Vehicle unlocking method, system, and apparatus, vehicle, and storage medium

1 . A vehicle unlocking method, wherein the vehicle unlocking method comprises:

obtaining a voice signal;

in response to determining that the voice signal comprises a wakeup keyword, extracting a voiceprint feature in the voice signal, wherein the wakeup keyword indicates that a vehicle needs to be unlocked;

in response to determining that the extracted voiceprint feature matches a stored voiceprint feature, determining an acoustic source direction based on the voice signal;

selecting, based on the acoustic source direction and from a plurality of cameras installed on the vehicle, a camera installed on the vehicle as a target camera to collect image data, wherein an angle of view of the target camera is configured to capture the acoustic source direction;

indicating, based on the acoustic source direction, the target camera to collect the image data; and

unlocking the vehicle based on the image data collected by the target camera, wherein:

unlocking the vehicle based on the image data collected by the target camera comprises matching the image data collected by the target camera comprises matching the image data collected by the target camera with stored figure image data,

wherein matching the image data collected by the target camera with stored figure image data comprises extracting a facial feature of a figure that is recognized in the image data and that is located in the acoustic source direction,

wherein extracting the facial feature of a figure that is recognized in the image data and that is located in the acoustic source direction comprises determining an image region from the image data based on the acoustic source direction, and

wherein determining the image region from the image data based on the acoustic source direction comprises:

determining a relative position of the acoustic source direction within the angle of view of the target camera;

determining a reference position on the image data based on the relative position; and

extending the reference position on the image data to each of a left side and a right side by a reference neighborhood range to obtain the image region.

2 . The vehicle unlocking method according to claim 1 , wherein the unlocking the vehicle based on the image data collected by the target camera comprises:

unlocking the vehicle in response to determining that the image data collected by the target camera matches the stored figure image data.

3 . The vehicle unlocking method according to claim 2 , wherein:

the stored figure image data comprises a facial feature;

the unlocking the vehicle in response to determining that the image data collected by the target camera matches the stored figure image data comprises:

unlocking the vehicle in response to determining that the extracted facial feature matches any stored facial feature.

4 . The vehicle unlocking method according to claim 3 , wherein the extracting a facial feature of a figure that is recognized in the image data and that is located in the acoustic source direction further comprises: extracting a facial feature of a figure recognized in the image region.

5 . The vehicle unlocking method according to claim 2 , wherein:

the image data collected by the target camera is a video, and the stored figure image data comprises a gait feature;

the matching the image data collected by the target camera with stored figure image data comprises:

extracting a gait feature of a figure recognized in the video; and

the unlocking the vehicle in response to determining that the image data collected by the target camera matches the stored figure image data comprises:

unlocking the vehicle in response to determining that the gait feature of the figure recognized in the image data matches any stored gait feature.

6 . The vehicle unlocking method according to claim 2 , wherein:

the image data collected by the target camera is a video, and the stored figure image data comprises a gait feature;

the matching the image data collected by the target camera with stored figure image data comprises:

extracting a facial feature of a figure recognized in the image data;

determining, as a target figure, a recognized figure whose facial feature matches any stored facial feature;

extracting a gait feature of the target figure recognized in the image data; and

unlocking the vehicle if the gait feature of the target figure matches any stored gait feature.

7 . A vehicle unlocking apparatus, wherein the vehicle unlocking apparatus comprises one or more memories and at least one processor, wherein:

the one or more memories are configured to store a wakeup keyword, wherein the wakeup keyword is used to indicate that a vehicle currently needs to be unlocked; and

the one or more memories store program instructions for execution by the at least one processor to:

obtain a voice signal;

in in response to determining that the voice signal comprises a wakeup keyword, extract a voiceprint feature in the voice signal, wherein the wakeup keyword indicates that a vehicle needs to be unlocked;

in response to determining that the extracted voiceprint feature matches a stored voiceprint feature, determine an acoustic source direction based on the voice signal;

select, based on the acoustic source direction and from a plurality of cameras installed on the vehicle, a camera installed on the vehicle as a target camera to collect image data, wherein an angle of view of the target camera is configured to capture the acoustic source direction; and

indicate, based on the acoustic source direction, the target camera to collect the image data; and

unlock the vehicle based on the image data collected by the target camera, wherein:

unlocking the vehicle based on the image data collected by the target camera comprises matching the image data collected by the target camera comprises matching the image data collected by the target camera with stored figure image data,

wherein matching the image data collected by the target camera with stored figure image data comprises extracting a facial feature of a figure that is recognized in the image data and that is located in the acoustic source direction,

wherein extracting the facial feature of a figure that is recognized in the image data and that is located in the acoustic source direction comprises determining an image region from the image data based on the acoustic source direction, and

wherein determining the image region from the image data based on the acoustic source direction comprises:

determining a relative position of the acoustic source direction within the angle of view of the target camera;

determining a reference position on the image data based on the relative position; and

extending the reference position on the image data to each of a left side and a right side by a reference neighborhood range to obtain the image region.

8 . The vehicle unlocking apparatus according to claim 7 , wherein unlocking the vehicle based on the image data collected by the target camera comprises:

unlocking the vehicle in response to determining that the image data collected by the target camera matches the stored figure image data.

9 . The vehicle unlocking apparatus according to claim 8 , wherein:

the stored figure image data comprises a facial feature;

unlocking the vehicle in response to determining that the image data collected by the target camera matches the stored figure image data comprises:

unlocking the vehicle in response to determining that the extracted facial feature matches any stored facial feature.

10 . The vehicle unlocking apparatus according to claim 9 , wherein extracting the facial feature of the figure that is recognized in the image data and that is located in the acoustic source direction further comprises:

extracting a facial feature of a figure recognized in the image region.

11 . A non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores program instructions for execution by at least one processor to:

obtain a voice signal;

in response to determining that the voice signal comprises a wakeup keyword, extract a voiceprint feature in the voice signal, wherein the wakeup keyword indicates that a vehicle needs to be unlocked;

in response to determining that the extracted voiceprint feature matches a stored voiceprint feature, determine an acoustic source direction based on the voice signal;

select, based on the acoustic source direction and from a plurality of cameras installed on the vehicle, a camera installed on the vehicle as a target camera to collect image data, wherein an angle of view of the target camera is configured to capture the acoustic source direction;

indicate, based on the acoustic source direction, the target camera to collect the image data; and

unlock the vehicle based on the image data collected by the target camera wherein:

unlocking the vehicle based on the image data collected by the target camera comprises matching the image data collected by the target camera comprises matching the image data collected by the target camera with stored figure image data,

wherein matching the image data collected by the target camera with stored figure image data comprises extracting a facial feature of a figure that is recognized in the image data and that is located in the acoustic source direction,

wherein extracting the facial feature of a figure that is recognized in the image data and that is located in the acoustic source direction comprises determining an image region from the image data based on the acoustic source direction, and

wherein determining the image region from the image data based on the acoustic source direction comprises:

determining a relative position of the acoustic source direction within the angle of view of the target camera;

determining a reference position on the image data based on the relative position; and

extending the reference position on the image data to each of a left side and a right side by a reference neighborhood range to obtain the image region.

12 . The non-transitory computer-readable storage medium according to claim 11 , wherein unlocking the vehicle based on the image data collected by the target camera comprises:

unlocking the vehicle in response to determining that the image data collected by the target camera matches the stored figure image data.

13 . The non-transitory computer-readable storage medium according to claim 12 , wherein:

the stored figure image data comprises a facial feature;

unlocking the vehicle in response to determining that the image data collected by the target camera matches the stored figure image data comprises:

unlocking the vehicle in response to determining that the facial feature of the figure recognized in the image data matches any stored facial feature.

14 . The vehicle unlocking method according to claim 1 , wherein determining that the voice signal comprises the wakeup keyword comprises:

determining that a spectrogram of the voice signal matches a spectrogram of the wakeup keyword.

15 . The vehicle unlocking method according to claim 1 , wherein determining that the voice signal comprises the wakeup keyword comprises:

determining that text information of the voice signal includes the wakeup keyword.

16 . The vehicle unlocking apparatus according to claim 7 , wherein the one or more memories store program instructions for execution by the at least one processor to:

determine that a spectrogram of the voice signal matches a spectrogram of the wakeup keyword.

17 . The vehicle unlocking apparatus according to claim 7 , wherein the one or more memories store program instructions for execution by the at least one processor to:

determine that text information of the voice signal includes the wakeup keyword.

18 . The vehicle unlocking method according to claim 1 , wherein the reference neighborhood range comprises a predetermined fraction of an entire width of an image of the image data.

19 . The vehicle unlocking apparatus according to claim 7 , wherein the reference neighborhood range comprises a predetermined fraction of an entire width of an image of the image data.

20 . The non-transitory computer-readable storage medium according to claim 11 , wherein the reference neighborhood range comprises a predetermined fraction of an entire width of an image of the image data.