by the reflectance calculation unit 91, and uses the approximate two defined by the following equation. Next derivative values der1 and der2 are calculated. der1 is a secondary difference by R1200, R1300, and R1500. Also, der2 is a secondary difference by R1100, R1200, R1500. der1 and der2 are represented by the following equations.
der1 = [{(R1500-R1300) / (R1500 + R1300)} / 200]-[{(R1300-R1200) / (R1300 + R1200)} / 100]
der2 = [{(R1500-R1200) / (R1500 + R1200)} / 300]-[{(R1200-R1100) / (R1200 + R1100)} / 100]

  The determination unit 94 determines the substance in the pixel P of the one-dimensional image based on the reflectances R1100, R1200, R1300, R1500, R1600, the normalization indices ND1 to ND4, and the secondary differential values der1 and der2.

  FIG. 8 is a flowchart defining an example of an operation procedure when the determination unit in the detection unit of the target detection device according to the embodiment of the present invention performs the substance determination process.

  Referring to FIG. 8, first, determination unit 94 determines whether or not the values of reflectances R1100, R1200, R1300, R1500, and R1600 are approximately zero. For example, when the reflectance value is less than a predetermined value (for example, 0.02), it is determined that the reflectance is approximately 0, and the reflectance value is greater than or equal to the predetermined value. Is determined not to be approximately zero. Determination unit 94 determines that glass is present when the values of reflectances R1100, R1200, R1300, R1500, and R1600 are all approximately 0 (NO in step S11).

  On the other hand, when any one of the reflectances R1100, R1200, R1300, R1500, and R1600 is not approximately 0 (YES in step S11), the determination unit 94 adds the normalized indicators ND5 and ND3 to the secondary differential value der1. It is determined whether the value “der1 × (ND5 + ND3)” multiplied by the sum is greater than 0 (step S12).

  If “der1 × (ND5 + ND3)” is greater than 0 (YES in step S12), the determination unit 94 determines whether the value of the normalized index ND1 is greater than 0 (step S13).

  When the value of the normalization index ND1 is greater than 0 (YES in step S13), the determination unit 94 determines that there is an animal or fabric.

  On the other hand, when the value of the normalization index ND1 is 0 or less (NO in step S13), the determination unit 94 determines whether the value of the secondary differential value der2 is greater than 0 (step S14).

  The determination unit 94 determines that human skin exists when the value of the secondary differential value der2 is greater than 0 (YES in step S14).

  On the other hand, the determination part 94 determines with a plant existing, when the value of the secondary differential value der2 is 0 or less (it is NO at step S14).

  If the value of “der1 × (ND5 + ND3)” is 0 or less (NO in step S12), the determination unit 94 determines whether the value of the normalization index ND3 is greater than 0 (step S15). ).

  If the value of the normalized index ND3 is greater than 0 (YES in step S15), the determination unit 94 determines whether the value of the normalized index ND2 is greater than 0 (step S16).

  The determination unit 94 determines that asphalt exists when the value of the normalization index ND2 is 0 or less (NO in step S16).

  On the other hand, when the value of the normalization index ND2 is greater than 0 (YES in step S16), the determination unit 94 determines that concrete is present.

  Further, when the value of the normalization index ND3 is 0 or less (NO in step S15), the determination unit 94 determines whether the value of the normalization index ND4 is greater than 0 (step S17).

  The determination unit 94 determines that concrete is present when the value of the normalization index ND4 is greater than 0 (YES in step S17).

  On the other hand, when the value of the normalization index ND4 is 0 or less (NO in step S17), the determination unit 94 determines that glass is present.

  In each of the above steps for determining whether or not it is greater than 0, a threshold value other than 0 may be set according to the physical properties of the determination target substance.

  When the determination unit 94 determines that human skin is present, the detection unit 12 can determine that an intruder exists in the monitoring area. For example, when the determination unit 94 determines that glass is present, the detection unit 12 can determine that a vehicle is present in the monitoring area.

  In addition, the situation where the human skin is not exposed because the intruder into the monitoring area wears a covering surface and gloves is also assumed. Assuming such a case, for example, in the target detection device 101, the threshold value and the wavelength are set so that a substance such as a chemical fiber worn only by a human can be detected. Thereby, when the determination unit 94 determines that the chemical fiber is present, the detection unit 12 can determine that an intruder exists in the monitoring area.

  It is also possible to detect the material of clothing worn by humans by setting a threshold and wavelength that can detect the type of fiber such as wool, cotton, nylon (registered trademark), and polyester.

  In addition, assuming that the object detection device 101 only detects a human, the configuration may be such that only the substance related to the human is separated.

  Moreover, although the object detection apparatus according to the embodiment of the present invention is used for the purpose of detecting an intruder in the monitoring area, the object detection apparatus is not limited to such an object, but for the purpose of detecting some object. Widely applicable.

  For example, the detection target may be a bicycle traveling on a general road or a highway, a blind person with a cane, or the like. Moreover, it is applicable also to the use which detects the growing state of the crop in a fixed area.

  In addition to the purpose of intrusion detection, the object detection device according to the embodiment of the present invention can also be used as a monitoring device that monitors a person or an object for purposes other than intrusion detection. For example, it may be used as a monitoring device that monitors traffic conditions at intersections.

  Further, in the object detection device according to the embodiment of the present invention, the object detection is performed without being noticed by a person in the dark and at night by providing the light irradiator that irradiates halogen light or infrared light as described above. It can also be used for the purpose.

  The above embodiment should be considered as illustrative in all points and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

1-3 Line sensors 11 Imaging unit 12 Detection unit 21 Optical system 22 Light receiving unit 31 Wavelength filter unit 32 Imaging element unit 61 Objective lens 62 Slit mechanism 63 Branch device 71 Beam splitter 72 Half mirror 73, 74 Mirror 91 Reflectance calculation unit 92 Normalization index calculation unit 93 Second derivative calculation unit 94 Determination unit 101 Target detection device P, SP Pixel F Wavelength selection filter R Image sensor

Claims (7)

An imaging unit for imaging a one-dimensional image for a plurality of wavelength bands selected in advance according to the physical properties of the detection target;
A target detection apparatus comprising: a detection unit for detecting a substance in each pixel of the one-dimensional image. The imaging unit
The target detection apparatus according to claim 1, comprising a plurality of line sensors provided corresponding to the wavelength band and detecting the intensity of the corresponding wavelength band. The imaging unit further includes:
The target detection device according to claim 2, further comprising a branching device for branching received light and irradiating the plurality of line sensors.   The detection unit detects a substance in the pixel based on an evaluation value calculated by combining some or all of the intensity values for the plurality of wavelength bands in one pixel. The object detection device according to any one of the above.   The target detection apparatus according to claim 1, wherein at least some of the plurality of wavelength bands are included in an infrared region. The pixel is a combination of a plurality of pixels respectively corresponding to the plurality of wavelength bands,
The object detection apparatus according to claim 5, wherein each of the pixels includes a wavelength selection filter for selecting a corresponding wavelength band and an image sensor for receiving transmitted light from the wavelength selection filter. Imaging a one-dimensional image for a plurality of wavelength bands preselected according to the physical properties of the detection target;
Detecting a substance in each pixel of the one-dimensional image.

Images