JP2009067240A - Departure notification device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a departure notification device capable of detecting a red region correctly and capable of preventing a driver from overlooking a notification that a vehicle should depart. <P>SOLUTION: Image data of RGB value formed by an imaging part 3 are converted to image data of HSV value, and a brake lamp of another vehicle (a vehicle ahead) stopping ahead of a vehicle 5, for example, is detected as a red region on the basis of the converted image data. Because red and yellow are more easily distinguished in HSV color space compared, for example, with RGB color space, mistaking yellow of a turn indicator for red of the brake lamp is inhibited. Then, on the basis of result of detection of the red region, whether the brake lamp under lighting has been put off or not is judged. If putting off of the brake lamp (that is, the starting motion of the vehicle ahead) has been confirmed, a voice output part 4 outputs that the vehicle should depart. The driver who has been notified that the vehicle should depart by voice instead of by an image can start the vehicle 5 at just the right time. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a departure notification device for notifying that a stopped vehicle should start at a timing when it should start.

Conventionally, when a vehicle is stopped at an intersection or railroad crossing, it is stopped such as a change from a red signal to a green signal or a start operation of another vehicle stopped in front of the vehicle (hereinafter referred to as a forward vehicle). There has been proposed a departure notification device that detects a change that serves as an index of timing at which a vehicle should depart (see Patent Document 1).
The in-vehicle display device disclosed in Patent Document 1 images the front of the vehicle, detects a red region of the traffic light from the captured image, and converts an image other than the detected red region into monochrome, It is configured to be displayed on a part of a navigation screen of an in-vehicle television or a car navigation system.

By using such an in-vehicle display device, a driver who is viewing a television screen or a navigation screen can easily change from a red signal to a green signal by looking at an image of a traffic light displayed on a part of the screen. And start the vehicle at an appropriate time.
JP 2006-115376 A

However, in the in-vehicle display device disclosed in Patent Document 1, a change in signal cannot be recognized unless the driver actively views an image of the traffic light. For this reason, even if the image of the traffic light is displayed, the driver may not be aware of the change from the red signal to the green signal, and the departure of the vehicle may be delayed. If the departure of the vehicle is delayed, the smooth flow of traffic may be hindered or the driver may panic due to pressure from the following vehicle.
In order to prevent such a problem, it is conceivable to prompt the driver to view the image, but in this case, the driver feels a burden and the convenience of the driver is impaired.

By the way, the brake lamp lights up in red when the vehicle is stopped, and the brake lamp is turned off when the vehicle departs. Therefore, in order to detect the start operation of the vehicle ahead, it is conceivable to detect a change from turning on the brake lamp to turning it off.
A conventional departure notification device such as the in-vehicle display device of Patent Document 1 obtains a red region using image data of RGB (Red, Green, Blue) values obtained by imaging the front of the vehicle. Specifically, if the R value of the target pixel is larger than a predetermined value, and R value> G value and R value> B value, the target pixel is regarded as a red pixel.
However, even for a yellow pixel, it is conceivable that the R value is larger than a predetermined value, and R value> G value and R value> B value (for example, R = 255, G = 200, B = 20). For this reason, a yellow turn signal may be erroneously detected instead of the brake lamp.

  The present invention has been made in view of such circumstances, and its main object is to obtain a plurality of image data of HSV (Hue, Saturation, Value) values by converting image data picked up while the vehicle is stopped, and to obtain the obtained image. A red area of a predetermined detection target is detected from each of the data, and whether or not a predetermined change has occurred in the detection target is determined based on the detection result, and if a predetermined change has occurred, a voice or An object of the present invention is to provide a departure notification device capable of accurately detecting a red region and appealing to a sense other than vision when notifying that a departure is to be made by adopting a configuration for notification by vibration.

  Another object of the present invention is to determine whether or not captured image data includes traffic signal image data. If included, a red signal is detected, and if not, a brake lamp is turned on. An object of the present invention is to provide a departure notification device that can accurately and easily determine the timing at which a vehicle should depart by detecting the presence or absence of image data of a brake lamp.

  Another object of the present invention is to provide a departure notification device capable of reducing arithmetic processing by adopting a configuration that binarizes a red region to be detected and other regions.

  Still another object of the present invention is to capture a captured image by adopting a configuration in which image processing is continued while warning that it is impossible to notify when a red region to be detected does not exist. Warning that it is impossible to notify that the vehicle should depart when there is no image of both the traffic light and the vehicle ahead, or when the red area is detected as a different color area due to the light beam. Thus, it is intended to provide a departure notification device that prompts the driver to measure the departure timing, and that can determine the departure timing when the state of light changes and a red region is detected.

  The departure notification device according to the first aspect of the present invention includes an imaging unit that generates image data in which the front of the vehicle is continuously or intermittently imaged in a time series while the vehicle is stopped. In the departure notification device for notifying that the image should be generated, the image conversion unit that converts the image data generated by the imaging unit into the image data of the HSV value, and the predetermined detection based on the image data converted by the image conversion unit Based on the detection results of the red detection means for detecting the target red region and the red detection means for each of a plurality of image data captured in time sequence, there is a change that serves as an index of the timing to depart. A change determination unit that determines whether or not the detection target has occurred, and a departure notification unit that notifies, by voice or vibration, that the vehicle should depart when it is determined that the change determination unit has occurred. And features.

  The departure notification device according to a second aspect of the present invention further includes signal determination means for determining whether the image data converted by the image conversion means or the image data generated by the imaging section includes image data of a traffic light. When it is determined that the signal determination unit is included, the red detection unit detects a red signal in which a traffic light is lit as a red region to be detected based on image data of the traffic signal. When the signal determining means determines NO, the red detecting means displays a brake lamp that is lit on a vehicle parked in front of the vehicle as a red area to be detected in an image within a predetermined range. The detection is based on data.

  In the departure notification device according to a third aspect of the present invention, the red color detecting means has a pixel having an H value in a predetermined range indicating red, an S value and a V value greater than each predetermined value indicating a bright color, and other than the pixel It is characterized in that the red region to be detected is detected by discriminating it from the two pixels in binary.

  A departure notification device according to a fourth aspect of the present invention is the presence determination means for determining whether or not a predetermined number or more of pixels in the red region to be detected are present in the detection result of the red detection means relating to one image data; When the determination means determines NO, it further comprises an impossible warning means that warns that it is impossible to notify that the vehicle should depart, and the red color detection means, regardless of the determination result of the presence determination means, The red region to be detected is detected based on image data after the image data.

In the first invention, an imaging unit is provided, and further an image conversion unit, a red detection unit, a change determination unit, and a departure notification unit are provided.
The imaging unit generates, in a time series, image data obtained by continuously or intermittently imaging the front of the vehicle while the vehicle on which the departure notification device is mounted stops at, for example, an intersection or a railroad crossing. The image data generated by the imaging unit is image data other than HSV image data, such as RGB value image data, YUV value image data, and NTSC project image data. In the following, for simplicity, image data of RGB values is illustrated.
The RGB value image data includes, for each pixel, R value indicating red, G value indicating green, and B value indicating blue. Each of the R value, the G value, and the B value has an integer value of 0 ≦ R, G, B ≦ 255, for example.

  The image conversion means converts RGB value image data into HSV value image data. The image data of the HSV value has, for each pixel, H value indicating hue, S value indicating saturation, and V value indicating lightness. For example, the H value has a value of 0 ° ≦ H ≦ 360 °, the S value has a value of 0 ≦ S ≦ 1, and the V value has a value of 0 ≦ V ≦ 255. In this case, when the H value is in the vicinity of 0 ° and 360 ° (for example, 0 ° ≦ H <20 ° and 340 ° <H ≦ 360 °), the pixel is red, and when both the S value and the V value are high, A bright red color (for example, a lit red signal or a brake lamp color) is displayed, and a dark red color (for example, an unlit red signal or a brake lamp color) is indicated when the S value and the V value are low. A yellow pixel such as a blinker color has an H value in the vicinity of 30 °, so that a red pixel and a yellow pixel can be easily and accurately distinguished.

The red detection means detects a predetermined red region to be detected based on the image data converted by the image conversion means.
Here, the conversion of the image data by the image conversion unit is executed for each of a plurality of image data generated in time series by the imaging unit, and the detection of the red region by the red detection unit is performed by a plurality of images changed by the image conversion unit. It is executed for each of the image data.

  The change determination means generates a change in the detection target as an index of the timing at which the stopped vehicle should depart based on the detection result of the red detection means regarding each of a plurality of image data captured in time sequence. It is determined whether or not. Here, when the detection target is, for example, a traffic light, the change that is an indicator of the timing to depart is the turn-off of the lit red signal, and when the detection target is a forward vehicle, the change is an indicator of the timing to depart The change that becomes is the extinction of the lit brake lamp.

  When the red signal or the brake lamp is lit, the pixel group detected as a bright red area in the image data captured earlier is detected as a bright red area in the image data captured later, so these two images There are no major differences in the data. However, when the red signal or the brake lamp changes from lighting to extinguishing, the pixel group detected as a bright red region in the image data captured earlier is detected as a dark red region in the image data captured later. A large difference occurs between the two image data. That is, for example, by comparing the image data captured earlier with the image data captured later, it is easily determined whether or not a change that is an index of the timing to depart has occurred in the detection target.

  When the change determination unit determines that a change that is an index of the timing to depart has occurred in the detection target, the departure notification unit notifies the user that the vehicle should depart by voice or vibration. Since the notification by voice or vibration is passively perceived by the driver's sense of hearing or touch, there is no need for the driver to actively perceive visually unlike the notification by image.

According to the second aspect of the present invention, signal determining means is further provided.
The signal determination unit uses, for example, a template matching method, and the image data of the HSV value converted by the image conversion unit or the image data of the RGB value generated by the imaging unit, that is, the image data of the entire landscape captured by the imaging unit (hereinafter referred to as “image data”). It is determined whether or not the image data of the traffic light is included in the landscape image data).
When the signal determination unit determines that the image data of the traffic signal is included in the landscape image data, the red detection unit detects the red region to be detected based on the image data of the traffic signal included in the landscape image data. To detect. Here, as the red region to be detected, a red signal in which a traffic light is lit is used. The extinguishing of the red signal that is lit means that the blue signal is lit, and means that the timing to depart from has arrived.

Further, there is no problem even if the image data other than the image data of the traffic light is ignored. Since this image data is omitted and the red area detection calculation is performed, the calculation process can be reduced and the calculation time can be shortened. .
On the other hand, when the signal determination unit determines that the image data of the traffic signal is not included in the scenery image data, the red detection unit detects the red region to be detected based on the image data in a predetermined range. Here, as the red region to be detected, a brake lamp that is lit by a vehicle that is stopped in front of the vehicle (that is, the preceding vehicle) is used. The extinguishing of the lit brake lamp is a start operation of the vehicle ahead, and means the arrival of the timing to depart.

  By performing the brake lamp detection calculation without determining whether or not the landscape image data includes the brake lamp image data, the calculation processing can be reduced and the calculation time can be shortened. Further, by omitting the image data other than the image data in the predetermined range and performing the red region detection calculation, the calculation processing can be further reduced and the calculation time can be further shortened.

  In the third invention, the red region to be detected and the other region are binarized. Specifically, the red detecting means has a pixel having an H value in a predetermined range indicating red and having an S value and a V value equal to or higher than each predetermined value indicating a bright color (that is, a bright red pixel), and other pixels. By distinguishing these pixels (that is, dark red, yellow, blue, green, etc.) by binary values, the detection target red region is detected.

Illuminated red signals, brake lamps, etc. are bright red areas. For this reason, the red detection means converts the image data of the HSV value into binary image data in which, for example, the value of a bright red pixel is “1” and the values of pixels other than this pixel are “0”. In other words, the color image indicated by the HSV value is converted into a monochrome image in which bright red areas such as a lit red signal and a brake lamp are indicated by white, for example, and areas other than the bright red area are indicated by black, for example. .
The detection result composed of such binary data has a small data amount. Therefore, the calculation process using the detection result is simple and the calculation time is short. Further, the storage capacity for storing the detection calculation result is reduced.

According to the fourth aspect of the present invention, there are further provided presence determination means and impossible warning means.
After detecting the red area for one image data by the red detection means, the presence determination means determines whether or not the detection target red area exists in the detection result of the red detection means for the one image data.
The impossibility warning means determines that the red region to be detected does not exist, that is, for example, when the image captured by the imaging unit does not include the image of the traffic light or the image of the preceding vehicle, If it cannot be detected as a red area due to the amount of increase or decrease, a warning is given that it is impossible to notify that the vehicle should depart.

  However, the red detection means detects the red region to be detected based on the image data after the image data subjected to the detection process, regardless of the determination result of the presence determination means. For this reason, in the previous image data, the red portion could not be detected as a red region due to the addition of light rays, but in this image data, the red portion can be detected as a red region because the condition of the light rays has changed. In this case, a red signal or a brake lamp that is lit is detected again, and it is determined whether or not the lamp is extinguished. If the lamp is extinguished, the driver is informed that the vehicle should depart.

In the case of the departure notification device according to the first aspect of the invention, the image data is processed in the HSV color space, so that erroneous recognition of red and yellow can be avoided and the red region can be detected accurately. As a result, it is possible to suppress misreporting based on misrecognition (that is, not informing that the vehicle should be departed at the timing at which departure is to be prompted, or informing the user to depart at the timing at which departure should not be prompted). .
In addition, in order to notify that the vehicle should be departed by voice or vibration, the driver can change the red signal to a green light, start the front vehicle, etc. without actively viewing images of traffic lights, front vehicles, etc. You can notice what happened. As a result, the burden on the driver can be reduced and convenience can be improved.

  Furthermore, since a driver who notices a change from a red signal to a green signal, a start operation of a preceding vehicle, etc. can start the vehicle at an appropriate timing, the traffic flow can be made smooth.

In the departure notification device according to the second aspect of the invention, if there is a traffic light in the captured scene, the detection of the red signal is performed preferentially, so that the timing for departure should be obtained accurately based on the change from turning on to turning off the red signal. be able to.
Even if the traffic light is located in the blind spot of the image pickup unit or there is no traffic light at the stop, the start operation of the vehicle ahead (specifically, the brake lamp) The timing to depart from the vehicle can be determined based on the change from turning on to turning off.

By the way, since the positional relationship between the brake lamp of the preceding vehicle and the vehicle is substantially constant, a range in which the image data of the brake lamp in the landscape image data is highly likely to be included is set in advance as a predetermined range. Can do. As a result, the process of confirming the presence or absence of the brake lamp in the captured image can be omitted, and the overall calculation time can be shortened.
In addition, since the red region is detected based on a part of the image data of the landscape image data instead of the entire landscape image data, the calculation process can be further reduced and the calculation time can be shortened.

  On the other hand, since the positional relationship between the traffic signal and the vehicle is indeterminate, the image is obtained by determining whether or not the image data of the traffic signal is included in the landscape image data without setting a predetermined range in advance. Although there is no traffic light in the scenery, it is possible to omit useless processing such as detecting a red signal. In addition, if it is confirmed that the image data of the traffic signal is included in the landscape image data, it is easy to specify the range in which the image data of the traffic signal is included. The red region can be detected based on the image data of the traffic light that is a part of the data. As a result, calculation processing can be reduced and calculation time can be shortened.

  In the case of the departure notification device according to the third aspect of the invention, it is possible to determine whether or not a change that is an indicator of timing to leave has occurred in the detection target using the binarized detection result. Can be reduced. In addition, since a plurality of detection results are stored, it is not necessary to provide a large-capacity storage unit.

  In the departure notification device according to the fourth aspect of the invention, when there is no traffic light or a preceding vehicle in the scene imaged by the imaging unit, or when the red area cannot be detected due to the light beam, it is not possible to notify that the vehicle should depart In addition, it is possible to warn the driver that notification is impossible and to prompt the driver to measure the departure timing. However, since the image processing is continued, for example, when the state of the light beam changes and a red region is detected, it is possible to notify that the vehicle should depart as usual.

  Hereinafter, the present invention will be described in detail with reference to the drawings illustrating embodiments thereof.

FIG. 1 is a block diagram showing a main configuration of a departure notification device according to an embodiment of the present invention.
In the figure, 1 is a departure notification device, and the departure notification device 1 is provided in a vehicle 5. The departure notification device 1 includes an ECU 2, an imaging unit 3, and an audio output unit 4, and the ECU 2 includes a CPU 20.
In addition to the departure notification device 1, the vehicle 5 includes a speed sensor 51, a transmission 52, a foot brake 53, a brake lamp 54, and a side brake 55.

The speed sensor 51 is for detecting the vehicle speed of the vehicle 5, and the detection result of the speed sensor 51 is given to the CPU 20. The vehicle speed is “0” when the vehicle stops at an intersection or railroad crossing, and exceeds “0” when the vehicle departs.
When the vehicle is stopped at an intersection or railroad crossing, the transmission gear of the transmission 52 is shifted to neutral (hereinafter referred to as shift N) by the driver's operation, and the foot brake 53 and the side brake 55 are respectively Turned on. On the other hand, when the vehicle departs, the transmission gear of the transmission 52 is shift-changed to other than neutral (hereinafter referred to as other than shift N), and the foot brake 53 and the side brake 55 are turned off by the driver.

From the above, in this embodiment, if the vehicle speed indicated by the detection result of the speed sensor 51 is “0” and at least the transmission 52 is shifted N, the foot brake 53 is turned on, or the side brake 55 is turned on. Assume that the vehicle 5 has stopped. On the other hand, if the vehicle speed detected by the speed sensor 51 exceeds “0”, or if the transmission 52 is other than Shift N and the foot brake 53 and the side brake 55 are turned off, the vehicle 5 starts. Shall be.
The CPU 20 is supplied with shift position information of the transmission 52 and signals indicating on / off of the foot brake 53 and the side brake 55, respectively.

  The brake lamps 54 are a pair of red lamps and are arranged on the left and right sides of the rear part of the vehicle 5. The brake lamps 54 are turned on when the foot brake 53 is turned on and turned off when the foot brake 53 is turned off. Therefore, a change from turning off to turning on the brake lamp 54 means that the vehicle 5 has stopped, and a change from turning on to turning off means that the vehicle 5 has started.

The imaging unit 3 includes an imaging element 31 and a DSP 32. The image pickup device 31 uses a CCD, a CMOS, or the like, and is disposed in the center in the lateral direction of the front portion of the vehicle 5, and continuously or intermittently captures the scenery in front of the vehicle 5. The output of the image sensor 31 is given to the DSP 32 for image processing, and the DSP 32 is an RGB value image indicating still images captured at predetermined time intervals (for example, 0.1 seconds) based on the output of the image sensor 31. Data is generated in time series. The generated image data is associated with the timing at which the image was taken and is given to the ECU 2.
In the present embodiment, the imaging unit 3 is controlled and activated by the CPU 20 while the vehicle 5 is stopped, and is stopped when the vehicle leaves the vehicle. However, the imaging unit 3 may be constantly operating.

The audio output unit 4 uses a speaker, and outputs predetermined sounds such as chimes, melodies, and messages, and sounds such as in-vehicle radios and in-vehicle TVs (not shown) into the vehicle.
In addition to the CPU 20, the ECU 2 has a nonvolatile memory 21, a work memory 22, and an image processing unit 23 using an ASIC.
The image data of RGB values given from the imaging unit 3 to the ECU 2 is received by the image processing unit 23, and image processing and detection processing described later are performed. The processing result of the image processing unit 23 is given to the CPU 20 and stored in the work memory 23.
The CPU 20 is a control center of the idling stop support device 1, and uses the work memory 22 as a work area and executes various processes according to the control program and data stored in the nonvolatile memory 21. For example, the CPU 20 controls the imaging unit 3 to start / end imaging, and controls the audio output unit 4 to output audio corresponding to audio data stored in advance in the nonvolatile memory 21.

  Based on the image data generated by the imaging unit 3, the departure notification device 1 configured as described above determines whether or not the stopped vehicle 5 is to be started at the ECU 2, and at the timing to start the departure. If there is, a predetermined notification sound that prompts the driver to start the vehicle 5 is output from the sound output unit 4 (hereinafter referred to as a departure notification).

  The driver of the vehicle 5 recognizes that it is time to start the vehicle 5 by listening to the notification sound output from the sound output unit 4 and starts the vehicle 5. For example, a notification image for notifying the driver that it is time to depart, such as a predetermined message or an image of a traffic light that changes from a red signal to a blue signal, is displayed on a screen such as an in-vehicle TV or a car navigation system (not shown) When doing so, the driver needs to visually recognize the screen until the notification image is notified. However, in this embodiment, since it is notified by voice that the vehicle should depart, there is no problem even if the driver is looking out of the window, for example. In addition, radio sound, television sound, or the like may be output from the sound output unit 4 as long as it can be output simultaneously with the notification sound or by switching to the notification sound.

  The image processing unit 23 of the ECU 2 is controlled by the CPU 20 to convert the RGB value image data generated by the imaging unit 3 in time series while the vehicle 5 is stopped into the HSV value image data, and the converted HSV value. On the basis of the image data, a red region to be detected is detected, and the detection result is given to the CPU 20. The CPU 20 given the detection result of the image processing unit 23 serves as an index of the timing at which the vehicle 5 should depart based on the detection result of the red region relating to each of a plurality of image data captured in time sequence. It is determined whether or not a change has occurred in the detection target.

Finally, the CPU 20 controls the audio output unit 4 to perform departure notification when it is determined that a change that becomes an index of timing to depart has occurred in the detection target, and a change that becomes an index of timing to depart is detected. If it is determined that the image has not occurred, the image processing unit 23 repeatedly performs image data conversion, red area detection, and change occurrence determination until the vehicle 5 departs.
However, the CPU 20 determines whether the image data of the traffic light is included in the image data of RGB value or the image data of HSV value before detecting the red region to be detected. Different detection targets are used depending on whether the image data of the traffic light is included or not.
The image processing unit 23 as described above functions as an image conversion unit and a red detection unit.

2 and 3 are schematic diagrams showing examples of images captured by the imaging unit of the departure notification device according to the embodiment of the present invention, respectively. FIG. 2 is a traffic light S1 in front of the vehicle 5 stopped at the intersection. , S2 is illustrated, and FIG. 3 illustrates an image PF including a pair of left and right brake lamps B of a vehicle (that is, a preceding vehicle) F that is stopped in front of the vehicle 5 that is stopped on a straight road. Yes.
In FIG. 2, since the vehicle 5 is the leading vehicle, the preceding vehicle is not shown in the image PS. In FIG. 3, there is no traffic signal in the vicinity where the vehicle 5 is stopped, or the traffic signal is hidden behind the front vehicle F, and therefore the traffic signal is not shown in the image PF.

  In the ECU 2, first, using a predetermined template T including an image of a traffic light, whether or not the traffic light is reflected in the image PS (or the image PF) is confirmed by a template matching method. Specifically, the image processing unit 23 calculates the similarity between the image data of the template T and the image data of a part of the image PS corresponding to the template T (hereinafter referred to as a corresponding image). The similarity is calculated for the entire image PS by sequentially changing the corresponding image to another part of the image PS (for example, by shifting the corresponding position by one column / one row). The degree is given to the CPU 20. The CPU 20 to which the calculation result of the similarity is given determines that the corresponding image having the highest calculated similarity and equal to or higher than the predetermined similarity is a traffic light image.

In the case of the image PS shown in FIG. 2, the corresponding image including the traffic signal S1 and the corresponding image including the traffic signal S2 have a similarity greater than or equal to a predetermined similarity, but the correspondence image including the traffic signal S1 has the highest similarity. The image is employed as an image of a traffic light (hereinafter referred to as a signal image) AS.
When the signal image AS is present at a certain position in the image PS, the signal image AS is present at the same position in the image PS captured at least after the imaging timing of the image PS. For this reason, template matching need only be executed once.

  When there is an image of the traffic light, the ECU 2 uses the traffic light as a detection target, and uses a red signal that is lit as the red region of the detection target. In the case of the image PS, a red signal in which the traffic light S1 is lit is detected using the signal image AS. For example, if a red signal that is lit is reflected in the signal image AS of the first image PS and a red signal that is lit is also reflected in the signal image AS of the second image PS, the traffic light S1 is It can be seen that the signal remains red, that is, the timing to depart has not yet arrived. On the other hand, if a red signal that is extinguished is reflected in the signal image AS of the third image PS, it can be understood that the traffic light S1 has changed from a red signal to a blue signal, that is, the timing to depart.

  However, in the case of the image PF shown in FIG. 3, there is no image having a similarity higher than the predetermined similarity. In this case, since the traffic light is not shown in the image PF and there is a high possibility that the brake lamp B of the preceding vehicle F is shown instead, the image in the center area in the vertical and horizontal directions of the image PF (hereinafter referred to as a predetermined range image) AF is referred to as an image of the brake lamp B of the front vehicle F.

  When such a range image is used, the ECU 2 uses a front vehicle as a detection target and uses a lit brake lamp as a red region of the detection target. In the case of the image PF, the brake lamp B in which the front vehicle F is lit is detected. For example, if the brake lamp B that is lit is reflected in the first image PF and the brake lamp B that is lit is also reflected in the second image PF, the preceding vehicle F is stopped. That is, it can be seen that the timing to depart has not yet arrived. On the other hand, if the brake lamp B which is turned off is reflected in the third image PF, it can be understood that the preceding vehicle F has started, that is, the timing for starting the vehicle has arrived.

  By the way, in the present embodiment, after the RGB value image data is changed to the HSV value image data, the HSV value of the signal image AS or the range image AF is performed by performing template matching using the HSV value image data. However, the present invention is not limited to this procedure. For example, the image data of the RGB value of the signal image AS or the range image AF may be obtained by performing template matching using the image data of the RGB value, and this image data may be converted into the image data of the HSV value.

  The image data of the HSV value (at least the image data of the HSV value of the signal image AS or the range image AF) is converted into binary image data by the image processing unit 23 and then given to the work memory 22 and stored therein. Since the converted binary image data has a smaller amount of data than RGB or HSV image data, various operations on the binary image data are light in processing, and the work occupied by the binary image data is also low. The storage capacity of the memory 22 is also small.

In order to convert HSV image data into binary image data, all pixels of the image data have an H value in a predetermined range indicating red, and an S value and V value greater than each predetermined value indicating a bright color. Is replaced with a pixel value “1” representing white, and pixel values of pixels other than the bright red pixel are replaced with a pixel value “0” representing black. .
Specifically, the H value has a value of 0 ° ≦ H ≦ 360 °, and the H value in a predetermined range indicating red is 0 ° ≦ H <20 ° and 340 ° <H ≦ 360 °. The S value has a value of 0 ≦ S ≦ 1, and an S value equal to or higher than a predetermined value indicating a bright color is set to S ≧ 0.5. Further, the V value has a value of 0 ≦ V ≦ 255, and a V value equal to or higher than a predetermined value indicating a bright color is V ≧ 180.

Therefore, the pixel value of a pixel having a pixel value of 0 ° ≦ H <20 ° or 340 ° <H ≦ 360 °, S ≧ 0.5, and V ≧ 180 is “1”, and 20 ° ≦ H ≦ 340. The pixel value of a pixel having a pixel value of °, S <0.5, or V <180 is “0”.
The entire landscape image data may be binarized. In this case, it is only necessary to set each pixel value to “0” even if bright red pixels are included other than the traffic light or the image data of the vehicle ahead.

  4 to 6 are each a schematic diagram showing an example of an image showing a red region detected by the departure notification device according to the embodiment of the present invention. In each of FIGS. 4 to 6, (a) and (b), the bright red color of the lit brake lamp B is represented as white, and other than the bright red color, that is, yellow, green, blue, and off An image in which dark red or the like, which is the color of the brake lamp B, is represented as black (hatched in the figure) is illustrated.

Each of FIGS. 4 to 6A shows a detection result of a red region based on image data captured at the time of imaging timing t (t is a natural number; hereinafter, imaging timing is referred to as time). Each of FIGS. 6 (b) to 6 (b) shows a detection result of a red region based on image data captured at time t + 1 when a predetermined time (0.1 second in the present embodiment) has elapsed from time t. ing.
4A and 5A, the red region of the brake lamp B that is lit at time t is shown in white, and the number of pixels in the red region (that is, the pixel value is “1”). the pixel number) RP _t is the predetermined number TH1 (TH1 is a natural number) or more.

In FIG. 4B, the brake lamp B that is lit at time t + 1 is shown in white as a red region, and the number of pixels RP _{t + 1 in the} red region is equal to or greater than the predetermined number TH1. Specifically, in FIGS. 4A and 4B, RP _t = RP _{t + 1} ≧ TH1.
That is, when RP _t (or RP _{t + 1} ) ≧ TH1, the lit brake lamp B is present in the detection result.

On the other hand, in FIG. 5B, the brake lamp B that is turned off at time t + 1 is shown in black, and the number of pixels RP _{t + 1 in the} red region is less than the predetermined number TH1. Specifically, in FIG. 5B, RP _{t + 1} = 0 <TH1.
That is, when RP _t (or RP _{t + 1} ) <TH1, the brake lamp B that is lit does not exist in the detection result.

  Each of FIG. 4 to FIG. 6C shows the difference between the detection result at time t and the detection result at time t + 1. However, in this specification, the difference between the two detection results is the exclusive OR of the two detection results, that is, the pixel value “1” or “0” of each pixel of one detection result and the other detection result. Means an exclusive OR with the pixel value “1” or “0” of each corresponding pixel, and is obtained as follows.

つまり、検出結果の画像において白い画素（又は黒い画素）が黒い画素（又は白い画素）に変化している場合、検出結果の差分の画像ではこの画素は白い画素となり、検出結果の画像において白い画素（又は黒い画素）が白い画素（又は黒い画素）のまま変化していない場合、検出結果の差分の画像ではこの画素は黒い画素となる。

That is, when a white pixel (or black pixel) is changed to a black pixel (or white pixel) in the detection result image, this pixel is a white pixel in the difference image of the detection result, and a white pixel is detected in the detection result image. When (or a black pixel) remains a white pixel (or a black pixel), the pixel is a black pixel in the difference image of the detection result.

The number of pixels DRP _{t + 1 in} the white area in FIG. 4C, which is the difference between the detection result shown in FIG. 4A and the detection result shown in FIG. 4B, is DRP _{t + 1} / RP _{t + 1.} Is less than a predetermined value TH2 (TH2 is a positive real number).
On the other hand, the number of pixels DRP _{t + 1 in} the white region in FIG. 5C, which is the difference between the detection result shown in FIG. 5A and the detection result shown in FIG. 5B, is DRP _{t + 1} / RP _t. The calculation result of ₊₁ is greater than or equal to the predetermined value TH2.

That is, when DRP _{t + 1} / RP _{t + 1} <TH2, the brake lamp B that has been lit remains lit even when time elapses from time t to time t + 1, and the timing to depart is still coming. You can see that they are not. On the other hand, in the case of DRP _{t + 1} / RP _{t + 1} ≧ TH2, it can be seen that the time to depart has arrived because time has elapsed from time t to time t + 1 and the lit brake lamp B is extinguished.
Here, the arrival of the timing to depart from the white area pixel number DRP _{t + 1} itself is not determined, and the white area pixel number DRP _{t + 1} is determined by the red area pixel number RP _{t + 1} at time _{t + 1} . The reason for determining using the division is to remove the influence of noise included in each of RP _t , RP _{t + 1} , and DRP _{t + 1} .

By the way, when the traffic light or the vehicle ahead is not shown in the image picked up by the image pickup unit 3 or when the color of the lit brake lamp is shown as a color other than bright red due to the light beam, pixel RP _t is absent or a predetermined number TH1. In such a case, it is not possible to accurately determine the timing for departure.
Therefore, the CPU 20 determines whether or not RP _t ≧ TH1, and if RP _t <TH1, for example, turns on or blinks an unillustrated indicator or outputs a predetermined message from the audio output unit 4 This warns the driver that departure notification is impossible.

However, whether RP _t ≧ TH1 or RP _t <TH1, the red region is detected based on the image data captured at time t + 1, so that RP _{t + 1} ≧ TH1 (that is, For example, if the light intensity changes so that the color of the lit brake lamp appears as bright red), it is possible to accurately determine and notify the timing to depart.

In FIG. 6A, the red region of the brake lamp B that is lit at time t is shown in white, and the number of pixels in the red region (that is, the number of pixels having a pixel value “1”) RP. _t is a predetermined number TH1 or more.
In FIG. 6B, the brake lamp B that has been turned off at time t + 1 even though the brake lamp B has changed from lighting to off is shown in white as a red area, and the number of pixels in the red area (That is, the number of pixels having a pixel value “1”) RP _{t + 1} is _{equal to} or greater than a predetermined number TH1.
Such a phenomenon occurs when, for example, the amount of light incident on the brake lamp B is large, the amount of reflected light from the brake lamp B is large, and the color of the brake lamp B that is turned off appears as bright red.

However, since the front vehicle F in which the brake lamp B is turned off starts, the lateral distance of the brake lamp B at the time t + 1 is apparently shorter than the lateral distance of the brake lamp B at the time t. Therefore, even if RP _t = RP _{t + 1} , as shown in FIG. 6C, the difference between the detection result shown in FIG. 6A and the detection result shown in FIG. The number of pixels DRP _{t + 1} in the white region is DRP _{t + 1} > 0.
From the above, in the present embodiment, the number of pixels in the white region is continuously the predetermined number TH4 (TH4 is a natural number) TH3 (TH3 is a natural number. In the present embodiment, TH3 = 3) times or more. If it is the above (that is, DRP _{t + 1} , DRP _t , DRP _t−1 ≧ TH4), it is considered that the preceding vehicle F has departed, that is, the timing to depart has arrived.

However, when the number of pixels in the red region exceeds TH5 (TH5 is a natural number of TH5> TH1) continuously (TH3 is more than TH3 times) (that is, RP _{t + 1} , RP _t , RP _t-1 > TH5). Since the red region is unnaturally large and the influence of noise is considered to be too great, no departure notification is performed even if DRP _{t + 1} , DRP _t , DRP _t−1 ≧ TH4.

FIG. 7 is a flowchart showing a procedure of departure notification processing executed by the CPU of the departure notification device according to the embodiment of the present invention.
The CPU 20 determines whether or not the vehicle speed VS is “0” based on the detection result of the speed sensor 51 (S11). If VS> 0 (NO in S11), the vehicle 5 is traveling. , S11 is repeated.

When the vehicle speed VS = 0 (YES in S11), the CPU 20 determines whether any one or more of the three types of stop conditions are satisfied (S12). Here, the stop conditions are the following three types.
1. Transmission 52 is shift N.
2. Foot brake 53 is on.
3. Side brake 55 is on.
If all the stop conditions are not satisfied (NO in S12), that is, if the transmission 52 is other than the shift N, the foot brake 53 is off, and the side brake 55 is off, the vehicle 5 has not stopped. The CPU 20 returns the process to S11.

If any one or more of the stop conditions are satisfied (YES in S12), the vehicle 5 is stopped, so the CPU 20 controls the image pickup unit 3 to start shooting the scenery in front of the vehicle 5 (S13). ).
Next, the CPU 20 causes the image processing unit 23 to execute template matching based on the RGB value image data generated by the imaging unit 3 at time t = 1 (S14). In this case, RGB image data is given from the imaging unit 3 to the image processing unit 23, and the image processing unit 23 converts the given image data into HSV value image data.
Here, in order to convert the RGB value of each pixel into an HSV value, the following conversion formula is used in the image processing unit 23.

H = 60 × (GB) / (MAX-MIN), if MAX = R
H = 60 × (BR) / (MAX−MIN) +120, if MAX = G
H = 60 × (RG) / (MAX−MIN) +240, if MAX = B
S = (MAX-MIN) / MAX
V = MAX
However, if 0 ≦ R, G, B ≦ 256, the maximum value among the R value, G value, and B value is the maximum value MAX, the minimum value is the minimum value MIN, and the calculation result by the above formula is H <0, In order to make the range of the H value 0 ≦ H ≦ 360, the H value is further obtained using the following equation.
H = 360-H

When the maximum value MAX = 0, the minimum value MIN = 0 (that is, black of R, G, B = 0) and the S value and the H value are not defined, but since V = 0, HSV When the value image data is converted into binary image data, the pixel value becomes “0”.
After the RGB value image data is converted to the HSV value image data, the image processing unit 23 uses the image data of the template T given in advance and applies the converted HSV value image data as described above. Template matching is performed, the similarity between the image data of the corresponding image and the image data of the template T is calculated, and the calculation result is given to the CPU 20.
Based on a plurality of calculation results given from the image processing unit 23, the CPU 20 determines the presence or absence of image data of the corresponding image having the highest similarity calculated by the image processing unit 23 and equal to or higher than the predetermined similarity. Thus, it is determined whether the image data of the traffic light is included in the image data of the HSV value (S15).

When the image data of the traffic light is included in the image data of the HSV value (YES in S15), the CPU 20 sets so as to detect the red signal in which the traffic light is lit as the red region to be detected (S16). For this reason, the CPU 20 sets the range in which the signal image exists (that is, the range in which the image data of the corresponding image having the highest similarity calculated by the image processing unit 23 occupies the original image data) as the range to be noted. The image processing unit 23 is set so that only the image data in the power range, that is, the image data of the signal image is used and image data other than the image data is ignored. Further, the CPU 20 sets the values of the predetermined values TH1 to TH4 used in a subroutine for vehicle determination processing, which will be described later, to values for traffic lights.
Hereinafter, the image data of the signal image is referred to as attention image data.

On the other hand, when the image data of the traffic light is not included in the image data of the HSV value (NO in S15), the CPU 20 sets so as to detect the brake lamp that the front vehicle is lit as the red region to be detected ( S17). Therefore, the CPU 20 uses only the image data of the range image (that is, the image data of the predetermined range of the image data of the HSV value converted by the image processing unit 23) and ignores the image data other than this image data. The image processing unit 23 is set. Further, the CPU 20 sets the values of the predetermined values TH1 to TH4 used in the vehicle determination process subroutine to values for the forward vehicle.
Hereinafter, the image data of the range image is also referred to as attention image data.

After the processing of S16 or S17 is completed, the CPU 20 based on the attention image data,
The detection of the red region to be detected is executed by the image processing unit 23 (S18).
In this case, the image processing unit 23 replaces the pixel value of a pixel having an H value in a predetermined range indicating red color and having an S value and a V value equal to or higher than each predetermined value indicating a bright color with “1”. By substituting the pixel values of the other pixels with “0”, the binarization processing of the target image data is performed. This is a red region in which a pixel having a pixel value “1” of the target image data binarized (hereinafter referred to as binarized data) is detected.

In addition, the image processing unit 23 counts the number of pixels RP of the detected red region, further calculates a difference between the binarized data for two times before and after, and counts the number of pixels DRP of the white region in the difference image data. To do. However, since there is no previous binarized data, the value of the number of pixels DRP counted here is a meaningless value. Thereafter, the counted pixel number RP and pixel number DRP are given to the CPU 20, and binarized data corresponding to the time t = 1 is given to the work memory 22 and stored therein.
After the completion of the process S18, CPU 20 substitutes "1" in the variable t (S19), the number of pixels RP supplied from the image processing unit 23 substitutes the pixel number RP _t (S20), the pixel number DRP pixels Substituting into the number DRP _t (S20), a subroutine for performing departure determination processing (see FIGS. 8 to 9) is called and executed (S22).
After the process of S22 is completed, the CPU 20 controls the imaging unit 3 to end the imaging of the scenery in front of the vehicle 5 (S23), and returns the process to S11.

8 and 9 are flowcharts showing a subroutine of departure determination processing procedures executed by the CPU of the departure notification device according to the embodiment of the present invention.
The CPU 20 causes the image processing unit 23 to detect the red region to be detected based on the RGB value image data generated by the imaging unit 3 at time t + 1 (S31).
In this case, image data of RGB values is given from the imaging unit 3 to the image processing unit 23, and the image processing unit 23 receives attention image data (that is, image data of a signal image or image data of a range image) included in the given image data. ) Are converted into image data of HSV values, and binarization processing is further performed to obtain binarized data, and the number of pixels RP in the red region is counted based on the obtained binarized data.

Further, the image processing unit 23 calculates the exclusive logical sum of the binarized data corresponding to the time t stored in the work memory 22 and the binarized data corresponding to the time t + 1 to obtain the difference image data. The number of pixels having the pixel value “1” in the obtained difference image data, that is, the number of pixels DRP _{t + 1} in the white area is counted. Thereafter, the counted pixel number RP and pixel number DRP are given to the CPU 20, and the binarized data corresponding to the time t + 1 is given to the work memory 22 and stored therein. Note that when storing the binarized data, the binarized data corresponding to the time t may be overwritten.
After completing the process of S31, the CPU 20 substitutes the pixel number RP given from the image processing unit 23 for the pixel number RP _{t + 1} (S32), and substitutes the pixel number DRP for the pixel number DRP _{t + 1} (S33).

Then CPU20 is the number of pixels RP _t of the detection result according to the time t is equal to or a predetermined number TH1 or more (S34).
If RP _t <TH1 (NO in S34), there are too few red regions detected from the target image data at time t (that is, the traffic light and the preceding vehicle are not shown in the image captured by the imaging unit 3). Or the color of the brake lamp that is lit is reflected as a color other than bright red due to the amount of light), and the timing at which to depart cannot be determined accurately, so the CPU 20 turns on the indicator, for example, The driver is warned that departure notification is impossible (S35).
After the process of S35 is completed, the CPU 20 moves the process to S57, which will be described later, and determines whether the vehicle 5 is still stopped or started, and executes the processes after S31 if the vehicle 5 is stopped.

If RP _t ≧ TH1 (YES in S34), the CPU 20 determines whether RP _{t + 1} is “0” (S51).
If RP _{t + 1} = 0 (YES in S51), the CPU 20 controls the audio output unit 4 to issue a departure notification (S52). This is because RP _{t + 1} = 0 means that the red region that existed at time t has completely disappeared at time t + 1, which means that the red signal or brake lamp that is lit This is because it is considered to have been turned off.
After completing the process of S52, the CPU 20 ends the departure determination process and returns the process to the routine that called the departure determination process subroutine.

When RP _{t + 1} > 0 (NO in S51), the CPU 20 determines whether or not the calculation result of DRP _{t + 1} / RP _{t + 1} is equal to or greater than a predetermined value TH2 (S53).
If DRP _{t + 1} / RP _{t + 1} ≧ TH2 (YES in S53), the CPU 20 moves the process to S52 and issues a departure notification. This is because DRP _{t + 1} / RP _{t + 1} ≧ TH2 means that there is a large difference between the detection result at time t and the detection result at time t + 1. This is because the red light or the brake lamp is turned off. Here, when the process of S53 is executed, since RP _{t + 1} > 0, the calculation result of DRP _{t + 1} / RP _{t + 1} does not diverge.

When DRP _{t + 1} / RP _{t + 1} <TH2 (NO in S53), the CPU 20 first executes the processes after S54 without executing the process of S52. This is because DRP _{t + 1} / RP _{t + 1} <TH2 means that there is no significant difference between the detection result at time t and the detection result at time t + 1. This is because there is a high possibility that the lighting is continued.
The CPU 20 determines whether or not t ≧ TH3 (S54). If t <TH3 (NO in S54), the CPU 20 still calculates the value of the number of pixels DRP _t−1 necessary for at least the processing of S55 described later. Therefore, the process proceeds to S57 without executing the processes of S55 and S56 described later.

If t ≧ TH3 (YES in S54), it is determined whether or not the number of pixels DRP _{t + 1} , DRP _t , DRP _{t-1 in} the past three white areas is greater than or _{equal to a} predetermined number TH4 (S55). , DRP _{t + 1} , DRP _t , DRP _t-1 <TH4 (NO in S55), there is no great difference in the difference in the detection results over the past three times (ie, the red signal or the brake lamp continues to be lit) Therefore, the process proceeds to S57 without executing the process of S56.
When DRP _{t + 1} , DRP _t , DRP _t−1 ≧ TH4 (YES in S55), the past three red region pixel counts RP _{t + 1} , RP _t , RP _t-1 are each a predetermined number TH5 or less. It is determined whether or not (S56).

When RP _{t + 1} , RP _t , RP _t−1 ≦ TH5 (YES in S56), the CPU 20 moves the process to S52 and issues a departure notification. This is because DRP _{t + 1} , DRP _t , DRP _t−1 ≧ TH4 and RP _{t + 1} , RP _t , RP _t−1 ≦ TH5 is not the influence of noise but the difference between the past three detection results. This means that a large difference has occurred, and it is considered that the vehicle ahead has departed and the brake lamp has moved away.
If RP _{t + 1} , RP _t , RP _t-1 > TH5 (NO in S56), the CPU 20 causes the process to proceed to S57 because there is a large difference in the difference between the past three detection results due to the influence of noise. Transfer.

After completion of the process of S35, or when NO in S54, S55, or S56, the CPU 20 determines whether or not the vehicle speed VS exceeds “0” based on the detection result of the speed sensor 51 (S57). When VS = 0 (NO in S57), it is determined whether or not all three of the three types of departure conditions are satisfied (S58). Here, the departure conditions are the following three types.
1. Transmission 52 is other than shift N.
2. Foot brake 53 is off.
3. The side brake 55 is off.
If at least one of the departure conditions is not satisfied (NO in S58), since the vehicle 5 is still stopped, the variable t is incremented by “1” (S59), and the CPU 20 returns the process to S31.

When vehicle speed VS> 0 (YES in S57) or when all departure conditions are satisfied (YES in S58), since vehicle 5 has started, CPU 20 ends the departure determination process, and a subroutine for this departure determination process Return processing to the routine that called.
The CPU 20 in S51, S53, and S55 and S56 functions as a change determination unit, and the audio output unit 4 controlled by the CPU 20 in S52 functions as a departure notification unit. Further, the CPU 20 in S15 functions as a signal determination unit, the CPU 20 in S34 functions as a presence determination unit, and the audio output unit 4 controlled by the CPU 20 in S35 functions as an impossible warning unit.

The departure notification device 1 as described above indicates that when a vehicle 5 is stopped at an intersection, a railroad crossing or the like, when a change from a red signal to a green signal, a start operation of a preceding vehicle, or the like is detected, the vehicle should be started. Since the driver can be notified by voice, the driver of the vehicle 5 can start the vehicle at an appropriate timing. As a result, the flow of traffic becomes smooth, and the driver does not panic because of pressure from the following vehicle.
Furthermore, since it is suppressed that the blinker that lights up in yellow is mistaken as a brake lamp, it is suppressed that the driver is notified of the departure by turning off the blinker. It is suppressed that the driver gets confused or causes the accident by causing the vehicle 5 to start according to the false alarm.

In addition to the output of the notification voice, for example, a vibration generating unit may be built in the driver's seat and a predetermined vibration may be given to the driver as long as the stopped vehicle 5 should depart.
The imaging unit 3 generates RGB value image data, and the image processing unit 23 is not limited to the configuration in which the image processing unit 23 converts the generated image data into HSV value image data. For example, the imaging unit 3 has YUV value image data. The ECU 2 may convert the generated image data into HSV value image data.

It is a block diagram which shows the principal part structure of the departure notification apparatus which concerns on embodiment of this invention. It is a schematic diagram which shows an example of the image containing the traffic signal imaged by the imaging part of the departure notification apparatus which concerns on embodiment of this invention. It is a schematic diagram which shows an example of the image containing the brake lamp of the front vehicle which the imaging part of the departure notification apparatus which concerns on embodiment of this invention imaged. It is a schematic diagram which shows an example of the image which shows the red area | region (specifically the brake lamp in lighting) detected by the departure notification apparatus which concerns on embodiment of this invention. It is a schematic diagram which shows an example of the image which shows the red area | region (specifically the brake lamp which changed from lighting to light extinction) detected by the departure notification apparatus which concerns on embodiment of this invention. It is a schematic diagram which shows an example of the image which shows the red area | region (specifically the brake lamp which moves ahead) detected by the departure notification apparatus which concerns on embodiment of this invention. It is a flowchart which shows the procedure of the departure notification process which CPU of the departure notification apparatus which concerns on embodiment of this invention performs. It is a flowchart which shows the subroutine of the departure determination processing procedure which CPU of the departure notification apparatus which concerns on embodiment of this invention performs. It is a flowchart which shows the subroutine of the departure determination processing procedure which CPU of the departure notification apparatus which concerns on embodiment of this invention performs.

Explanation of symbols

1 departure notification device 20 CPU
23 Image processing unit (image conversion means, red detection means)
3 imaging unit 4 audio output unit (departure notification means, impossible warning means)
5 Vehicle B Brake lamp F Front vehicle (vehicle stopped in front of the vehicle)
S1 traffic light

Claims (4)

An imaging unit that generates time-series image data obtained by continuously or intermittently imaging the front of the vehicle while stopped,
In the departure notification device for notifying that the vehicle to be stopped should depart at the timing to depart,
Image conversion means for converting the image data generated by the imaging unit into image data of HSV values;
A red detection means for detecting a red region of a predetermined detection target based on the image data converted by the image conversion means;
A change for determining whether or not a change serving as an index of the timing to depart has occurred in the detection target based on a detection result of the red detection unit relating to each of a plurality of pieces of image data captured in time sequence. A determination means;
A departure notification device, comprising: a departure notification means for notifying by voice or vibration that a departure should be made when it is determined that the change determination means has occurred. A signal determination unit for determining whether the image data of the traffic light is included in the image data converted by the image conversion unit or the image data generated by the imaging unit;
When it is determined that the signal determination unit is included, the red detection unit detects a red signal in which a traffic light is lit as a red region to be detected based on image data of the traffic signal. Yes,
When the signal determination means determines NO, the red detection means uses a brake lamp that is lit on a vehicle parked in front of the vehicle as a red area to be detected based on a predetermined range of image data. The departure notification device according to claim 1, wherein the departure notification device is detected.   The red detection means distinguishes a pixel having an S value and a V value having a predetermined range of H values indicating red and having a bright color or more, and a pixel other than the pixel by binary values. The departure notification device according to claim 1, wherein a red region of the detection target is detected. Presence determination means for determining whether a predetermined number or more of pixels in the red region to be detected are present in the detection result of the red detection means for one image data;
An impossibility warning means that warns that it is impossible to notify that the vehicle should be departed when the presence determination means determines NO
The red color detecting unit detects a red region to be detected based on image data after the image data regardless of a determination result of the presence determining unit. 4. The departure notification device according to any one of items 3.

Images