JP2005301719A - Occupant detection apparatus and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an occupant detection apparatus and method capable of detecting occupant information regardless of the movements of occupants. <P>SOLUTION: This occupant detection apparatus includes an edge information creation means 12 using images picked up by an image pickup means 11 that picks up images of the inside of a vehicle; an edge area extraction means 13 that depends on the edge information; an edge template creation means 14 that depends on edge areas; a difference information creation means 15 for the previous and current images; a difference area extraction means 16 that depends on difference information; a difference template creation means 17 that depends on difference areas; a difference template update means 18; a template integration means 20 for an edge template and an updated difference template; and an occupant information extraction means 21 for extracting the occupant information on the basis of position information and images obtained from the integrated templates. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an occupant detection apparatus and method for a vehicle such as a vehicle.

In recent years, in order to improve the safety and comfort of a vehicle such as an automobile, it has been proposed to detect the presence or absence of a passenger, a position such as a position, and perform control according to the information. As an apparatus for detecting such information, for example, there is one described in Patent Document 1 below.
This device creates edge images of two original images captured with a time difference for the same subject, performs difference processing of the edge images, and detects the head position of the occupant based on the moved edge component . As a result, the background edge can be eliminated and information on a moving subject can be obtained efficiently.

JP 2000-113164 A

However, in the above-described conventional apparatus, since information about the occupant is detected only by the movement information of the edge component, the movement information is generated even when the occupant moves his hand near the head, and the occupant state And misjudgment of head position. In addition, since it is assumed that the occupant always moves slightly due to the movement of the vehicle and other natural movements, the occupant information can be obtained by excluding the background edge when the occupant is moving. . However, detection is impossible when the occupant does not move or in the initial state.
The present invention has been made in view of the above points, and an object thereof is to provide an occupant detection apparatus and method capable of detecting occupant information regardless of the movement of the occupant.

  In order to solve the above-described problems, the present invention provides an occupant detection device that detects occupant information in a vehicle interior, and includes edge information using an image capturing unit that acquires an image including thermal image information in the vehicle interior, and the acquired image. Edge information creating means for creating an edge area extracting means for extracting an edge area surrounded by pixels satisfying a predetermined condition from the created edge information, and setting a predetermined value to the extracted edge area, An edge template creating means for creating an edge template by setting another predetermined value in a region other than the above, a difference information creating means for creating difference information by taking a difference between the previous image and the current image, A difference area extraction unit that extracts a difference area surrounded by pixels that satisfy a predetermined condition from the created difference information, and a predetermined value is set in the extracted difference area. A difference template creating means for creating a difference template by setting another predetermined value in a region other than the above, a difference template updating means for updating the difference template, and a memory for storing the acquired image and the updated difference template A template for creating an integrated template by extracting a common area between the means and the edge template and the updated difference template, setting a predetermined value in the common area, and setting another predetermined value in the other area And an occupant information extracting unit that extracts occupant information based on the position information acquired by the template integrating unit and the temperature information from the acquired image.

  ADVANTAGE OF THE INVENTION According to this invention, the passenger | crew detection apparatus and method which can detect passenger | crew information irrespective of a passenger | crew's movement can be provided.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In the drawings described below, components having the same function are denoted by the same reference numerals, and repeated description thereof is omitted.
FIG. 1 is a configuration diagram illustrating hardware of an occupant detection device according to an embodiment of the present invention. As shown in FIG. 1, the occupant detection device according to the present embodiment includes an IR camera (infrared camera) 1 and a microcomputer 2. Subsequent devices include a monitor 3, an airbag control device 4, and an air conditioner control device 5. Is connected.

FIGS. 2A and 2B are diagrams showing a camera mounting state in the present embodiment, where FIG. 2A is a side view and FIG. 2B is a top view.
The IR camera 1 is installed in, for example, a map lamp unit or a center cluster unit of a vehicle, captures a thermal image in the vehicle compartment, and outputs temperature information. Further, one IR camera 1 may be installed to detect the occupants of the entire vehicle, or a plurality of IR cameras 1 may be used for each occupant. In the present embodiment, as shown in FIG. 2, the case where the right side of the vehicle is imaged with the IR camera 1 installed in the center cluster portion of the vehicle will be described. As temperature information output from the IR camera 1, there may be a case where an analog signal obtained by an IR detection element (heat detection element) in the IR camera 1 is represented by, for example, two-dimensional information digitized from 0 to 255. In this embodiment, the two-dimensional information obtained by temperature conversion of the digital value of 0 to 255 with the temperature conversion constant is used as thermal image information.

  FIG. 3 is a diagram showing an imaging region in the present embodiment. In FIG. 3, 30 is a driver's seat occupant, 30a is the head of the driver's seat occupant 30, 31 is a side glass, 32 is a rear glass, and 33 is a ceiling.

  The microcomputer 2 processes the thermal image information acquired by the IR camera 1 to identify the seat where the driver's seat occupant 30 of FIG. 3 is present, that is, the presence or absence of the occupant for each seat. At the same time, the position of the head 30a and the representative temperature of the face are detected from the surface temperature distribution of the face of the head 30a of the driver seat occupant 30 in FIG.

Then, the monitor 3 as a subsequent device displays the presence / absence of an occupant for each seat, the head position, the representative temperature of the face, and the thermal image information itself, which are outputs of the microcomputer 2.
The airbag control device 4 controls the deployment of the airbag based on the presence / absence of the occupant for each seat and the position of the head, which are outputs of the microcomputer 2, for example, not opening the airbag of the seat without the occupant. .
Further, the air conditioner control device 5 performs, for example, comfortable air conditioning control using the surface temperature of the occupant in the air conditioner control based on the presence / absence of the occupant for each seat, the head position, and the representative temperature of the face, which are the outputs of the microcomputer 2. Do.

  FIG. 4 is a functional block diagram functionally expressing the occupant detection device of the present embodiment. In FIG. 4, 11 is an image pickup means for acquiring an image including thermal image information representing the temperature distribution in the passenger compartment. Reference numeral 12 denotes edge information creating means for creating edge information using the image acquired by the imaging means 11. Reference numeral 13 denotes an edge area extracting means for extracting an edge area surrounded by pixels satisfying a predetermined condition from the edge information created by the edge information creating means 12. 14 is an edge template creation in which an edge template is created by setting a predetermined value for the edge area based on the edge area extracted by the edge area extracting means 13 and setting another predetermined value for the other area. Means. Reference numeral 15 denotes difference information creation means for creating difference information by taking the difference between the previous image acquired by the imaging means 11 and the current image. Reference numeral 16 denotes difference area extraction means for extracting a difference area surrounded by pixels satisfying a predetermined condition from the difference information created by the difference information creation means 15. 17 is a difference template creation for creating a difference template based on the difference area extracted by the difference area extraction means 16 by setting a predetermined value in the difference area and setting another predetermined value in the other area. Means. Reference numeral 18 denotes difference template update means for updating the difference template. Reference numeral 19 denotes storage means (memory) for storing the image acquired by the imaging means 11 and the difference template updated by the difference template update means 18. 20 extracts a common area between the edge template created by the edge template creation means 14 and the difference template updated by the difference template update means 18 as a new area of interest, and sets a predetermined value in the common area. This is a template integration unit that creates an integrated template based on edge information and difference information by setting another predetermined value in the other area. Reference numeral 21 denotes an occupant information extraction unit that calculates and outputs occupant information based on the position information acquired by the template integration unit 20 and the temperature information from the image acquired by the imaging unit 11.

  When the functional blocks shown in FIG. 4 are compared with those expressed as hardware shown in FIG. 1, the IR camera 1 in FIG. 1 becomes the imaging means 11 in FIG. 4, and the microcomputer 2 in FIG. Information creation means 12, edge region extraction means 13, edge template creation means 14, difference information creation means 15, difference area extraction means 16, difference template creation means 17, difference template update means 18, storage means 19, template integration means 20, Each corresponds to the occupant information extraction means 21.

  FIG. 5 is a functional block diagram showing the contents of the template integration means 20 shown in FIG. That is, the template integration unit 20 includes a common area extraction unit 201, a label numbering unit 202, and a label number updating unit 203 as shown in FIG. The common area extracting unit 201 extracts a common area between the edge template created by the edge template creating unit 14 in FIG. 4 and the difference template updated by the difference template updating unit 18 in FIG. The label numbering means 202 attaches an individual label number to the common area. Based on the label number assigned by the label numbering means 202, the label number updating means 203 determines that the areas in which the horizontal component of each common area is common in the vertical direction are the same target, and the same label Update the label number to be a number. Note that the occupant information extraction unit 21 in FIG. 4 processes, as one piece of target information, the position in the xy direction with the same label number when referring to the temperature information from the image acquired by the imaging unit 11. It is like that. Hereinafter, each means will be described in detail.

FIG. 6 is a flowchart showing the flow of processing in the occupant detection device and method of the present embodiment.
Hereinafter, the operation in the present embodiment will be described with reference to FIG.
As shown in FIG. 6, when the occupant detection process is started, first, in step 1 (S1), an initialization process for initializing counters and flags used for the subsequent processes is executed. For example, as the initial value of the difference template, the template value is set to 1 only for the area around the headrest position of the seat, and the other values are set to 0. The difference template only needs to be able to understand information on the candidate position where the occupant's head is supposed to exist in the two-dimensional information. If the difference between the area that is supposed to exist and the area that does not exist is known, the difference template has a value of 1 or 0. It is not limited.

Next, in step 2 (S2), an image including thermal image information indicating the temperature distribution in the passenger compartment is acquired by the IR camera 1 of FIG. The two-dimensional thermal image information acquired at this time becomes an image having a higher density value as the surface temperature is higher. In order to obtain such information, for example, all the passengers in the passenger compartment may be shown in one image, or each person may be shown in the IR camera 1 set for each seat. Good. In the present embodiment, the case where the IR camera 1 is installed at the position shown in FIG. 2 and the driver's seat occupant 30 is copied by one IR camera 1 as shown in FIG. 3 will be described as an example.
FIG. 7 is a diagram for explaining the movement of the occupant in the present embodiment. (A) shows the temperature image in the passenger compartment at time T-1, and (b) shows the case where the head is tilted at time T. A case where the driver's seat occupant 30 tilts his / her head as shown by an arrow at time T as shown in FIG. 7B with respect to the state at the previous time T-1 as shown in FIG. 7A will be described. When processing an image including other passengers, for example, if the image is divided into regions where the passengers of each seat are considered to be present from the installed state of the IR camera 1 and extraction is performed for each region, the subsequent processing is performed. The same processing as in this embodiment may be performed.

Next, in step 3 (S3), edge information is created using the two-dimensional thermal image information captured in step 2 (S2) (edge information creation means 12 in FIG. 4). For example, a vertical edge image is created by extracting a density gradient in the horizontal direction using a Sobel operator. FIG. 8 is a diagram showing an example of creating an edge template in the present embodiment. (A) shows edge processing (edge information), (b) shows a region template (edge region), and (c) shows an edge template.
FIG. 8A shows an example of edge information when the input image is FIG. 7B. This edge information has signs of + edge and -edge according to the density gradient. In the present embodiment, vertical edge information is created, but of course, horizontal edge information may be used in combination. That is, the edge information includes at least vertical edge information.

  Next, in step 4 (S4), based on the edge information created in step 3 (S3), a region of interest surrounded by pixels that satisfy a predetermined condition, that is, an edge region sandwiched between the edges of interest is illustrated. Extraction is performed as shown in FIG. 8B (edge region extracting means 13 in FIG. 4). Note that the combination of the signs of the edges of interest changes depending on the background temperature. That is, the edge region extraction unit 13 and the difference region extraction unit 16 in FIG. 4 extract the average temperature of a predetermined region of the thermal image information acquired by the imaging unit 11 as the background temperature, and extract each region based on the background temperature. The predetermined condition for each is changed. Therefore, an average temperature of a region where no occupant is supposed to be present is calculated, and when the average temperature is higher than a predetermined value Th, it is determined that the background temperature is higher than the surface temperature of the occupant, and the predetermined value Th In the following cases, the background temperature is judged to be lower than the passenger's surface temperature. Then, the edge information of the occupant can be surely extracted even if the background temperature changes by switching the combination of the edge codes depending on whether the background is low or high.

  Next, in step 5 (S5), based on the edge region extracted in step 4 (S4), a predetermined value is set in the edge region, and another predetermined value is set in the other regions. An edge template is created (edge template creation means 14 in FIG. 4). For example, an edge template as shown in FIG. 8C is created by setting the value of the position corresponding to the extracted edge region to 1, for example, and setting the others to 0. This edge template only needs to be able to know the information of the candidate position where the occupant's head is supposed to exist in the two-dimensional information. If the difference between the area that seems to exist and the area that does not exist is known, the value is 1 or 0. It is not limited to.

Next, in step 6 (S6), difference information is created by taking the difference between the previous image acquired by the imaging unit 11 and the current image (difference information creating unit 15 in FIG. 4). That is, motion information extraction processing is performed on the two-dimensional thermal image information captured in step 2 (S2). This is a process for obtaining a difference between two pieces of thermal image information at different times, that is, a process for extracting only a portion where a motion has occurred by the difference process to obtain difference information. Difference processing between the previous thermal image information stored in the storage unit 19 and the two-dimensional thermal image information captured in step 2 (S2) is performed. When the output variation range of the temperature information of the IR detection element is ± Td, a portion having a temperature change of + Td or more or −Td or less is set as motion information. That is, a position where the temperature information has changed by + Td or −Td which is a predetermined amount from the previous process is extracted. FIG. 9 is a diagram showing an example of creating a difference template in the present embodiment. (A) shows difference processing (difference information), (b) shows an area template (difference area), and (c) shows a difference template.
By such a process, for example, as shown in FIG. 7B, when the driver's seat occupant 30 tilts the head 30a (neck), the position where the temperature change is large is as shown in FIG. 9A. Movement information can be obtained. In addition to the position information of the pixel where the temperature change is large, this motion information includes temperature change information indicating whether the temperature at the position has changed to a low temperature or a high temperature, depending on the sign of the value of the pixel. Yes. Such motion information is extracted, and the process proceeds to the next step.

  Next, in step 7 (S7), based on the difference information created in step 6 (S6), a difference area sandwiched between the information of interest is extracted (difference area extraction means 16 in FIG. 4). That is, a difference area surrounded by pixels satisfying a predetermined condition is extracted from the created difference information. For example, as shown in FIG. 7B, when the head 30a of the driver's seat occupant 30 moves, the temperature change on both sides of the head 30a extracted as the difference information has a higher temperature change on one side. In some cases, the temperature change on the opposite side is low. Using this feature, if there is motion information as shown in FIG. 9A as a candidate for the position where the head 30a exists, a difference area as shown in FIG. 9B is extracted. . At the time of extracting the difference area, if it is determined in step 7 (S7) that the background temperature is lower, a higher temperature area is extracted by scanning from the smaller x-coordinate. If it is determined that the temperature is higher, scanning is performed from the smaller x coordinate to extract a low temperature region.

  Next, in step 8 (S8), based on the difference area extracted in step 7 (S7), a predetermined value is set in the difference area, and another predetermined value is set in the other areas. A template is created (difference template creation means 17 in FIG. 4). For example, by setting the position value corresponding to the extracted area to 1, for example, and setting the others to 0, a difference template as shown in FIG. 9C is created. This difference template only needs to be able to know information on the candidate position where the occupant's head is supposed to exist in the two-dimensional information. If the difference between the area that is supposed to exist and the area that does not exist is known, the difference template has a value of 1 or 0. It is not limited to.

Next, in step 9 (S9), the difference template is updated based on the amount of movement of the occupant (difference template updating means 18 in FIG. 4). FIGS. 10A to 10C are diagrams showing an example of updating the difference template in the present embodiment.
The basic idea in this process is that if the amount of motion information extracted in step 6 (S6) is larger than a predetermined value Tm, the difference template extracted in step 8 (S8) is used as a new difference template, When the amount of information is less than or equal to the predetermined value Tm, the current difference template (FIG. 10) extracted in step 8 (S8) is added to the previous difference template (corresponding to FIG. 10 (a)) stored in the storage means 19. (Corresponding to (b)) is added (OR operation of both templates) to form a new difference template (corresponding to FIG. 10C). In this embodiment, the update of the difference template based on the amount of motion information is described. However, in addition to this process, the difference template update method may be changed according to the position of the extracted difference information. It doesn't matter. In other words, whether to add to the previous difference template may be determined based on the position of the motion information. In this way, the difference template is updated by adding the information of the current difference template to the previous difference template created in step 8 (S8) according to the amount of difference information created in step 6 (S6). This is performed by selecting either the processing to be performed or the processing for completely updating the current difference template. Then, the updated difference template is stored in the storage means 19 and the process proceeds to the next step.

Next, in step 10 (S10), the edge template created in step 5 (S5) and the difference template created in step 9 (S9) are integrated (template integration means 20 in FIG. 4). In other words, a common area between the created edge template and the updated difference template is extracted as a new area of interest, a predetermined value is set in the common area, and another predetermined value is set in the other areas To create an integrated template based on the edge information and the difference information. FIG. 11 is a diagram showing an example of template integration processing in the present embodiment. (A) is an edge template, (b) is a difference template, and (c) is an integrated template.
For example, when an edge template is extracted as shown in FIG. 11A and a difference template is extracted as shown in FIG. 11B, a common area is shown by performing a logical product operation of both templates. It is extracted as an occupant position reference integrated template shown in FIG. 11C (common area extracting means 201 in FIG. 5).

Next, in step 11 (S11), the occupant position reference template for labeling the common area created in step 10 (S10) is labeled (label numbering means 202 in FIG. 5). FIGS. 12A and 12B are diagrams showing an example of label update in the present embodiment.
In the example shown in FIG. 11 (c), there is no problem because the extracted area is one. For example, when the occupant is wearing glasses, the forehead area is shown in FIG. 12 (a). And the lower area of the face is separated. In such a case, a different label number is attached to each area.

  Next, in step 12 (S12), a relabeling process is performed on the area numbered in step 11 (S11) (label number updating unit 203 in FIG. 5). For example, when the occupant is wearing glasses, the forehead region and the lower region of the face are separated as shown in FIG. 12A as described above. Since it is occupant face information, it is desirable to treat them as the same. Accordingly, as shown in FIG. 12 (b), it is determined that regions having a common part in the horizontal direction of each common region are the same target, and regions having different label numbers are the same. Use a label number. In this way, based on the assigned label number, it is determined that the regions in which the horizontal component of each common region is common in the vertical direction are the same target, and the label number is updated so that the same label number is obtained. . With the above processing, a label numbered occupant position reference template is created and the process proceeds to the next step.

  Next, in step 13 (S13), occupant information is calculated and output based on the position information acquired in step 10 (S10) and the temperature information from the image acquired in step 2 (S2). That is, based on the occupant position reference integrated template shown in FIG. 11C created in step 12 (S12), temperature information is referred to from the two-dimensional thermal image information obtained in step 2 (S2). Presence / absence, head position, and representative face temperature are extracted as occupant information (occupant information extraction means 21 in FIG. 4). Specifically, the occupant position reference integrated template is scanned, the temperature at the pixel position having a value of 1 for each label number is read from the two-dimensional thermal image information, and the average temperature for each label number is calculated. Then, the occupant position reference integrated template is scanned again, the temperature of the pixel position having a value of 1 is read from the two-dimensional thermal image information for each label number, and temperature information higher than this average value is extracted for occupant information. Information. The average value of the information extracted in this process is obtained. For example, if the motion information is in the past and the average temperature is within a predetermined range, it is determined that there is an occupant. to decide. Further, when it is determined that there is an occupant, for example, the center of gravity position of the occupant extraction information distribution is extracted as the occupant head position, and the average value of the occupant extraction information is extracted as the representative temperature of the occupant. Such occupant information is extracted and the process proceeds to the next step. In the extraction of occupant information, at least one of the presence / absence of the occupant, the position of the occupant, and the face temperature of the occupant is extracted.

  Finally, in step 14 (S14), the occupant information extracted in step 13 (S13) is output to subsequent devices such as the monitor 3, the airbag control device 4, and the air conditioner control device 5, and then step 2 (S2). Then, the same processing is repeated thereafter.

  As described above, the occupant detection apparatus and method according to the present embodiment acquires an image including thermal image information representing the temperature distribution in the passenger compartment, creates edge information using the image, and includes the edge information. A predetermined value is set in the edge region surrounded by pixels satisfying the predetermined condition, and another predetermined value is set in the other region to create an edge template, and the acquired previous image and current image A difference template is created by taking difference and setting a predetermined value in a difference area surrounded by pixels satisfying a predetermined condition in the difference information and setting another predetermined value in other areas. This difference template is updated, a predetermined value is set in the common area of the edge template and the updated difference template, and another predetermined value is set in the other areas to set the edge information and the difference information. Create an integrated template based on, based on the temperature information from the position information and the image in the integrated template, and outputs to calculate the occupant information. Whereas the conventional technique uses two original images at different times to create an edge image and then creates a differential image of the edge image to perform position detection, in the present invention, the two images at different times are used. The original image is used to create a difference image between the edge image and the original image, and after the consistency is determined, position detection is performed. In this way, the region where the occupant is present is specified from both the edge processing for extracting the boundary information and the difference processing for extracting the motion information, and the occupant information is detected from the captured image information. It is possible to reliably detect the presence or absence of an occupant without being affected by the movement other than. Further, even when there is no initial state where no motion information is generated or when there is no short-time motion information, the position can be specified by edge processing each time, so that the presence or absence of a passenger can be reliably detected. That is, instead of holding the previous output when there is no motion information, the result is obtained every time. Based on the presence / absence of an occupant output signal, it is possible to perform on / off control of the airbag deployment and control of the air-conditioning outlet based on the presence / absence of the occupant.

In addition, in each extracted area, an area that overlaps when viewed in the vertical direction is treated as the same target information, so the face area of the occupant wearing glasses, for example, is divided into two vertically Even if it is done, it can be extracted as the same information.
Further, since the occupant's head position is calculated and output, it is possible to perform airbag deployment force control and air conditioning outlet and outlet direction control based on the occupant's head position.
Further, since the thermal image information is acquired and the occupant information is extracted based on the temperature information, it is possible to easily determine the occupant and the luggage.
In addition, the temperature change in the passenger compartment is detected based on the background temperature distribution, and the combination of edge codes at the time of boundary detection is changed. Regardless of whether it is high or low, it is possible to stably detect the position and movement of the occupant.
Furthermore, since the representative temperature of the occupant is calculated and output, it is possible to control the temperature of the air conditioning based on the occupant's feeling of temperature.

Because of the effects as described above, an area where an occupant is present is specified by edge processing for extracting the boundary of the object and difference processing for extracting the motion of the object, and the occupant is based on image information at a position corresponding to the area. By extracting the state, it is possible to provide an apparatus and a method that can detect the state of the presence, position, temperature, etc. of the occupant regardless of the movement of the occupant.
The embodiment described above is described in order to facilitate understanding of the present invention, and is not described in order to limit the present invention. Therefore, each element disclosed in the above embodiment includes all design changes and equivalents belonging to the technical scope of the present invention.

It is a block diagram which shows the hardware of the passenger | crew detection apparatus of embodiment of this invention. (A) is a side view which shows the attachment state of the camera in embodiment of this invention, (b) is a top view. It is a figure which shows the imaging area in embodiment of this invention. It is a functional block diagram functionally expressing the occupant detection device of the embodiment of the present invention. It is a functional block diagram which shows the content of the template integration means of FIG. It is a flowchart which shows the flow of a process in the passenger | crew detection apparatus and method of embodiment of this invention. It is a figure for demonstrating a passenger | crew's movement in embodiment of this invention, (a) is a temperature image in a vehicle interior in the time T-1, (b) is a figure which shows the case where the head is inclined at the time T. . It is a figure which shows the creation example of the edge template in embodiment of this invention, (a) is an edge process, (b) is an area | region template, (c) is a figure which shows an edge template. It is a figure which shows the example of creation of the difference template in embodiment of this invention, (a) is a difference process, (b) is an area | region template, (c) is a figure which shows a difference template. (A)-(c) is a figure which shows the example of an update of the difference template in embodiment of this invention. It is a figure which shows the template integration process example in embodiment of this invention, (a) is an edge template, (b) is a difference template, (c) is a figure which shows an integrated template. (A), (b) is a figure which shows the example of a label update in embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... IR camera 2 ... Microcomputer 3 ... Monitor 4 ... Airbag control apparatus 5 ... Air-conditioner control apparatus 11 ... Imaging means 12 ... Edge information creation means 13 ... Edge area extraction means 14 ... Edge template creation means 15 ... Difference information creation means 16 ... Difference area extraction means 17 ... Difference template creation means 18 ... Difference template update means 19 ... Storage means 20 ... Template integration means 21 ... Passenger information extraction means 30 ... Driver's seat occupant 30a ... Driver's seat occupant's head 31 ... Side glass 32 ... rear glass 33 ... ceiling 201 ... common area extraction means 202 ... label numbering means 203 ... label number updating means

Claims (9)

In an occupant detection device that is installed in the passenger compartment and detects occupant information in the passenger compartment,
Imaging means for acquiring an image including thermal image information representing a temperature distribution in the passenger compartment;
Edge information creating means for creating edge information using an image acquired by the imaging means;
Edge region extraction means for extracting an edge region surrounded by pixels satisfying a predetermined condition from the edge information created by the edge information creation means;
Edge template creation means for creating an edge template by setting a predetermined value for the edge area based on the edge area extracted by the edge area extraction means and setting another predetermined value for the other area; ,
Difference information creating means for creating difference information by taking the difference between the previous image acquired by the imaging means and the current image;
A difference area extracting means for extracting a difference area surrounded by pixels satisfying a predetermined condition from the difference information created by the difference information creating means;
A difference template creating means for creating a difference template by setting a predetermined value in the difference area based on the difference area extracted by the difference area extracting means and setting another predetermined value in the other area; ,
Difference template update means for updating the difference template;
Storage means for storing the image acquired by the imaging means and the difference template updated by the difference template update means;
A common area between the edge template created by the edge template creation means and the difference template updated by the difference template update means is extracted, and a predetermined value is set in the common area. Template integration means for creating an integrated template based on edge information and difference information by setting a predetermined value;
Occupant information extraction means for calculating and outputting occupant information based on position information acquired by the template integration means and temperature information from the image acquired by the imaging means;
An occupant detection device comprising:   The occupant detection device according to claim 1, wherein the edge information includes at least vertical edge information. In the edge area extraction means and the difference area extraction means,
The average temperature of a predetermined area of the thermal image information acquired by the imaging means is extracted as a background temperature, and the predetermined condition for extracting each area is changed based on the background temperature, respectively. 2. The occupant detection device according to 2. The difference template update means includes:
A process of adding information on the current difference template to the previous difference template created by the difference template creation unit according to the amount of difference information created by the difference information creation unit,
4. The occupant detection apparatus according to claim 1, wherein the difference template is updated by selecting any of the processes for updating all of the current difference template. The template integration means includes
Label numbering means for assigning individual label numbers to the common area;
Based on the label number assigned by the label numbering means, the horizontal component of each common area is determined to be the same object in the vertical direction, and the label number is set so as to be the same label number. Label updating means for updating,
The occupant information extracting means includes
5. When referring to temperature information from an image acquired by the imaging means, a position with the same label number is processed as one piece of target information. Occupant detection device. The occupant information extracting means includes
6. The occupant detection device according to claim 1, wherein presence / absence of an occupant is extracted. The occupant information extracting means includes
The occupant detection device according to any one of claims 1 to 6, wherein a position of the occupant is extracted. The occupant information extracting means includes
The occupant detection device according to any one of claims 1 to 7, wherein the occupant face temperature is extracted. In the passenger detection method for detecting passenger information in the passenger compartment,
An imaging step of acquiring an image including thermal image information representing a temperature distribution in the passenger compartment;
Using the image acquired by the imaging step, an edge information creation step for creating edge information;
An edge region extraction step for extracting an edge region surrounded by pixels satisfying a predetermined condition from the edge information created by the edge information creation step;
An edge template creation step of creating an edge template by setting a predetermined value in the edge region based on the edge region extracted in the edge region extraction step and setting another predetermined value in the other region; ,
A difference information creating step of creating difference information by taking the difference between the previous image acquired by the imaging step and the current image;
From the difference information created by the difference information creation step, a difference region extraction step for extracting a difference region surrounded by pixels that satisfy a predetermined condition;
A difference template creating step of creating a difference template by setting a predetermined value in the difference area and setting another predetermined value in the other area based on the difference area extracted by the difference area extracting step; ,
A difference template update step for updating the difference template;
A common area between the edge template created by the edge template creation process and the difference template updated by the difference template update process is extracted, and a predetermined value is set in the common area. A template integration step of creating an integrated template based on edge information and difference information by setting a predetermined value;
An occupant information extraction step of calculating and outputting occupant information based on the position information acquired by the template integration step and the temperature information from the image acquired by the imaging step. .

Images