Environmental recognition method and system for assisting blind people in traveling

Abstract

The present invention provides an environmental recognition method and system for assisting blind people in their travel. This system involves acquiring environmental monitoring images and object information, including the object's target location and current location. The system then determines the object's local trajectory based on the environmental monitoring images and object information. The object then moves according to the local trajectory. Obstacles along the local trajectory are captured from the environmental monitoring images. Obstacle information is acquired in real time to obtain obstacle information. The system then determines the risk of the obstacle relative to the object based on the object and obstacle information. When the risk value is equal to or greater than a preset risk threshold, a risk avoidance strategy is generated for the object to execute. Preemptive measures are taken to help blind people proactively avoid obstacles, rather than passively responding to them after the danger occurs, thereby reducing the probability of accidents occurring during travel.

Description

Technical Field

[0001] The present invention relates to the field of environmental recognition technology, and in particular to an environmental recognition method and system for assisting blind people in traveling. Background Art

[0002] Although the coverage rate of urban blind paths is high, the actual utilization rate is insufficient. The main reason is that shared bicycles, motor vehicles, trash cans and other objects have occupied the blind paths for a long time.

[0003] Blind people not only have to deal with static obstacles while driving, but also with dynamic environments, such as pedestrians, vehicles, and temporary construction. Traditional guidance devices (such as canes) can only detect local static obstacles and cannot warn of fast-moving objects. For example, if a blind person is in the driver's blind spot when a vehicle starts, a collision could occur. Summary of the Invention

[0004] The purpose of the present invention is to provide an environment recognition method and system for assisting blind people in traveling. The technical problem to be solved is how to assist blind people in traveling and reduce the risks of blind people traveling.

[0005] The present invention is achieved through the following technical solutions:

[0006] A first aspect provides an environment recognition method for assisting blind people in traveling, comprising the following steps:

[0007] Acquire environmental monitoring images and object information, wherein the object information includes a target location and a current location of the object;

[0008] Determining a local movement trajectory of the object through the above-mentioned environmental monitoring image and object information; the above-mentioned object moves according to the local movement trajectory;

[0009] Obtain obstacles on the local movement trajectory from the above-mentioned environmental monitoring image;

[0010] Obtain information of the above obstacles in real time to obtain obstacle information;

[0011] Determine the risk value of the obstacle relative to the object based on the object information and obstacle information;

[0012] When the risk value is equal to or greater than a preset risk threshold, a risk avoidance strategy is generated for execution by the subject.

[0013] By acquiring environmental monitoring images and object information including the target and current locations of objects, the system determines the specific location of the blind person in the environment, providing basic data for subsequent trajectory planning and obstacle avoidance. A local trajectory is planned based on the actual environment, guiding the blind person to avoid known dangerous areas (such as walls, ditches, and motorway lanes), providing preliminary safety for the blind person's travel. Obstacles along the local trajectory are captured from environmental monitoring images, and obstacle information is obtained in real time, allowing the blind person to promptly identify obstacles ahead and avoid collisions and other hazards during movement. For example, if a moving vehicle or pedestrian is in front of the blind person, the system provides timely updates of the obstacle's status, allowing the blind person to understand the latest obstacle movements and take appropriate countermeasures. By comprehensively considering multiple factors such as the blind person's location, movement speed, and distance to the obstacle, the system quantifies risks that are previously difficult to intuitively assess, generating a risk value. Risk avoidance strategies can be developed based on different risk scenarios, such as guiding the blind person to change direction, slow down, or stop moving. This improves the blind person's ability to respond to sudden dangers and effectively reduces travel risks.

[0014] In summary, the above methods are to take measures in advance before danger occurs to help blind people actively avoid obstacles, rather than passively respond after danger occurs, thereby reducing the probability of accidents occurring during the travel of blind people.

[0015] Furthermore, the local movement trajectory of the object is determined by using the above-mentioned environmental monitoring image and object information. The specific steps include:

[0016] Generate a global action trajectory based on the current position and target position of the above object information;

[0017] Determine the monitoring area through the above environmental monitoring images;

[0018] The action trajectory passing through the monitoring area is intercepted from the above global action trajectory to obtain the local action trajectory.

[0019] The global motion trajectory is generated by the current position and target position of the object information, providing the blind with a complete path planning from the starting point to the end point; the monitoring area is determined by the environmental monitoring image, and the local area is focused on the specific environment where the blind person is currently located. The monitoring area changes according to the blind person's movement and changes in the environment, and obstacles and potential risks around the blind person (uncovered sewage wells, bicycles, electric vehicles, etc.) are perceived in the monitoring area; the interception of the local motion trajectory enables the blind person to focus more on the movement path within the current monitoring area, avoiding interference from other irrelevant parts of the global trajectory, and improving the targetedness and safety of the action; the local motion trajectory can be dynamically adjusted according to the real-time changes in the monitoring area (i.e., risk avoidance strategy), ensuring that the blind person moves along a safe and reasonable path during movement and effectively avoids sudden obstacles or risks.

[0020] Furthermore, the obstacles on the above-mentioned local motion trajectory include animals and still objects; wherein the above-mentioned animals and objects are in relative motion, and the still objects and objects are relatively stationary.

[0021] Since animals move relative to objects, their movement direction and speed are uncertain, and the risks they pose to blind people are relatively complex; while static objects are relatively still relative to objects, their movement direction and speed are consistent with those of the objects, and they will not pose any danger to the objects. Therefore, static objects are not considered when calculating the risk value.

[0022] Furthermore, the risk value of the obstacle relative to the object is determined based on the object information and obstacle information. The specific steps include:

[0023] The obstacle information includes the obstacle's moving direction, moving speed, and current position; the object information also includes the object's moving direction, moving speed, and current position;

[0024] Determine the distance between the obstacle and the object based on the current positions of the obstacle and the object;

[0025] The probability of collision between the obstacle and the object is predicted based on their current positions and movement directions; the above probability is the risk value.

[0026] Risk assessment is performed based on real-time object and obstacle information, so the risk value can be dynamically adjusted as the environment and object status change, reflecting the current risk situation in real time, providing the blind with timely and accurate safety tips and risk avoidance guidance, and ensuring that the blind are always in a relatively safe state during travel.

[0027] Furthermore, when the risk value is greater than a preset risk threshold, a risk hedging strategy is generated for execution by the subject, and the specific steps include:

[0028] Determine the collision position of the obstacle and the object based on the moving direction, moving speed and separation distance of the obstacle and the object;

[0029] Determine the collision area based on the collision position and the preset collision increment;

[0030] Determine the avoidance time based on the collision position, the current position of the obstacle, and the moving speed of the obstacle;

[0031] The object's avoidance position is determined by moving away from the collision area based on the avoidance time, the object's current position, and the object's moving speed.

[0032] Based on the relative movement trend of obstacles and objects in space, potential dangerous situations are predicted. When the risk value is greater than the risk threshold, the collision position, collision area and safe area are determined, and a risk avoidance strategy is formulated based on the collision position. The risk avoidance strategy can be: when the object is in the safe area, the object waits for the obstacle to pass before moving; when the object is in the collision area, the object's movement direction is changed so that the object enters the safe area or leaves the center of the collision area to reduce the severity of the injury.

[0033] Furthermore, the collision area is determined by the collision position and the preset collision increment. The specific steps include:

[0034] Obtain the front, rear, left, and right endpoints of the obstacle, and superimpose the collision increments based on their positions to obtain the width and length of the collision area. Using the collision position as the midpoint, amplify the calculated width and length to obtain the collision area. Areas other than the collision area are considered safe zones.

[0035] Consider the range of obstacles that could pose a threat to the blind in all directions. An obstacle isn't just a point or a simple geometric shape; potential risks exist in all directions. Ensure that the blind are effectively protected within a defined safety zone around the obstacle, avoiding missing risk areas due to incomplete safety zone definition. The safety zone provides a clear safety boundary for the blind. The blind can adjust their movement path and speed based on the safety zone to avoid entering dangerous areas where collisions could occur. For example, when a blind person approaches an obstacle, understanding the scope of the safety zone allows them to plan a detour in advance, ensuring they pass within the safe zone and effectively reducing the risk of collision.

[0036] A second aspect provides an environment recognition system for assisting blind people in traveling, the system adopting the above-mentioned environment recognition method;

[0037] The system includes: an image acquisition device, a positioning device, a processing device and a voice device, wherein the image acquisition device, the positioning device and the voice device are connected to the processing device;

[0038] The above-mentioned image acquisition device is used to acquire environmental monitoring images;

[0039] The above positioning device is used to obtain the current position of the object in real time;

[0040] The above-mentioned voice device is used to receive the target position input by the subject and output voice information;

[0041] The processing device is used to perform the following operations:

[0042] Determine the local movement trajectory of the object through the above-mentioned environmental monitoring image and object information; and transmit the above-mentioned local movement trajectory to the object through a voice device;

[0043] Obtain obstacles on the local movement trajectory from the above-mentioned environmental monitoring image;

[0044] Obtain information of the above obstacles in real time to obtain obstacle information;

[0045] Determine the risk value of the obstacle relative to the object based on the object information and obstacle information;

[0046] When the risk value is equal to or greater than a preset risk threshold, a risk avoidance strategy to be executed by the subject is generated; and the risk avoidance strategy is transmitted to the subject via a voice device.

[0047] Furthermore, the processing device is used to determine the local movement trajectory of the object through the environmental monitoring image and object information, and the specific steps include:

[0048] Generate a global action trajectory based on the current position and target position of the above object information;

[0049] Determine the monitoring area through the above environmental monitoring images;

[0050] The action trajectory passing through the monitoring area is intercepted from the above global action trajectory to obtain the local action trajectory.

[0051] Furthermore, the processing device is used to determine the risk value of the obstacle relative to the object based on the object information and the obstacle information, and the specific steps include:

[0052] The obstacle information includes the obstacle's moving direction, moving speed, and current position; the object information also includes the object's moving direction, moving speed, and current position;

[0053] Determine the distance between the obstacle and the object based on the current positions of the obstacle and the object;

[0054] The probability of collision between the obstacle and the object is predicted based on their current positions and movement directions; the above probability is the risk value.

[0055] Furthermore, when the risk value is greater than a preset risk threshold, a risk hedging strategy is generated for execution by the subject, and the specific steps include:

[0056] Determine the collision position of the obstacle and the object based on the moving direction, moving speed and separation distance of the obstacle and the object;

[0057] Determine the collision area based on the collision position and the preset collision increment;

[0058] Determine the avoidance time based on the collision position, the current position of the obstacle, and the moving speed of the obstacle;

[0059] The object's avoidance position is determined by moving away from the collision area based on the avoidance time, the object's current position, and the object's moving speed.

[0060] Furthermore, the image acquisition device obtains the front endpoint, rear endpoint, left endpoint and right endpoint of the obstacle, and superimposes the collision increments based on the positions of the front endpoint, rear endpoint, left endpoint and right endpoint to obtain the width and length of the collision area; with the collision position as the midpoint, the calculated width and length are amplified to obtain the collision area; the other areas except the collision area are safe areas.

[0061] A third aspect provides a computer-readable storage medium having computer program instructions stored thereon, which implement the above-mentioned environment recognition method when the computer program instructions are executed by a processor.

[0062] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0063] By acquiring environmental monitoring images and object information including the target and current locations of objects, the system determines the specific location of the blind person in the environment, providing basic data for subsequent trajectory planning and obstacle avoidance. A local trajectory is planned based on the actual environment, guiding the blind person to avoid known dangerous areas (such as walls, ditches, and motorway lanes), providing preliminary safety for the blind person's travel. Obstacles along the local trajectory are captured from environmental monitoring images, and obstacle information is obtained in real time, allowing the blind person to promptly identify obstacles ahead and avoid collisions and other hazards during movement. For example, if a moving vehicle or pedestrian is in front of the blind person, the system provides timely updates of the obstacle's status, allowing the blind person to understand the latest obstacle movements and take appropriate countermeasures. By comprehensively considering multiple factors such as the blind person's location, movement speed, and distance to the obstacle, the system quantifies risks that are previously difficult to intuitively assess, generating a risk value. Risk avoidance strategies can be developed based on different risk scenarios, such as guiding the blind person to change direction, slow down, or stop moving. This improves the blind person's ability to respond to sudden dangers and effectively reduces travel risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the examples. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be considered as limiting the scope. A person of ordinary skill in the art can also derive other relevant drawings based on these drawings without inventive effort. In the drawings:

[0065] Figure 1 As the main flow chart;

[0066] Figure 2 This is a schematic diagram of the relative movement between a blind person and an obstacle;

[0067] Figure 3 Schematic diagram of the collision area, safety area, and object before and after avoiding danger;

[0068] Figure 4 for Figure 3 A partial enlarged view of middle A. DETAILED DESCRIPTION

[0069] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with examples and drawings. The exemplary embodiments of the present invention and their descriptions are only used to explain the present invention and are not intended to limit the present invention.

[0070] First embodiment:

[0071] Combine Figure 1 , an environment recognition method for assisting blind people in traveling, comprising the following steps:

[0072] Acquire environmental monitoring images and object information, wherein the object information includes a target location and a current location of the object;

[0073] Determining a local movement trajectory of the object through the above-mentioned environmental monitoring image and object information; the above-mentioned object moves according to the local movement trajectory;

[0074] Obstacles on the local motion trajectory are obtained from the above-mentioned environmental monitoring images; the above-mentioned obstacles include animals and still objects; among them, the above-mentioned animals move relative to the object, and the still objects are relatively stationary. Since the animals move relative to the objects, their movement direction and speed are uncertain, and the risk posed to the blind is relatively complex; while the still objects are relatively stationary and their movement direction and speed are consistent with the objects, they will not pose a danger to the objects, and thus the still objects are not considered when calculating the risk value.

[0075] Obtain information of the above obstacles in real time to obtain obstacle information;

[0076] Determine the risk value of the obstacle relative to the object based on the object information and obstacle information;

[0077] When the risk value is equal to or greater than a preset risk threshold, a risk avoidance strategy is generated for execution by the subject.

[0078] Taking the privacy of pedestrians into consideration, the collected images cannot identify their identities, nor are their images disseminated on the Internet, ensuring a reasonable browsing range of the images and avoiding the risk of privacy leakage.

[0079] By acquiring environmental monitoring images and object information including the target and current locations of objects, the system determines the specific location of the blind person in the environment, providing basic data for subsequent trajectory planning and obstacle avoidance. A local trajectory is planned based on the actual environment, guiding the blind person to avoid known dangerous areas (such as walls, ditches, and motorway lanes), providing preliminary safety for the blind person's travel. Obstacles along the local trajectory are captured from environmental monitoring images, and obstacle information is obtained in real time, allowing the blind person to promptly identify obstacles ahead and avoid collisions and other hazards during movement. For example, if a moving vehicle or pedestrian is in front of the blind person, the system provides timely updates of the obstacle's status, allowing the blind person to understand the latest obstacle movements and take appropriate countermeasures. By comprehensively considering multiple factors such as the blind person's location, movement speed, and distance to the obstacle, the system quantifies risks that are previously difficult to intuitively assess, generating a risk value. Risk avoidance strategies can be developed based on different risk scenarios, such as guiding the blind person to change direction, slow down, or stop moving. This improves the blind person's ability to respond to sudden dangers and effectively reduces travel risks.

[0080] In summary, the above methods are to take measures in advance before danger occurs to help blind people actively avoid obstacles, rather than passively respond after danger occurs, thereby reducing the probability of accidents occurring during the travel of blind people.

[0081] For reference, suppose a blind person is walking on a park path using a cane equipped with an environmental monitoring camera and a navigation system. The camera captures the environmental image of the park path in real time, and the blind person's current position is obtained through GPS. Based on the blind person's current position (park center, coordinate (0, 0)) and the target position (park exit, coordinate (100, 0)), the global action trajectory is planned and the current local action trajectory is intercepted. Using image recognition technology, it is detected that there is a dog, a pedestrian and a bench on the local action trajectory, and the dog and the blind person are moving towards each other (such as Figure 2 As shown in the figure, the pedestrian and the blind person move in the same direction and at the same speed. The blind person passes by a bench when moving towards the exit. Therefore, the dog and the bench are animals, and the pedestrian is a static object. At this time, the pedestrian poses no threat to the blind person and is not considered a risk (this example is an ideal state).

[0082] Second embodiment:

[0083] Based on the first embodiment, the local movement trajectory of the object is determined using the above-mentioned environmental monitoring image and object information. The specific steps include:

[0084] Generate a global action trajectory based on the current position and target position of the above object information;

[0085] Determine the monitoring area through the above environmental monitoring images;

[0086] The action trajectory passing through the monitoring area is intercepted from the above global action trajectory to obtain the local action trajectory.

[0087] The global motion trajectory is generated by the current position and target position of the object information, providing the blind with a complete path planning from the starting point to the end point; the monitoring area is determined by the environmental monitoring image, and the local area is focused on the specific environment where the blind person is currently located. The monitoring area changes according to the blind person's movement and changes in the environment, and obstacles and potential risks around the blind person (uncovered sewage wells, bicycles, electric vehicles, etc.) are perceived in the monitoring area; the interception of the local motion trajectory enables the blind person to focus more on the movement path within the current monitoring area, avoiding interference from other irrelevant parts of the global trajectory, and improving the targetedness and safety of the action; the local motion trajectory can be dynamically adjusted according to the real-time changes in the monitoring area (i.e., risk avoidance strategy), ensuring that the blind person moves along a safe and reasonable path during movement and effectively avoids sudden obstacles or risks.

[0088] Third embodiment:

[0089] Based on any of the above embodiments, the risk value of the obstacle relative to the object is determined using the above object information and obstacle information. The specific steps include:

[0090] The obstacle information includes the obstacle's moving direction, moving speed, and current position; the object information also includes the object's moving direction, moving speed, and current position;

[0091] Determine the distance between the obstacle and the object based on the current positions of the obstacle and the object;

[0092] The probability of collision between the obstacle and the object is predicted based on their current positions and movement directions; the above probability is the risk value.

[0093] For reference, the current coordinates of the blind man and the dog are (0, 0) and (10, 0) respectively. If the blind man and the dog move relative to each other in the same straight line, the collision probability is 100%; if the blind man moves to (-10, 0) and the dog moves to (100, 0), the collision probability is 0%; if the blind man moves to (10, 10) and the dog moves to (0, 10), and the blind man and the dog move at the same speed, the collision probability is 100%.

[0094] In a specific embodiment, when the risk value is greater than a preset risk threshold, a risk hedging strategy is generated for execution by the subject, and the specific steps include:

[0095] Determine the collision position of the obstacle and the object based on the moving direction, moving speed and separation distance of the obstacle and the object;

[0096] Determine the collision area based on the collision position and the preset collision increment;

[0097] Determine the avoidance time based on the collision position, the current position of the obstacle, and the moving speed of the obstacle;

[0098] The object's avoidance position is determined by moving away from the collision area based on the avoidance time, the object's current position, and the object's moving speed.

[0099] The avoidance position is output through voice information. When there is a moving object in front and an empty space on the right, the subject is prompted to turn right to avoid the moving object in front.

[0100] Based on the relative movement trend of obstacles and objects in space, potential dangerous situations are predicted. When the risk value is greater than the risk threshold, the collision position, collision area and safe area are determined, and a risk avoidance strategy is formulated based on the collision position. The risk avoidance strategy can be: when the object is in the safe area, the object waits for the obstacle to pass before moving; when the object is in the collision area, the object's movement direction is changed so that the object enters the safe area or leaves the center of the collision area to reduce the severity of the injury.

[0101] Since the risk assessment process is based on real-time object and obstacle information, the risk value can be dynamically adjusted as the environment and object status change, enabling the assistance system to reflect the current risk situation in real time, providing the blind with timely and accurate safety tips and risk avoidance guidance, ensuring that the blind are always in a relatively safe state during travel.

[0102] In a specific embodiment, the collision area is determined by the collision position and the preset collision increment, and the specific steps include:

[0103] Obtain the front, rear, left, and right endpoints of the obstacle, and superimpose the collision increments based on their positions to obtain the width and length of the collision area. Using the collision position as the midpoint, amplify the calculated width and length to obtain the collision area. Areas other than the collision area are considered safe zones.

[0104] Consider the range of obstacles that could pose a threat to the blind in all directions. An obstacle isn't just a point or a simple geometric shape; potential risks exist in all directions. Ensure that the blind are effectively protected within a defined collision zone around the obstacle, avoiding missing risk areas due to incomplete safety zone definition. The safety zone provides a clear safety boundary for the blind. The blind can adjust their movement path and speed based on the safety zone to avoid entering dangerous areas where collisions could occur. For example, when a blind person approaches an obstacle, by understanding the scope of the safety zone, they can plan a detour in advance to ensure they pass within the safe zone, effectively reducing the risk of collision with the obstacle.

[0105] For reference, track the dog's trajectory and obtain the dog's position (10, 0) and movement speed of 1.5m / s in real time. The blind person's movement speed is 0.5m / s. Calculation shows that the distance between the blind person and the dog is 10 meters, and they meet at (2.5, 0) after 5 seconds. The preset collision increment is 0.5 meters. Based on the four endpoints of the dog, it can be seen that the dog's width is 0.2 meters and the dog's length is 0.6 meters. Taking the collision position as the base point, add the dog's length in the length direction and the dog's width in the width direction to obtain a rectangular area. Then, add the collision increment to obtain the collision area. This collision area is extended outward by 0.5 meters (i.e., the collision increment) based on the rectangular area. The rest of the area except the collision area is the safe area, such as Figure 3 shown.

[0106] Since the collision position is (2.5, 0), the length of the dog is 0.6, the width of the dog is 0.2, and the collision increment is 0.5, the four endpoints of the collision area are (2.2, 0.1), (2.2, -0.1), (2.8, 0.1), and (2.8, -0.1).

[0107] The blind person moves in a direction perpendicular to the obstacle and calculates the avoidance position (0, 2.5) based on the avoidance time of 5 seconds.

[0108] Fourth embodiment:

[0109] An environment recognition system for assisting blind people in traveling, the system adopts the above-mentioned environment recognition method;

[0110] The system includes: an image acquisition device, a positioning device, a processing device and a voice device, wherein the image acquisition device, the positioning device and the voice device are connected to the processing device;

[0111] The above-mentioned image acquisition device is used to acquire environmental monitoring images;

[0112] The above positioning device is used to obtain the current position of the object in real time;

[0113] The above-mentioned voice device is used to receive the target position input by the subject and output voice information;

[0114] The processing device is used to perform the following operations:

[0115] Determine the local movement trajectory of the object through the above-mentioned environmental monitoring image and object information; and transmit the above-mentioned local movement trajectory to the object through a voice device;

[0116] Obtain obstacles on the local movement trajectory from the above-mentioned environmental monitoring image;

[0117] Obtain information of the above obstacles in real time to obtain obstacle information;

[0118] Determine the risk value of the obstacle relative to the object based on the object information and obstacle information;

[0119] When the risk value is equal to or greater than a preset risk threshold, a risk avoidance strategy to be executed by the subject is generated; and the risk avoidance strategy is transmitted to the subject via a voice device.

[0120] Fifth embodiment:

[0121] Based on the fourth embodiment, the processing device is used to determine the local movement trajectory of the object through the environmental monitoring image and object information. The specific steps include:

[0122] Generate a global action trajectory based on the current position and target position of the above object information;

[0123] Determine the monitoring area through the above environmental monitoring images;

[0124] The action trajectory passing through the monitoring area is intercepted from the above global action trajectory to obtain the local action trajectory.

[0125] Sixth embodiment:

[0126] Based on the fourth or fifth embodiment, the processing device is used to determine the risk value of the obstacle relative to the object based on the object information and the obstacle information. The specific steps include:

[0127] The obstacle information includes the obstacle's moving direction, moving speed, and current position; the object information also includes the object's moving direction, moving speed, and current position;

[0128] Determine the distance between the obstacle and the object based on the current positions of the obstacle and the object;

[0129] The probability of collision between the obstacle and the object is predicted based on their current positions and movement directions; the above probability is the risk value.

[0130] In a specific embodiment, when the risk value is greater than a preset risk threshold, a risk hedging strategy is generated for execution by the subject, and the specific steps include:

[0131] Determine the collision position of the obstacle and the object based on the moving direction, moving speed and separation distance of the obstacle and the object;

[0132] Determine the collision area based on the collision position and the preset collision increment;

[0133] Determine the avoidance time based on the collision position, the current position of the obstacle, and the moving speed of the obstacle;

[0134] The object's avoidance position is determined by moving away from the collision area based on the avoidance time, the object's current position, and the object's moving speed.

[0135] In a specific embodiment, the image acquisition device obtains the front endpoint, rear endpoint, left endpoint and right endpoint of the obstacle, and superimposes the collision increments based on the positions of the front endpoint, rear endpoint, left endpoint and right endpoint to obtain the width and length of the collision area; with the collision position as the midpoint, the calculated width and length are amplified to obtain the collision area; the other areas except the collision area are safe areas.

[0136] Sixth embodiment:

[0137] A computer-readable storage medium stores computer program instructions, which implement the above-mentioned environment recognition method when executed by a processor.

[0138] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An environment recognition method for assisting blind people in traveling, characterized in that: The following steps are involved: Acquiring an environmental monitoring image and object information, wherein the object information includes a target location and a current location of the object; Determining a local motion trajectory of the object through the environmental monitoring image and the object information; and moving the object according to the local motion trajectory; Obtaining obstacles on the local movement trajectory from the environmental monitoring image; Obtaining information of the obstacle in real time to obtain obstacle information; Determining a risk value of the obstacle relative to the object based on the object information and the obstacle information; When the risk value is equal to or greater than a preset risk threshold, a risk hedging strategy is generated for execution by the subject, specifically including: Determine the collision position of the obstacle and the object based on the moving direction, moving speed and interval distance of the obstacle and the object; Determine the collision area based on the collision position and the preset collision increment; Determine the avoidance time based on the collision position, the current position of the obstacle, and the moving speed of the obstacle; The object's avoidance position is determined by moving away from the collision area based on the avoidance time, the object's current position, and the object's moving speed. The specific steps to determine the collision area include: Obtain the front, rear, left, and right endpoints of the obstacle, and superimpose the collision increments based on their positions to obtain the width and length of the collision area. Using the collision position as the midpoint, amplify the calculated width and length to obtain the collision area. The area other than the collision area is the safe area.

2. The environment recognition method according to claim 1, characterized in that: Determining the local movement trajectory of the object using the environmental monitoring image and object information, specifically the steps include: Generate a global action trajectory using the current position and target position of the object information; Determining a monitoring area through the environmental monitoring image; The motion trajectory passing through the monitoring area is intercepted from the global motion trajectory to obtain a local motion trajectory.

3. The environment recognition method according to claim 1, characterized in that: The obstacles on the local motion trajectory include animals and still objects; wherein the animals and the object move relative to each other, and the still objects and the object are relatively stationary.

4. The environment recognition method according to claim 1, characterized in that: Determining a risk value of the obstacle relative to the object using the object information and the obstacle information, specifically comprising: The obstacle information includes the moving direction, moving speed and current position of the obstacle; the object information also includes the moving direction, moving speed and current position of the object; Determine the distance between the obstacle and the object based on the current positions of the obstacle and the object; The probability of collision between the obstacle and the object is predicted based on the current positions and movement directions of the obstacle and the object; the probability is the risk value.

5. An environment recognition system for assisting blind people in traveling, characterized in that: The system adopts the environment recognition method according to any one of claims 1 to 4; The system includes: an image acquisition device, a positioning device, a processing device and a voice device, wherein the image acquisition device, the positioning device and the voice device are connected to the processing device; The image acquisition device is used to acquire environmental monitoring images; The positioning device is used to obtain the current position of the object in real time; The voice device is used to receive the target position input by the subject and output voice information; The processing device is configured to perform the following operations: Determining a local motion trajectory of the object through the environmental monitoring image and the object information; transmitting the local motion trajectory to the object through a voice device; Obtaining obstacles on the local movement trajectory from the environmental monitoring image; Obtaining information of the obstacle in real time to obtain obstacle information; Determining a risk value of the obstacle relative to the object based on the object information and the obstacle information; When the risk value is equal to or greater than a preset risk threshold, a risk avoidance strategy to be executed by the subject is generated; and the risk avoidance strategy is transmitted to the subject via a voice device.

6. The environment recognition system according to claim 5, characterized in that: The processing device is used to determine the local movement trajectory of the object through the environmental monitoring image and object information, and the specific steps include: Generate a global action trajectory using the current position and target position of the object information; Determining a monitoring area through the environmental monitoring image; The motion trajectory passing through the monitoring area is intercepted from the global motion trajectory to obtain a local motion trajectory.

7. The environment recognition system according to claim 5, characterized in that: The processing device is used to determine the risk value of the obstacle relative to the object based on the object information and the obstacle information, and the specific steps include: The obstacle information includes the moving direction, moving speed and current position of the obstacle; the object information also includes the moving direction, moving speed and current position of the object; Determine the distance between the obstacle and the object based on the current positions of the obstacle and the object; The probability of collision between the obstacle and the object is predicted based on the current positions and movement directions of the obstacle and the object; the probability is the risk value.

8. A computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the environment recognition method according to any one of claims 1 to 4 is implemented.

Images