Aerial walking type automatic guided vehicle equipment and operation method thereof

Abstract

A method of operating an air traveling automated guided vehicle apparatus, the air traveling automated guided vehicle apparatus comprising at least first and second carriages. The operating method includes capturing, by an image capturing device on a second compartment, a first image of a first compartment as an input of a convolutional neural network model, a regression distance between the first compartment and the second compartment is calculated to adjust a real distance between the first compartment and the second compartment, and convolutional neural network regression is performed to adjust a focus of the image capture device on the second compartment. And the distance between the first carriage and the second carriage at the time point when the image is captured is obtained through convolutional neural network regression. Regardless of the driving direction, the distance between the first compartment and the second compartment can be detected by convolutional neural network regression or a combination of convolutional neural network regression and a distance sensor module. Therefore, the operation method can be applied to any type of track and carriage configuration as long as the test image is used for training the convolutional neural network model.

Description

Technical Field

[0001] The invention relates to an aerial unmanned guided vehicle device and an operating method thereof. Background Art

[0002] In the field of aerial unmanned guided vehicle equipment, linear distance sensors and curved distance sensors are commonly used. However, the sensor can only detect the distance between two adjacent carriages when the carriages move along a linear track or a curved track, and the angle between the two carriages must be within the detectable range of the curved distance sensor. Therefore, when the angle difference between the driving directions of two adjacent carriages is not within the detectable range of the curved distance sensor, the distance between the two carriages cannot be detected.

[0003] Therefore, providing an operating method for an aerial unmanned guided vehicle device that can solve the above-mentioned problems is still a direction that needs to be developed in the industry. Summary of the invention

[0004] The invention relates to an operating method of an aerial unmanned guided vehicle equipment.

[0005] In one embodiment, an aerial unmanned guided vehicle device includes at least a first carriage and a second carriage. An operation method includes capturing a first image of the first carriage by an image capture device on the second carriage, calculating a regression distance between the first carriage and the second carriage by using the first image as an input of a convolutional neural network model, adjusting a real distance between the first carriage and the second carriage according to the regression distance, and performing convolutional neural network regression to adjust a focus of the image capture device on the second carriage according to the regression distance.

[0006] In one embodiment, the operating method of the aerial unmanned guided vehicle equipment further includes capturing image data of the first compartment by multiple cameras on the second compartment before capturing the first image of the first compartment by the image capture device on the second compartment, capturing distance data between the first compartment and the second compartment by the distance sensor module on the second compartment, and training a convolutional neural network model based on the distance data and the image data.

[0007] In one embodiment, the distance sensor module includes a linear distance sensor and a curved distance sensor, and the camera on the second car captures image data of the first car and the distance sensor module on the second car captures distance data between the first car and the second car. It also includes sensing the straight-line distance between the first car and the second car by the linear distance sensor on the second car when the first car and the second car move in a straight section of the track; and sensing the curved distance between the first car and the second car by the curved distance sensor on the second car when the first car and the second car move in a curved section of the track.

[0008] In one embodiment, the operating method of the aerial unmanned guided vehicle equipment further includes using a distance sensor module to calculate a detection distance between the first compartment and the second compartment; and comparing the regression distance and the detection distance to determine a minimum value.

[0009] In one embodiment, adjusting the actual distance between the first carriage and the second carriage according to the regression distance further includes adjusting the actual distance between the first carriage and the second carriage according to a minimum value.

[0010] In one embodiment, performing convolutional neural network regression to adjust the focus of the image capture device on the second compartment according to the regression distance further includes performing convolutional neural network regression to adjust the focus of the image capture device on the second compartment according to the minimum value.

[0011] In one embodiment, capturing image data of the first car by a camera on the second car and capturing distance data between the first car and the second car by a distance sensor module on the second car also includes capturing multiple test images of the first car by each of the cameras on the first car, wherein the test images are captured at different focal lengths.

[0012] In one embodiment, calculating the regression distance between the first carriage and the second carriage using the convolutional neural network model further includes transmitting the regression distance to the control system by a microcontroller of the image capture device after receiving a first signal sent by the control system.

[0013] In one embodiment, the first time rate at which the regression distance is calculated is different from the second time rate at which the first signal is sent to the microcontroller.

[0014] In one embodiment, the operating method of the AAV equipment further includes sending an emergency signal from the microcontroller of the image capture device to the control system of the AAV equipment when the regression distance is less than a threshold.

[0015] In one embodiment, the operating method of the aerial unmanned guided vehicle device further includes enhancing the test image, comparing the enhanced test image to obtain a comparison result, and determining one of the cameras as the first image capture device according to the comparison result.

[0016] The invention relates to an aerial unmanned transport vehicle.

[0017] In one embodiment, an aerial unmanned guided vehicle device includes a first carriage and a second carriage, each of which includes a control system, an image capture device, and a distance sensor module. The image capture device includes a microcontroller, wherein the image capture device is electrically connected to the control system. The distance sensor module is electrically connected to the control system.

[0018] In one embodiment, the aerial unmanned guided vehicle equipment further includes a linear distance sensor and a curved distance sensor electrically connected to the linear distance sensor.

[0019] In one embodiment, the first carriage and the second carriage respectively include a first portion and a second portion. The first portion includes a front side, wherein the curve distance sensor is disposed on the first portion. The second portion includes a front side, wherein the linear distance sensor is disposed on the second portion.

[0020] In one embodiment, the aerial unmanned guided vehicle equipment further includes a first reflector disposed on the first portion.

[0021] In one embodiment, the aerial unmanned guided vehicle equipment further includes a second reflector disposed on the second portion.

[0022] In one embodiment, the aerial unmanned guided vehicle equipment further includes: a plurality of cameras, which are disposed at different positions on the front side of the first carriage and the front side of the second carriage.

[0023] In the above embodiment, the image capture device on the second car is used to capture the image of the first car. The image of the first car is used as the input of the convolutional neural network model. The distance between the first car and the second car at the time point when the image is captured can be obtained by convolutional neural network regression. Therefore, regardless of the direction of travel, the distance between the first car and the second car can be detected by convolutional neural network regression or a combination of convolutional neural network regression and a distance sensor module. Therefore, as long as the convolutional neural network model is trained using a test image, the operation method of convolutional neural network regression can be applied to any type of track and car configuration. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 FIG. 1 is a side view of an aerial unmanned guided vehicle device according to an embodiment of the present invention.

[0025] Figure 2 yes Figure 1 Front view of the cabin of an Aerial Unmanned Guided Vehicle equipment.

[0026] Figure 3 yes Figure 1 Rear view of the cabin of the mid-air walking unmanned guided vehicle equipment.

[0027] Figure 4 1 is a side view of a first carriage and a second carriage of an aerial unmanned guided vehicle device according to an embodiment of the present invention.

[0028] Figure 5 yes Figure 4 Top view of the .

[0029] Figure 6 FIG. 1 is a top view of a first vehicle and a second vehicle compartment of an aerial unmanned guided vehicle device according to an embodiment of the present invention.

[0030] Figure 7 yes Figure 1 Flow chart of the operation method of the aerial unmanned guided vehicle equipment.

[0031] Figure 8 is a schematic diagram of a test image of a first compartment according to an embodiment of the present invention.

[0032] Fig. 9 yes Figure 1 A flowchart of another operating method of an aerial unmanned guided vehicle device.

[0033] Fig.10 yes Figure 1 A flowchart of another method for operating an aerial unmanned guided vehicle device. DETAILED DESCRIPTION

[0034] Figure 1 1 is a side view of an overhead unmanned guided vehicle (OHT) 100 according to an embodiment of the present invention. The overhead unmanned guided vehicle 100 includes at least one carriage 110, a control system 120, an image capture device 130, a linear distance sensor 140, and a curved distance sensor 150. The control system 120 is electrically connected to the image capture device 130, the linear distance sensor 140, and the curved distance sensor 150. The image capture device 130 includes a microcontroller 132, and the microcontroller 132 is configured to train a convolutional neural network model based on image data and perform convolutional neural network regression.

[0035] Figure 2 yes Figure 1 A front view of a carriage 110 of an aerial unmanned guided vehicle 100. The control system 120 is Figure 2 Omitted. Figure 1 and Figure 2 The carriage 110 includes a first portion 112 and a second portion 114. The first portion 112 is an upper portion of the carriage 110, and the second portion 114 is a lower portion of the carriage 110. The first portion 112 of the carriage 110 includes a front side 1122 and a rear side 1124. The second portion 114 of the carriage 110 includes a front side 1142 and a rear side 1144. The linear distance sensor 140 is disposed at the front side 1142 of the second portion 114. The curved distance sensor 150 is disposed at the front side 1122 of the first portion 112.

[0036] Figure 3 yes Figure 1 A rear view of the vehicle compartment 110 of the mid-air traveling automated guided vehicle equipment 100. Figure 1 and Figure 3 The aerial unmanned guided vehicle equipment 100 further includes a first reflector 160 and a second reflector 170 . The first reflector 160 is disposed at a rear side 1124 of the first portion 112 of the carriage 110 . The second reflector 170 is disposed at a rear side 1144 of the second portion 114 of the carriage 110 .

[0037] like Figure 2 As shown, the aerial unmanned guided vehicle device 100 also includes a plurality of cameras 180, which are arranged on the front side 1142 of the second part 114. The cameras 180 are located at different positions on the front side 1142 of the first carriage 110A and the second carriage 110B. In the present embodiment, both the camera 180 and the image capture device 130 can be used to capture images of the other carriage 110. From another perspective, the image capture device 130 is one of the plurality of cameras 180. These images are used to train the convolutional neural network model. The image capture device 130 is one of the plurality of cameras 180 that can capture the highest quality images for the convolutional neural network model. Therefore, in the convolutional neural network regression process, the image capture device 130 is used for image capture. The operation method will be described in the following paragraphs.

[0038] Figure 4 FIG. 1 is a side view of a first carriage 110A and a second carriage 110B of an aerial unmanned guided vehicle device 100 according to an embodiment of the present invention. Figure 5 yes Figure 4 A top view of the Figure 5 The control system 120, the image capture device 130 and the camera 180 are omitted. The aerial unmanned guided vehicle device 100 includes a plurality of carriages 110, each of which has the same configuration, such as Figure 1 shown. Figure 4 and Figure 5 The relative positions of the first carriage 110A and the second carriage 110B when the first carriage 110A and the second carriage 110B move in a straight section of the track are shown. The first direction D1 indicates the travel direction of the first carriage 110A and the second carriage 110B.

[0039] Reference Figure 5The first part 112 and the second part 114 can rotate relatively. When the first carriage 110A and the second carriage 110B move in the first direction D1, the front side 1142 of the second part 114 of the second carriage 110B directly faces the rear side 1144 of the second part 114 of the first carriage 110A. Therefore, the linear distance sensor 140 of the second carriage 110B directly faces the second reflector 170 of the first carriage 110A. The linear distance sensor 140 of the second carriage 110B receives the reflected light beam from the second reflector 170 of the first carriage 110A to detect the distance L1 between the first carriage 110A and the second carriage 110B.

[0040] Figure 6 FIG. 1 is a top view of a first carriage 110A and a second carriage 110B of an aerial unmanned guided vehicle device 100 according to an embodiment of the present invention. Figure 6 1 shows the relative position between the first carriage 110A and the second carriage 110B when the first carriage 110A and the second carriage 110B move in the curved section of the track. The second direction D2 indicates the travel direction of the second carriage 110B, and the third direction D3 indicates the travel direction of the first carriage 110A. The front side 1142 of the second portion 114 of the second carriage 110B does not face the rear side 1144 of the second portion 114 of the first carriage 110A. Therefore, the linear distance sensor 140 of the second carriage 110B does not face the second reflector 170 of the first carriage 110A. In this case, the front side 1122 of the first portion 112 of the second carriage 110B may face the rear side 1124 of the first portion 112 of the first carriage 110A, but the present invention is not limited thereto. Specifically, when the first carriage 110A and the second carriage 110B move in the curved section of the track, the first portion 112 rotates. When the angle difference between the second direction D2 and the third direction D3 is within the detectable range of the curve distance sensor 150, the curve distance sensor 150 of the first carriage 110A can receive the reflected light beam from the first reflector 160 of the second carriage 110B to detect the distance L2 between the first carriage 110A and the second carriage 110B.

[0041] As mentioned above, the linear distance sensor 140 and the curve distance sensor 150 constitute a distance sensor module. However, the distance sensor module can detect the distance between the first carriage 110A and the second carriage 110B only when the first carriage 110A and the second carriage 110B move along a straight section or along a curved section within the detectable range of the curve distance sensor 150. In other words, when the angle difference between the second direction D2 and the third direction D3 is not within the detectable range of the curve distance sensor 150, the distance between the first carriage 110A and the second carriage 110B cannot be detected.

[0042] Therefore, the image capture device 130 on the second carriage 110B is used to capture the image of the first carriage 110A. The image of the first carriage 110A is used as the input of the convolutional neural network model. The distance between the first carriage 110A and the second carriage 110B at the time point of capturing the image of the carriage 110 can be obtained through convolutional neural network regression. Therefore, regardless of the driving direction, the distance between the first carriage 110A and the second carriage 110B can be detected by convolutional neural network regression or a combination of convolutional neural network regression and distance sensor module.

[0043] Figure 7 yes Figure 1 The operation method 300 of the aerial unmanned guided vehicle equipment 100 is a flowchart of the operation method 300 of the aerial unmanned guided vehicle equipment 100. The operation method 300 of this embodiment uses convolutional neural network regression to detect and adjust the distance between the carriages 110 of the aerial unmanned guided vehicle equipment 100.

[0044] Reference Figure 4 and Figure 7 The operation method 300 starts at step S310 , where the plurality of cameras 180 on the second carriage 110B capture image data of the first carriage 110A, and the distance sensor module on the second carriage 110B captures distance data between the first carriage 110A and the second carriage 110B. The distance sensor module includes a linear distance sensor 140 and a curved distance sensor 150 .

[0045] In step S310, image data is obtained by capturing a plurality of test images of the first carriage 110A by each camera 180 on the second carriage 110B. The test images are obtained using different focal lengths of the camera 180. As described above, distance data is obtained by the aforementioned linear distance sensor 140 and curved distance sensor 150.

[0046] The operation method 300 continues to step S320, training a convolutional neural network model based on the distance data and the image data. The label data of the convolutional neural network model is the correlation between the image data and the corresponding distance data. The test image is used as training data for the convolutional neural network model. The test image is processed by image enhancement, such as feature detection, noise detection and sharpening process.

[0047] Figure 82 is a schematic diagram of a test image 200 of a first compartment 110A according to an embodiment of the present invention. The test image 200 includes an image of the first compartment 110A. The kernel corresponding to the feature region 210 has a larger weight to improve the feature detection result. After the image processing, the quality of the processed test images 200 is compared to obtain a comparison result. The comparison test determines the best quality test image 200, and in the following steps, the camera 180 used to capture this best quality test image 200 is selected as the image capture device 130.

[0048] The operation method 300 proceeds to step S330, where the image capturing device 130 on the second carriage 110B captures the first image of the first carriage 110A. The first image of the first carriage 110A is captured at time point 1. In this embodiment, the travel directions of the first carriage 110A and the second carriage 110B are arbitrary.

[0049] The operation method 300 continues to step S340, by using the first image as an input to the convolutional neural network model to calculate the regression distance between the first carriage 110A and the second carriage 110B. The trained convolutional neural network model provides the regression distance as an output. The regression distance is recorded in the microcontroller 132 at time point 2.

[0050] The operation method 300 continues to step S350, and the operation method 300 continues to adjust the actual distance between the first carriage 110A and the second carriage 110B according to the regression distance. In step S350, the control system 120 sends a first signal to the microcontroller 132 of the image capture device 130 at time point 3, and the microcontroller 132 transmits the regression distance recorded at time point 2 to the control system 120. The control system 120 adjusts the speed of the second carriage 110B to adjust the actual distance. Therefore, the distance between the first carriage 110A and the second carriage 110B can be adjusted to a preferred distance that can ensure the safety of the carriage 110.

[0051] The convolutional neural network regression is processed by the microcontroller 132 at a first rate. That is, the value of the regression distance is updated after a period of time. For example, the first time rate is about 40 milliseconds. The control system 120 sends the first signal at a second time rate. For example, the second time rate is about 10 milliseconds. Therefore, the first time rate for calculating the regression distance is different from the second time rate for receiving the regression distance. In addition, the operation of the microcontroller 132 and the control system 120 is not synchronized. Alternatively, the operation of the microcontroller 132 is independent of the operation of the control system 120. Whenever the actual distance needs to be adjusted, the latest updated regression distance value is obtained.

[0052] The method 300 of operation continues to step S360, where the method 300 of operation continues to perform the convolutional neural network regression to adjust the focus of the image capture device 130 on the second car 110B according to the regression distance. After the control system 120 adjusts the actual distance between the first car 110A and the second car 110B, the control system 120 sends a second signal to the microcontroller 132 to adjust the focus of the image capture device 130. The focus is updated so that the image capture device 130 can remain in a state where a higher quality image can be obtained. The control system 120 sends the second signal at a third time rate. The third time rate can be different from the first time rate and the second time rate. Whenever the focus needs to be adjusted, the most recently updated regression distance value is obtained.

[0053] During the operation of the aerial unmanned guided vehicle device 100, steps S330 to S360 are repeated. The operation method 300 can be performed regardless of the movement direction of the first carriage 110A and the second carriage 110B. Therefore, as long as the convolutional neural network model is trained using the test image 200, the operation method 300 using the convolutional neural network regression can be applied to any type of track and carriage 110 configuration.

[0054] Fig. 9 yes Figure 1 Flow chart of the operation method 400 of the aerial unmanned guided vehicle device 100. Steps S410 to S440 of the operation method 400 are substantially the same as steps S310 to S340 of the operation method 300.

[0055] In step S450 of the operation method 400 , the detection distance between the first carriage 110A and the second carriage 110B is calculated by using the distance sensor module. The detection distance can be calculated by the distance sensor module, except when the angle difference between the travel directions is not within the detectable range of the curve distance sensor 150 .

[0056] In step S460 of the operation method 400, the regression distance and the detection distance are compared and the minimum value is determined. Specifically, a shorter distance indicates a higher possibility that the second carriage 110B will collide with the first carriage 110A. Therefore, it is more reliable to use the minimum value between the regression distance and the detection distance as a criterion for adjusting the actual distance.

[0057] The convolutional neural network regression is processed by the microcontroller 132 at a first time rate. That is, the regression distance value is updated after a period of time. For example, the first time rate is about 40 milliseconds. The distance sensor module has a fourth time rate, so the detection distance value is updated after a period of time. For example, the fourth time rate is about 10 milliseconds. Therefore, the first time rate for calculating the regression distance is different from the fourth time rate for calculating the detection distance. In other words, the operation of the microcontroller 132 and the control system 120 is not synchronized. Alternatively, the operation of the microcontroller 132 is independent of the operation of the control system 120.

[0058] In step S470 of the operation method 400, the actual distance between the first carriage 110A and the second carriage 110B is adjusted according to the minimum value. The control system 120 sends a first signal to the microcontroller 132 of the image capture device 130, and the microcontroller 132 transmits the minimum value to the control system 120. The control system 120 adjusts the speed of the first carriage 110A to adjust the actual distance. Therefore, the distance between the first carriage 110A and the second carriage 110B can be adjusted to a preferred distance that can ensure the safety of the carriages 110.

[0059] In step S480 of the operation method 400, a convolutional neural network regression is performed to adjust the focus of the image capture device 130 according to the regression value. The focus is updated so that the image capture device 130 can be maintained in a state where a higher quality image can be captured.

[0060] After the control system 120 adjusts the actual distance between the first carriage 110A and the second carriage 110B, the control system 120 sends a second signal to the microcontroller 132 to adjust the focal length of the image capture device 130. The focal length is updated to the value of the newly acquired regression distance so that the image capture device 130 can be maintained in a condition where a higher quality image can be obtained.

[0061] During the operation of the AAV device 100, steps S430 to S480 are repeated. The operation method 300 can be performed regardless of the movement direction of the first carriage 110A and the second carriage 110B. Therefore, as long as the convolutional neural network model is trained using the test image 200, the operation method 300 using regression can be applied to any type of track and configuration of the carriage 110.

[0062] Fig.10 yes Figure 1Flowchart of the operating method 500 of the aerial unmanned guided vehicle equipment 100. Steps S510 to S540 are substantially the same as steps S310 to S340 of the operating method 300. In step S550 of the operating method 500, when the regression distance is less than a threshold value, the microcontroller 132 of the image capture device 130 sends an emergency signal to the control system 120 of the aerial unmanned guided vehicle equipment 100. The threshold value here represents a specific distance when the first carriage 110A and the second carriage 110B are considered to be too close. In some embodiments, when the control system 120 receives the emergency signal, the first carriage 110A and the second carriage 110B may stop to avoid a collision. In some embodiments, the speed of the first carriage 110A and the speed of the second carriage 110B are adjusted to ensure that the distance between them is at least above the threshold value. In some embodiments, step S550 of the operating method 500 may be applied to Figure 7 The operating method 300 and Fig. 9 The operating method 400 is shown.

[0063] In summary, the image capture device on the second car is used to capture the image of the first car. The image of the first car is used as the input of the convolutional neural network model. The distance between the first car and the second car at the time point when the image is captured can be obtained by convolutional neural network regression. Therefore, regardless of the direction of travel, the distance between the first car and the second car can be detected by convolutional neural network regression or a combination of convolutional neural network regression and a distance sensor module. Therefore, as long as the convolutional neural network model is trained using a test image, the regression operation method can be applied to any type of track and car configuration. In some embodiments, the minimum value between the regression distance provided by the image capture device and the detection distance provided by the distance sensor module is used as a standard to improve the reliability of the operation method. In some embodiments, an emergency signal is sent to the control system to avoid a collision.

[0064] Although the present invention has been disclosed as above in the form of implementation modes, it is not intended to limit the present invention. Those skilled in the art may make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention shall be determined by the definition of the attached claims.

[0065]

Explanation of symbols

[0066] 100: Aerial unmanned guided vehicle equipment

[0067] 110: Carriage

[0068] 110A: Carriage 1

[0069] 110B: Carriage 2

[0070] 112: Part 1

[0071] 1122: front side

[0072] 1124: Rear

[0073] 114: Part 2

[0074] 1142:Front

[0075] 1144: Rear

[0076] 120: Control system

[0077] 130: Image capture device

[0078] 132:Microcontroller

[0079] 140: Linear distance sensor

[0080] 150: Curve distance sensor

[0081] 160: First reflector

[0082] 170: Second reflector

[0083] 180: Camera

[0084] 200: Test image

[0085] 210: Feature area

[0086] 300, 400, 500: Operation method

[0087] D1: First direction

[0088] D2: Second direction

[0089] D3: Third direction

[0090] L1, L2: distance

[0091] S310, S320, S330, S340, S350, S360, S410, S420, S430, S440, S450, S460, S470, S480, S510, S520, S530, S540, S550: steps.

Claims

1. A method for operating an aerial unmanned guided vehicle, characterized in that: The aerial unmanned guided vehicle equipment includes at least a first carriage and a second carriage, and the operation method includes: capturing a first image of the first carriage by an image capturing device on the second carriage; Calculating a regression distance between the first carriage and the second carriage by using the first image as an input to a convolutional neural network model; adjusting the actual distance between the first carriage and the second carriage according to the regression distance; as well as A convolutional neural network regression is performed to adjust a focus of the image capture device on the second compartment according to the regression distance.

2. The method for operating an aerial unmanned guided vehicle according to claim 1, characterized in that: Also includes: Before the image capturing device on the second carriage captures the first image of the first carriage, the plurality of cameras on the second carriage capture image data of the first carriage, and the distance sensor module on the second carriage captures distance data between the first carriage and the second carriage; and The convolutional neural network model is trained based on the distance data and the image data.

3. The method for operating an aerial unmanned guided vehicle according to claim 2, characterized in that: The distance sensor module includes a linear distance sensor and a curved distance sensor, and the plurality of cameras on the second carriage capture the image data of the first carriage and the distance sensor module on the second carriage captures the distance data between the first carriage and the second carriage further includes: When the first carriage and the second carriage move in a straight section of the track, the linear distance between the first carriage and the second carriage is sensed by the linear distance sensor on the second carriage; as well as When the first carriage and the second carriage move in the curved section of the track, the curved distance between the first carriage and the second carriage is sensed by the curve distance sensor on the second carriage.

4. The method for operating an aerial unmanned guided vehicle according to claim 3, characterized in that: Also includes: Calculating the detection distance between the first carriage and the second carriage using the distance sensor module; and The regression distance and the detection distance are compared to determine a minimum value.

5. The method for operating an aerial unmanned guided vehicle according to claim 4, characterized in that: Wherein adjusting the actual distance between the first carriage and the second carriage according to the regression distance further comprises: The actual distance between the first carriage and the second carriage is adjusted according to the minimum value.

6. The method for operating an aerial unmanned guided vehicle according to claim 4, characterized in that: The step of executing the convolutional neural network regression to adjust the focus of the image capture device on the second carriage according to the regression distance further comprises: The convolutional neural network regression is performed to adjust the focus of the image capture device on the second carriage according to the minimum value.

7. The method for operating an aerial unmanned guided vehicle according to claim 2, characterized in that: The method of capturing the image data of the first carriage by the plurality of cameras on the second carriage and capturing the distance data between the first carriage and the second carriage by the distance sensor module on the second carriage further comprises: A plurality of test images of the first compartment are captured by each of the plurality of cameras on the first compartment, wherein the plurality of test images are captured at different focal lengths.

8. The method for operating an aerial unmanned guided vehicle according to claim 2, characterized in that: The step of calculating the regression distance between the first carriage and the second carriage using a convolutional neural network model further comprises: After receiving the first signal sent by the control system, the microcontroller of the image capture device transmits the regression distance to the control system.

9. The method for operating an aerial unmanned guided vehicle according to claim 8, characterized in that: The first time rate of calculating the regression distance is different from the second time rate of sending the first signal to the microcontroller.

10. The method for operating an aerial unmanned guided vehicle according to claim 8, characterized in that: Also includes: When the regression distance is less than a threshold, the microcontroller of the image capture device sends an emergency signal to the control system of the aerial unmanned guided vehicle device.

11. The method for operating an aerial unmanned guided vehicle according to claim 7, characterized in that: Also includes: enhancing the plurality of test images; comparing the enhanced test images to obtain a comparison result; and According to the comparison result, it is determined that one of the plurality of cameras is the first image capturing device.

12. An aerial unmanned guided vehicle device, characterized in that: Include: The first carriage and the second carriage respectively include: Control systems; an image capture device, comprising a microcontroller, wherein the image capture device is electrically connected to the control system; and A distance sensor module is electrically connected to the control system.

13. The aerial unmanned guided vehicle equipment according to claim 12, characterized in that: Also includes: Linear distance sensor; and The curved distance sensor is electrically connected to the linear distance sensor.

14. The aerial unmanned guided vehicle equipment according to claim 13, characterized in that: The first carriage and the second carriage respectively include: a first portion including a front side, wherein the curve distance sensor is disposed on the first portion; and The second portion includes a front side, wherein the linear distance sensor is disposed on the second portion.

15. The aerial unmanned guided vehicle equipment according to claim 14, characterized in that: Also includes: The first reflector is arranged on the first part.

16. The aerial unmanned guided vehicle equipment according to claim 14, characterized in that: Also includes: The second reflector is arranged on the second part.

17. The aerial unmanned guided vehicle equipment according to claim 14, characterized in that: Also includes: A plurality of cameras are disposed at different positions on the front side of the first carriage and the front side of the second carriage.

Images