Sighting control method based on human detection, product, device and medium

1 . A sighting control method based on human detection, comprising:

acquiring a current scene image in a current field of view of a sighting device;

determining whether there is a human target currently being aimed at based on a target recognition result of the current scene image and a position of an aiming mark in the current scene image;

detecting whether a shooting action occurs if it is determined that there is a human target currently being aimed at; and

controlling the sighting device to self-destroy if a shooting action occurs.

2 . The sighting control method based on human detection according to claim 1 , wherein, before detecting whether a shooting action occurs if it is determined that there is a human target currently being aimed at, the sighting control method comprises:

triggering the sighting device to enter an abnormal sighting mode; or

triggering the sighting device to enter an abnormal sighting mode if the human target keeps being aimed at or the human target remains within a preset range of the aiming mark for over a preset time or a preset number of consecutive frames from the moment of the human target being currently detected to be aimed at.

3 . The sighting control method based on human detection according to claim 1 , wherein an inertial sensor is arranged in the sighting device, and detecting whether a shooting action occurs comprises:

determining whether a shooting action occurs by analyzing sensor data acquired in real time by the inertial sensor.

4 . The sighting control method based on human detection according to claim 3 , wherein the inertial sensor comprises an accelerometer and/or a gyroscope, and determining whether a shooting action occurs by analyzing sensor data acquired in real time by the inertial sensor comprises:

analyzing sensor data acquired in real time by the accelerometer and/or the gyroscope to determine whether the sensor data matches a shooting behavior template to determine whether a shooting action occurs, wherein the shooting behavior template is generated by analyzing first sensor data and second sensor data, respectively corresponding to historical shooting actions, acquired by the accelerometer and/or the gyroscope.

5 . The sighting control method based on human detection according to claim 1 , wherein an inertial sensor is arranged in the sighting device, and detecting whether a shooting action occurs comprises:

determining whether a shooting action occurs, by analyzing sensor data acquired in real time by the inertial sensor and by determining whether the human target satisfies a preset downward deviation condition with respect to the aiming mark according to multiple consecutive frames of the current scene image.

6 . The sighting control method based on human detection according to claim 5 , wherein, determining whether a shooting action occurs, by analyzing sensor data acquired in real time by the inertial sensor and by determining whether the human target satisfies a preset downward deviation condition with respect to the aiming mark according to multiple consecutive frames of the current scene image, comprises:

analyzing sensor data acquired in real time by the inertial sensor to determine whether the sensor data matches a shooting behavior template; and

while it is determined that the sensor data of the inertial sensor matches the shooting behavior template, determining whether the human target satisfies a quick downward deviation and disappearing condition with respect to the aiming mark according to a change of the position of the human target with respect to the aiming mark in the multiple consecutive frames of the current scene image so as to determine whether a shooting action occurs.

7 . The sighting control method based on human detection according to claim 1 , wherein the sighting device comprises a fuse element connected between a power module and a main control module, and controlling the sighting device to self-destroy comprises:

outputting a self-destroy current-rise instruction to control the power module to output a self-destroy current causing the fuse element to melt to form an open circuit.

8 . The sighting control method based on human detection according to claim 1 , wherein determining whether there is a human target currently being aimed at comprises:

performing a global target detection and recognition with respect to the current scene image based on a target recognition algorithm, to recognize and mark all human targets contained in the current scene image;

based on a position of the aiming mark in the current scene image, determining whether the position of the aiming mark overlaps a position of any one of the human targets that have been marked; and

determining whether there is a human target being currently aimed at based on whether a position of the human target overlaps with a position of the aiming mark.

9 . The sighting control method based on human detection according to claim 1 , wherein determining whether there is a human target currently being aimed at comprises:

determining a predetermined area image with a position of the aiming mark in the current scene image as a reference position;

performing a local target detection and recognition with respect to the predetermined area image based on a target recognition algorithm to determine whether there is a human target within the predetermined area image within a set time; and

determining whether there is a human target being currently aimed at based on whether there is a human target present within the predetermined area image within a set time.

10 . The sighting control method based on human detection according to claim 1 , wherein the target recognition is a deep learning-based target recognition method, comprising:

determining whether there is a target present in the current scene image using a target detection model;

if a target exists in the current scene image, performing feature extraction on a current target using a feature extraction model to obtain a target feature vector of the current target;

comparing the target feature vector with human feature vectors in a human feature vector database to determine whether the current target is a human target; wherein the human feature vector database has been obtained by inputting sample images containing human targets into the feature extraction model.

11 . The sighting control method based on human detection according to claim 10 , wherein comparing the target feature vector with human feature vectors in a human feature vector database to determine whether the current target is a human target comprises:

calculating a distance between the target feature vector and each of the human feature vectors in the human feature vector database, and determining whether the current target is a human target based on the distance.

12 . A sighting device comprising:

a memory configured to store a computer program therein;

a processor configured to execute the computer program to implement the sighting control method based on human detection according to claim 1 ;

an image sensor configured to acquire a current scene image in a current field of view; and

a display module configured to display the current scene image and an aiming mark.

13 . The sighting device according to claim 12 , further comprising an inertial sensor configured to acquire, in real time, sensor data to be analysed to determine whether a shooting action occurs.

14 . The sighting device according to claim 13 , wherein the inertial sensor comprises at least one of an accelerometer and a gyroscope.

15 . The sighting device according to claim 12 , further comprising a fuse element connected between a power module and a main control module, the processor further configured to output a self-destroy current-rise instruction to control the power module to output a self-destroy current causing the fuse element to melt to form an open circuit.

16 . A computer-readable storage medium, wherein a computer program is stored in the computer-readable storage medium, and the computer program is executable by a processor to implement the sighting control method based on human detection according to claim 1 .