TWI773958B - Method, apparatus and system for display images of blind spot of vehicle - Google Patents

Abstract

A method, an apparatus and a system for displaying images of a blind spot of vehicle are provided. The apparatus is coupled to at least one external camera disposed outside the vehicle, at least one internal camera disposed inside the vehicle with a lens facing a driver, and at least one display disposed inside the vehicle. In the method, external images outside the vehicle are captured by using the external camera, and internal images including the driver are captured by using the internal camera. A face of the driver in the internal images and a displacement of the face are recognized, and a location of a region of interest (ROI) in the external images are adjusted according to the recognized displacement, so as to display images of the ROI on the corresponding display.

Description

Vehicle blind spot image display method, device and system

The present invention relates to an image processing method and device, and more particularly, to a vehicle blind spot image display method, device and system.

The driver of the vehicle cannot see the complete scene outside the vehicle because the line of sight is blocked by the rear mirror and the column of the vehicle. This obscured field of vision is the blind spot or blind spot of driving. When the vehicle is moving, accidents often occur because people or vehicles in the blind spot or blind spot cannot be seen.

For example, from the left front view of the bus driver shown in FIG. 1A , part of the outside scene of the vehicle is blocked by the left A-pillar 12 and the rear mirror 14 . It can be seen from the blind spot area of bus driving shown in FIG. 1B , the outside scene of the bus 10 seen by the driver will be shielded by the left A-pillar 12 and the rear mirror 14 respectively. The blind spot area covers part of the Banma line. , driving may be involved in an accident because you cannot see pedestrians in the blind spot area.

In order to solve the above problem of blind spots in the field of view, vehicle manufacturers have proposed many solutions, such as making use of side sensing technology to make blind spot warning devices, using camera Take the side image of the car and project it on the screen, or change the design of the column to project the outside image on the column through optical refraction, or even make a hole in the column to allow the driver to see through the outside of the car.

However, changing the column design may weaken the strength of the column and affect the safety, while the direct projection of the side image of the vehicle will not be able to correctly and naturally present the image corresponding to the blind spot of the current field of view due to changes in the driving posture or viewing angle.

In view of this, the present invention provides a vehicle blind spot image display method, device and system, which can adjust the displayed outside image in response to the change of the driver's sitting posture or the viewing angle, so as to correctly present the image of the blind spot of the current driving field.

A vehicle blind spot image display method according to an embodiment of the present invention is applicable to an electronic device having a connection device and a processor. Wherein, the connecting device is used for coupling at least one external camera disposed outside the vehicle, at least one internal camera disposed inside the vehicle and having a lens facing the driver, and at least one display disposed inside the vehicle. The method includes using an external camera to capture an external image of the outside of the vehicle, using an internal camera to capture an internal image including the driver, identifying the driver's face and its displacement in the internal image, and correspondingly adjusting the external image according to the identified displacement. The position of the region of interest (ROI) is displayed to display the image of the region of interest on the corresponding display.

A vehicle blind spot image display device according to an embodiment of the present invention includes a connection device, a storage device, and a processor. Wherein, the connecting device is used for coupling at least one external camera disposed outside the vehicle, at least one external camera disposed inside the vehicle and the lens facing the driver An interior camera and at least one display disposed inside the vehicle. The storage device is used to store programs. The processor is coupled to the connection device and the storage device, and is used for loading and executing the program in the storage device to capture the external image of the outside of the vehicle by using the external camera, capture the internal image including the driving by using the internal camera, and identify the driving in the internal image. face and its displacement, and correspondingly adjust the position of the region of interest in the external image according to the identified displacement, so as to display the image of the region of interest on the corresponding display.

A vehicle blind spot image display system according to an embodiment of the present invention includes at least one exterior camera, at least one interior camera, at least one display, and an electronic device. Among them, the exterior camera is disposed outside the vehicle, the interior camera is disposed inside the vehicle with the lens facing the driver, and the display is disposed inside the vehicle. The electronic device includes a connecting device and a processor, wherein the connecting device is used for coupling the external camera, the internal camera and the display. The processor is configured to capture an external image of the outside of the vehicle using the external camera, capture an internal image including the driver using the internal camera, identify the driver's face and displacement in the internal image, and adjust the external image accordingly based on the identified displacement to display the image of the region of interest on the corresponding display.

In order to make the above-mentioned features and advantages of the present invention more obvious and easy to understand, the following embodiments are given and described in detail with the accompanying drawings as follows.

10: Bus

12. P1, P2: A pillar

14: Rear view mirror

100: Vehicle Blind Spot Image Display Device

110: Connection device

120: Storage Device

130: Processor

300: Vehicle Blind Spot Image Display System

510, 510’: External image

520, 520’: Internal image

512, 512’: area of interest images

522, 522': face box

C1, C2: External cameras

C3: Internal Camera

D: driving

FC, FC': Center reference point

M1, M2: Display

ROI, ROI': Area of Interest

S402~S406: Steps

FIG. 1A shows the left front view of driving a bus.

FIG. 1B illustrates a blind spot area for bus driving.

FIG. 2 is a block diagram of a vehicle blind spot image display device according to an embodiment of the present invention.

FIG. 3 is a schematic diagram of a vehicle blind spot image display system according to an embodiment of the present invention.

FIG. 4 is a flowchart of a method for displaying a blind spot image of a vehicle according to an embodiment of the present invention.

FIG. 5A and FIG. 5B are examples of a vehicle blind spot image display method according to an embodiment of the present invention.

In the present invention, in addition to arranging a camera on the outside of the vehicle (for example, outside the column) to capture external images, a camera is additionally installed inside the vehicle to capture driving images, and the driver's face and its displacement are identified by a deep learning algorithm. And it is fed back to the display of the external image, so the image of the area where the driver's field of view is blocked can be appropriately displayed to assist the driver in judging the external situation, thereby reducing the probability of an accident.

FIG. 2 is a block diagram of a vehicle blind spot image display device according to an embodiment of the present invention. The vehicle blind spot image display device 100 in this embodiment at least includes a connection device 110 , a storage device 120 and a processor 130 . The vehicle blind spot image display device 100 is, for example, an electronic device with computing functions, such as a personal computer, a server, a tablet computer, a navigation device, a video monitor box (Cambox), and a car computer, which is not limited to this embodiment.

The connecting device 110 is used for coupling to at least one external camera disposed outside of at least one pillar of the vehicle (eg, left A-pillar, right A-pillar), at least one internal camera disposed inside the vehicle and having a lens facing the driver, and at least one internal camera disposed on the pillar. at least one display inside the body to receive images from the external camera and the internal camera and to the display Output the processed external image. Specifically, the connection device 110 can be connected to the external camera and the internal camera in any wired or wireless manner, and receives images captured by the external camera and the internal camera, such as a Universal Serial Bus (USB) , RS232, Bluetooth (Bluetooth, BT), wireless fidelity (Wireless fidelity, Wi-Fi) and other transmission interfaces, which are not limited in this embodiment. On the other hand, the connection device 110 may also be an image output interface such as a Video Graphics Array (VGA), a High Definition Multimedia Interface (HDMI), a Thunderbolt, etc., for outputting the video signal to the display. The external image to be processed is not limited in this embodiment.

For example, FIG. 3 is a schematic diagram of a vehicle blind spot image display system according to an embodiment of the present invention (the vehicle blind spot image display device 100 is not shown). Referring to FIG. 3 , the system 300 of the present embodiment illustrates the configuration of cameras and displays connected to the vehicle blind spot image display device 100 of FIG. 2 . Wherein, the system 300 of this embodiment disposes cameras C1 and C2 on the outer sides of the right A-pillar P1 and the left A-pillar P2 of the vehicle, respectively, to capture the external images of the right and left sides of the vehicle, respectively. The cameras C1 and C2 can be, for example, wide-angle cameras (eg, 120-150 degrees), fisheye cameras (eg, 190 degrees) or pinhole cameras with wider field of view (FOV), which are not limited in this embodiment. In one embodiment, the cameras C1 and C2 are, for example, disposed on the A-pillars P1 and P2 of the vehicle about 120-150 cm (especially 140 cm) above the ground to avoid being hit by people or foreign objects. In other embodiments, the above-mentioned configuration height can also be adjusted according to factors such as vehicle type, driving body shape, external environment, etc., which is not limited herein. In other embodiments, the cameras C1 and C2 may also be arranged at other positions outside the vehicle.

In addition, the system 300 of this embodiment also configures displays M1 and M2 on the inner sides of the right A-pillar P1 and the left A-pillar P2 of the vehicle, respectively, to display the right and left external images captured by the cameras C1 and C2, respectively. In one embodiment, the displays M1 and M2 are, for example, a Liquid Crystal Display (LCD), a Light-Emitting Diode (LED) display, or other types of flat-panel displays or flexible displays. set limits. In one embodiment, the displays M1 and M2 are, for example, projection screens of a projector, and can display images projected by the projector. In other embodiments, the cameras C1 and C2 may also be arranged at other positions in the vehicle. In other embodiments, there may be only one display and only one of the right A-pillar P1 and the left A-pillar P2 may be arranged, or there may be more than two displays.

It should be noted that, the vehicle blind spot image display device 100 of the present embodiment uses the camera C3 disposed inside the A-pillar P1 on the right side of the vehicle to capture the interior image including the driver D, and uses the deep learning algorithm to detect the face of the driver D. and its displacement, thereby adjusting the external images displayed on the displays M1 and M2. In one embodiment, the camera C3 may also be disposed at other positions in the vehicle at a suitable distance (eg, about 1-2 meters away from the driving D). In one embodiment, the camera C3 is, for example, a camera with a relatively small field of view (eg, 30-60 degrees), which is not limited herein.

The storage device 120 is, for example, any type of fixed or removable random access memory (Random Access Memory, RAM), read-only memory (ROM), flash memory (flash memory), A hard disk or similar element or a combination of the above elements is used to store programs executable by the processor 130 .

The processor 130 is coupled to the connection device 110 and the storage device 120 , and can load and execute programs stored in the storage device 120 . In different embodiments, the processor 130 is, for example, a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphics Processing Unit, GPU), or other programmable general-purpose or special-purpose microprocessors ( Microprocessor), Digital Signal Processor (DSP), Programmable Controller, Application Specific Integrated Circuits (ASIC), Programmable Logic Device (PLD) or other similar device or a combination of these devices, but not limited thereto.

FIG. 4 is a flowchart of a method for displaying a blind spot image of a vehicle according to an embodiment of the present invention. Please refer to FIG. 2 and FIG. 4 at the same time. The method of this embodiment is applicable to the above-mentioned vehicle blind spot image display device 100 . The following describes the detailed steps of the vehicle blind spot image display method of this embodiment in combination with various elements of the vehicle blind spot image display device 100 . .

In step S402, the processor 130 uses external cameras (eg, C1 and C2) to capture external images of the outside of the vehicle column. The external camera is, for example, a wide-angle camera or a fisheye camera with a wide field of view (FOV), which can be used to capture external images that can cover all possible blind spots of the field of view when driving on the driver's seat.

In step S404, the processor 130 uses an internal camera (eg, C3) to capture an internal image including the driver, and recognizes the driver's face and its displacement in the internal image. In one embodiment, the blind spot image display device 100 may be used for the first time (for example, when the vehicle is just acquired by driving) or when the device is initially used (for example, when the vehicle is just started when driving). In the initial setting, the processor 130, for example, uses an internal camera to capture an initial image of the driver sitting in the driver's seat in advance, uses a deep learning method to detect the facial features of the driver in the initial image, and records the face position. Those skilled in the art may know a variety of existing deep learning algorithms for face recognition, which will not be described here. During the initial setting, the driver can manually adjust (for example, but not limited to, by operating the user interface or the remote control of the vehicle blind spot image display device 100 ) the external images captured by the external cameras are correspondingly displayed on the monitors. The viewing angle or scale factor of each region of interest (ROI) on the vehicle is adjusted to meet its requirements, and the vehicle blind spot image display device 100 can record the initial setting of each ROI corresponding to the face position. In the subsequent actual driving process, the processor 130 may continue to use the deep learning method to identify the faces in the multiple frames of the internal image for multiple frames of the internal image.

In some embodiments, such as a bus, multiple human faces (eg, a driver and a bus passenger) may appear in the internal image at the same time, and the processor 130 may recognize multiple human faces in the internal image. At this time, the processor 130 may determine that the face in the internal image that is closest to the driver's face identified in the previous frame is the driver's face in this frame. In this way, even if multiple faces (eg, passengers) appear in the interior image at the same time, the processor 130 can still accurately identify the driving, so as to correctly execute the vehicle blind spot image display method according to the embodiment of the present invention.

It should be noted that, in order to solve the problem of poor image quality and recognition caused by insufficient driving lights at night, in one embodiment, the internal camera may use a camera with an active light source (such as an infrared fill light) to increase the Night camera quality. In other embodiments, the processor 130 may not be able to communicate with each frame of the internal image The face is recognized by the deep learning algorithm of face recognition (for example, when the light source is insufficient), at this time, the processor 130 of the vehicle blind spot image display device 100 can use the deep learning method with the feature points of the face region to continuously identify each image in the internal image If a face is identified in several of the frames, the processor 130 can lock the area around the position in the image of the adjacent subsequent frames according to the position of the face in the frame, and continue Identify the object area with the same feature points as the face area in the subsequent image frame as the face of the subsequent image frame; through the combination of the above face recognition and motion estimation technology, the low light source can be used. Continue to track faces.

In step S406, the processor 130 correspondingly adjusts the position of the region of interest (ROI) in the external image according to the identified displacement, so as to display the image of the region of interest on the corresponding display. In the embodiment of the present invention, for example, an area in which the driver's current field of view is blocked is defined as the area of interest in the external image captured by the external camera, and the change between the displacement of the driver's head and the movement of the area of interest is determined. Calculating the movement amount of the area of interest when the current driver's face is displaced, so that the final captured area of interest image can conform to or cover the blind spot area of the driver's current field of view. In some embodiments, during driving, the process of FIG. 4 may be repeatedly performed every one or more frames to display the image of the region of interest on the display in real time.

In detail, the processor 130 uses the recognized displacement to calculate the displacement of the human face (for example, in units of pixels of the internal image), so that the displacement of the human face and the movement of the region of interest (for example, in the external image) are calculated according to the displacement of the human face. The change relationship between the pixels is the unit), calculate the movement of the area of interest corresponding to the movement of the face, and finally move the area of interest in the opposite direction of the displacement of the face by the calculated movement. determine the final location of the area of interest.

,

)。在其他實施例中，辨識出來的涵括人臉輪廓的外框可以是其他形狀，而臉部中心參考點的座標可以是根據其他方式決定(例如外框形狀的質心)。 For example, in the initial setting, using the deep learning algorithm to recognize the face can obtain the coordinates of the face frame that can include the outline of the face (for example, in units of pixels of the internal image), including the face frame. The upper left corner vertex coordinate and the lower right corner vertex coordinate, through these two coordinates, the center point coordinate of the face frame can be calculated, which is used as the coordinate of the face center reference point FC ( FC _x , FC _y ). When driving, the deep learning algorithm is used to continuously identify the face of each image frame, and the coordinates of the displaced face center reference point FC' can be obtained (

,

). In other embodiments, the identified outer frame including the outline of the human face may be in other shapes, and the coordinates of the face center reference point may be determined according to other methods (eg, the center of mass of the outer frame shape).

-FC _x )來推算各圖幀的關注區域(ROI)相對於初始設定的關注區域的移動量，包括顯示於右側A柱顯示器的右側ROI移動量S _R 及顯示於左側A柱顯示器的左側ROI移動量S _L ，其計算公式如下：

Based on the change of the field of view occlusion area during driving is mainly affected by the head/face change in the horizontal direction (ie, the x-axis direction), in one embodiment, only the head/face in the horizontal direction of each picture frame is taken. part movement (

- FC _x ) to estimate the movement amount of the region of interest (ROI) of each frame relative to the initial set region of interest, including the right ROI movement amount _SR displayed on the right A-pillar monitor and the left ROI displayed on the left A-pillar monitor The movement amount _SL is calculated as follows:

-FC_x)|大於該臉部位移閥值時才對應調整外部影像中關注區域的位置，否則就不調整關注區域的位置，以節省運算資源。 Among them, n is the scale coefficient, a is the ROI displacement coefficient of the left A-pillar, and b is the ROI displacement coefficient of the right A-pillar. In one embodiment, n is 0.0005, a is 1.5, and b is 1. The values of the above coefficients can be adjusted according to different situations, which are not limited herein. In one embodiment, when the display on the A-pillar is closer to the driver, its corresponding ROI displacement coefficient is larger (for example, the ROI displacement coefficient of the left A-pillar closer to the driver is larger than the right A-pillar farther away from the driver). ROI displacement coefficient), i.e. the amount of ROI movement in response to the displacement of the driving head/face is larger. Although the above-mentioned calculation of the movement amount of the region of interest during driving is performed according to the displacement of the face center reference point of each picture frame relative to the initially set face center reference point, in another embodiment, it can also be calculated according to each picture frame. The displacement of the face center reference point of the frame relative to the face center reference point of the previous image frame is used to calculate the movement amount of the region of interest corresponding to each image frame relative to the region of interest corresponding to the previous image frame. In yet another embodiment, the scale coefficient n , the ROI displacement coefficient a of the left A-pillar, and the ROI displacement coefficient b of the right A-pillar may be proportional to the scaling ratio set on each display in the above-mentioned initial setting. In yet another embodiment, a face displacement threshold can be set, when |(

When -FC _x )| is greater than the face displacement threshold, the position of the region of interest in the external image is adjusted accordingly, otherwise the position of the region of interest will not be adjusted to save computing resources.

It should be noted that the above embodiment is based on the fact that the driver seldom moves the head/face in the vertical direction during driving, so the vertical displacement is not taken into consideration to calculate the ROI movement amount. However, in other embodiments, the change amount of the head/face in the vertical direction can also be used to calculate the movement amount of the ROI and adjust the vertical position of the ROI area, which is not limited herein.

FIG. 5A and FIG. 5B are examples of a vehicle blind spot image display method according to an embodiment of the present invention. Please refer to FIG. 3 and FIG. 5A at the same time, FIG. 5A shows the external image 510 and the internal image 520 captured by the external camera C2 and the internal camera C3 at the time point t, wherein the internal image can be detected by the deep learning algorithm 520 medium driving Drive the face of D, and obtain the coordinates of the center reference point FC of the face frame 522, and based on the coordinates of the center reference point FC, the position of the region of interest ROI in the external image 510 can be calculated according to the pre-determined variation relationship (that is, the blind spot area of the driver D's field of view), and captures the image 512 of the region of interest ROI accordingly, and displays it on the display M2 disposed on the pillar P2. Those skilled in the art should understand that if the resolution of the region of interest ROI is different from the resolution of the display screen, the vehicle blind spot image display device 100 or the display can adjust the resolution accordingly.

Next, please refer to FIG. 3 and FIG. 5B at the same time, FIG. 5B shows the external image 510 ′ and the internal image 520 ′ captured by the external camera C2 and the internal camera C3 at the time point t+Δt, in which the deep learning algorithm is used The face of the driver D in the internal image 520' can be detected, and the coordinates of the center reference point FC' of the face frame 522' can be obtained. Based on this central reference point FC' and the coordinates of the central reference point FC at the previously obtained time point t, the movement amount of the driver D's head in the time Δt can be calculated, and according to the displacement amount of the face and the movement amount of the area of interest The change relationship between the ROI can be calculated to move the ROI in the opposite direction of the face displacement by the calculated movement amount, and the position of the new ROI' can be obtained and extracted accordingly. The image 512' of the region of interest ROI' is displayed on the display M2 disposed on the pillar P2. The external image 510' and the internal image 520' captured at the time point t+Δt are, for example, the next image captured after the external camera C2 and the internal camera C3 capture the external image 510 and the internal image 520, or The images captured after an interval of a fixed time interval can be set by the user, which is not limited here.

In one embodiment, the vehicle blind spot image display device 100 may be based on the It has a face recognition function, which additionally performs face recognition on the external images captured by the external camera, so as to identify the pedestrians appearing in the external images. When the pedestrian is detected, the vehicle blind spot image display device 100 can use the in-vehicle audio or display to issue a warning message to remind the driver to pay attention to the pedestrian and reduce accidents. In one embodiment, the vehicle blind spot image display device 100 may issue a warning message only when the vehicle blind spot image display device 100 determines that the distance between the pedestrian and the vehicle in the external image is lower than a predetermined value (eg, 2.5 meters), wherein the distance between the pedestrian and the vehicle is It can be estimated based on, for example, the number of pixels occupied by the pedestrian in the external image or the position of the pedestrian's foot in the image of the external image. In other embodiments, the vehicle blind spot image display device 100 may also detect pedestrians appearing in external images based on other pedestrian detection algorithms, which is not limited herein. In some embodiments, when the vehicle is stopped, no pedestrian detection is performed or/and the above-mentioned position adjustment of the region of interest is not performed.

In summary, the vehicle blind spot image display method, device and system of the present invention can adjust the position of the region of interest in the external image correspondingly according to the movement amount and direction of the driver's head, so as to respond to changes in the driver's sitting posture or viewing angle, Correctly display the image of the blind spot of the driver's current field of vision, assist the driver in judging the external situation, and then reduce the probability of an accident. In addition, the vehicle blind spot image display method, device and system of the present invention can be used only by installing a camera and a display on the vehicle body, which is convenient for use on traditional vehicle styles and does not affect the robustness of the vehicle body structure. Although embodiments of the present invention are primarily described with respect to buses or trucks, they may also be used in other vehicle types such as sedans or recreational vehicles.

Although the present invention has been disclosed above with the embodiments, it is not intended to limit the present invention. Invention, anyone with ordinary knowledge in the technical field can make some changes and modifications without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention should be regarded as defined by the appended patent application scope as follows: allow.

100: Vehicle Blind Spot Image Display Device

110: Connection device

120: Storage Device

130: Processor

Claims (18)

A vehicle blind spot image display method, which is suitable for an electronic device having a connecting device and a processor, the connecting device is coupled to at least one external camera disposed outside the vehicle, and at least one interior camera disposed inside the vehicle with a lens facing the driver A camera and at least one display disposed inside the vehicle, the method includes the following steps: capturing an external image outside the vehicle by using the external camera; defining an external image in the external image captured by the external camera a region of interest (ROI); using the interior camera to capture an interior image including the driver, and identifying the driver's face and the displacement of the face in the interior image; and according to The recognized displacement correspondingly adjusts the position of the region of interest in the external image, so as to display the image corresponding to the adjusted region of interest on the display, wherein the corresponding adjustment is made according to the recognized displacement. The step of the position of the region of interest in the external image includes: calculating the displacement of the region of interest corresponding to the displacement of the human face according to the change relationship between the displacement of the human face and the movement amount of the region of interest. a movement amount; and moving the region of interest in the external image by the movement amount in a direction opposite to the displacement of the human face. The method of claim 1, wherein before the step of recognizing the driver's face and the displacement of the human face in the internal image, the method further comprises: using the internal camera to capture images in advance Take an initial image of the driver sitting in the driver's seat; and detect the face of the driver in the initial image, and record the position of the face and the region of interest corresponding to the position of the face. Initial setting. The method of claim 1, wherein the step of recognizing the driver's face and the displacement of the human face in the internal image includes: when multiple faces are identified in the internal image, determining The face closest to the driver's face identified in the previous frame in the internal image is the driver's face in this frame, so as to calculate the displacement of the face. The method of claim 1, wherein the step of identifying the driver's face in the internal image and the displacement of the human face comprises: detecting the driver's face in the internal image; and according to the position of the face in the identified picture frame, identify the object located in the surrounding area of the position of the face and having the same feature points as the area of the face in the images of the adjacent subsequent picture frames The region is used as the face of the subsequent frame to calculate the displacement of the face. The method of claim 1, wherein when the displacement of the human face is greater than a face displacement threshold, the region of interest is moved by the moving amount in the opposite direction of the displacement of the human face, otherwise The region of interest is not moved. The method of claim 1, further comprising: identifying a pedestrian appearing in the external image captured by the external camera; and sending a warning message when the pedestrian is detected. A vehicle blind spot image display device, comprising: a connecting device coupled to at least one exterior camera disposed outside the vehicle, at least one interior camera disposed inside the vehicle with a lens facing driving, and at least one interior camera disposed inside the vehicle a display; a storage device that stores a program; and a processor, coupled to the connection device and the storage device, to load and execute the program in the storage device to capture an external image of the outside of the vehicle by using the external camera ; Define a region in the external image captured by the external camera as a region of interest (ROI); use the internal camera to capture the internal image including the driver, and identify the internal image The driving face and the displacement of the human face described in; and correspondingly adjusting the position of the region of interest in the external image according to the identified displacement, and displaying the adjusted image corresponding to the region of interest In the display, wherein the step of correspondingly adjusting the position of the region of interest in the external image according to the identified displacement comprises: according to the change between the displacement of the human face and the movement amount of the region of interest relationship, calculate the movement amount of the region of interest corresponding to the displacement of the face; and moving the region of interest in the external image by the moving amount in a direction opposite to the displacement of the human face. The vehicle blind spot image display device as claimed in claim 7, wherein the processor comprises in advance using the internal camera to capture an initial image of the driver sitting in the driver's seat, and to detect any information in the initial image. the driving face, and recording the position of the face and the initial setting of the region of interest corresponding to the position of the face. The vehicle blind spot image display device according to claim 7, wherein when a plurality of human faces are recognized in the internal image, the processor determines whether the internal image is the same as the one recognized in the previous frame. The closest face of the driver's face is the driver's face in this image frame, so as to calculate the displacement of the face. The vehicle blind spot image display device according to claim 7, wherein the processor includes detecting the driver's face in the interior image, and according to the recognized position of the face in the frame , identify the object area that is located in the surrounding area of the position of the face and has the same feature points as the area of the face in the image of the adjacent subsequent frames as the face of the subsequent frame, to Calculate the displacement of the face. The vehicle blind spot image display device according to claim 7, wherein the processor shifts the region of interest toward the face when the displacement of the face is greater than a face displacement threshold Move the movement amount in the opposite direction, otherwise the region of interest is not moved. The vehicle blind spot image display device according to the claim 7, wherein the processor further recognizes a pedestrian appearing in the external image captured by the external camera, and when detecting the pedestrian, A warning message is issued. A vehicle blind spot image display system, comprising: at least one external camera disposed on the outside of the vehicle; at least one internal camera disposed inside the vehicle with a lens facing the driver; at least one display disposed on the inside of the vehicle; and an electronic device , having a connection device coupled to the external camera, the internal camera and the display, and a processor configured to: capture an external image of the outside of the vehicle using the external camera ; Define a region in the external image captured by the external camera as a region of interest (ROI); use the internal camera to capture the internal image including the driver, and identify the internal image The driving face and the displacement of the human face described in; and correspondingly adjusting the position of the region of interest in the external image according to the identified displacement, and displaying the adjusted image corresponding to the region of interest In the display, wherein the step of correspondingly adjusting the position of the region of interest in the external image according to the identified displacement comprises: according to the change between the displacement of the human face and the movement amount of the region of interest relationship, calculate the movement amount of the region of interest corresponding to the displacement of the face; and calculate the displacement of the region of interest towards the face in the external image Move the movement amount in the opposite direction. The vehicle blind spot image display system as claimed in claim 13, wherein the processor comprises in advance using the internal camera to capture an initial image of the driver sitting in the driver's seat, and to detect any information in the initial image. the driving face, and recording the position of the face and the initial setting of the region of interest corresponding to the position of the face. The vehicle blind spot image display system as claimed in claim 13, wherein when multiple faces are recognized in the interior image, the processor determines whether the interior image is the same as that identified in the previous frame. The closest face of the driver's face is the driver's face in this image frame, so as to calculate the displacement of the face. The vehicle blind spot image display system of claim 13, wherein the processor includes detecting the driver's face in the interior image, and identifying the position of the face in the frame according to the image , identify the object area that is located in the surrounding area of the position of the face and has the same feature points as the area of the face in the image of the adjacent subsequent frames as the face of the subsequent frame, to Calculate the displacement of the face. The vehicle blind spot image display system according to claim 13, wherein the processor shifts the region of interest toward the human face only when the displacement of the human face is greater than a face displacement threshold. Move the movement amount in the opposite direction, otherwise the region of interest is not moved. The vehicle blind spot image display system as described in claim 13, wherein the processor further recognizes a pedestrian appearing in the external image captured by the external camera, and when detecting the pedestrian, A warning message is issued.

Images