Limits...
Monocular Vision System for Fixed Altitude Flight of Unmanned Aerial Vehicles.

Huang KL, Chiu CC, Chiu SY, Teng YJ, Hao SS - Sensors (Basel) (2015)

Bottom Line: The UAV flight system can be set to fly at a fixed and relatively low altitude to obtain the same resolution of ground images.A forward-looking camera is mounted on the upside of the aircraft's nose.Experimental results show that the proposed system enables UAVs to obtain terrain images at constant resolution, and to detect the relative altitude along the flight path.

View Article: PubMed Central - PubMed

Affiliation: Department of Electrical and Electronic Engineering, Chung Cheng Institute of Technology, National Defense University, Taoyuan 33551, Taiwan. 1040510304@ndu.edu.tw.

ABSTRACT
The fastest and most economical method of acquiring terrain images is aerial photography. The use of unmanned aerial vehicles (UAVs) has been investigated for this task. However, UAVs present a range of challenges such as flight altitude maintenance. This paper reports a method that combines skyline detection with a stereo vision algorithm to enable the flight altitude of UAVs to be maintained. A monocular camera is mounted on the downside of the aircraft's nose to collect continuous ground images, and the relative altitude is obtained via a stereo vision algorithm from the velocity of the UAV. Image detection is used to obtain terrain images, and to measure the relative altitude from the ground to the UAV. The UAV flight system can be set to fly at a fixed and relatively low altitude to obtain the same resolution of ground images. A forward-looking camera is mounted on the upside of the aircraft's nose. In combination with the skyline detection algorithm, this helps the aircraft to maintain a stable flight pattern. Experimental results show that the proposed system enables UAVs to obtain terrain images at constant resolution, and to detect the relative altitude along the flight path.

No MeSH data available.


Partial calibration results under different baselines.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4541911&req=5

sensors-15-16848-f009: Partial calibration results under different baselines.

Mentions: Because the real altitudes during these experiments are very difficult to measure, an error analysis of the relative altitude calculation is provided to decrease the calculation error. Because the camera and lens are low-priced products, the captured images have a slight deformation. This image deformation causes detection errors in the distance estimation. The image deformation problem can be calibrated by the image calibration method or the calibration table. Because this study considers the problem of the execution speed, the calibration table is used to correct the measurement errors. During calibration, the same camera and lens are used to capture two images with different baselines, and the measuring targets are placed between 20 m and 40 m from the camera system. The baselines are varied every 2 cm and they range from 20 cm to 50 cm. Figure 9 presents the partial calibration results of the measurements. The x coordinates of the calibration tables represent the detected distance, which is estimated by the image detection algorithm, and the y coordinates of the calibration tables represent the actual distance. When the detected distance is obtained by the proposed system, the detected distance is the x coordinates to find the corresponding value of the y coordinates with the closest baseline.


Monocular Vision System for Fixed Altitude Flight of Unmanned Aerial Vehicles.

Huang KL, Chiu CC, Chiu SY, Teng YJ, Hao SS - Sensors (Basel) (2015)

Partial calibration results under different baselines.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4541911&req=5

sensors-15-16848-f009: Partial calibration results under different baselines.
Mentions: Because the real altitudes during these experiments are very difficult to measure, an error analysis of the relative altitude calculation is provided to decrease the calculation error. Because the camera and lens are low-priced products, the captured images have a slight deformation. This image deformation causes detection errors in the distance estimation. The image deformation problem can be calibrated by the image calibration method or the calibration table. Because this study considers the problem of the execution speed, the calibration table is used to correct the measurement errors. During calibration, the same camera and lens are used to capture two images with different baselines, and the measuring targets are placed between 20 m and 40 m from the camera system. The baselines are varied every 2 cm and they range from 20 cm to 50 cm. Figure 9 presents the partial calibration results of the measurements. The x coordinates of the calibration tables represent the detected distance, which is estimated by the image detection algorithm, and the y coordinates of the calibration tables represent the actual distance. When the detected distance is obtained by the proposed system, the detected distance is the x coordinates to find the corresponding value of the y coordinates with the closest baseline.

Bottom Line: The UAV flight system can be set to fly at a fixed and relatively low altitude to obtain the same resolution of ground images.A forward-looking camera is mounted on the upside of the aircraft's nose.Experimental results show that the proposed system enables UAVs to obtain terrain images at constant resolution, and to detect the relative altitude along the flight path.

View Article: PubMed Central - PubMed

Affiliation: Department of Electrical and Electronic Engineering, Chung Cheng Institute of Technology, National Defense University, Taoyuan 33551, Taiwan. 1040510304@ndu.edu.tw.

ABSTRACT
The fastest and most economical method of acquiring terrain images is aerial photography. The use of unmanned aerial vehicles (UAVs) has been investigated for this task. However, UAVs present a range of challenges such as flight altitude maintenance. This paper reports a method that combines skyline detection with a stereo vision algorithm to enable the flight altitude of UAVs to be maintained. A monocular camera is mounted on the downside of the aircraft's nose to collect continuous ground images, and the relative altitude is obtained via a stereo vision algorithm from the velocity of the UAV. Image detection is used to obtain terrain images, and to measure the relative altitude from the ground to the UAV. The UAV flight system can be set to fly at a fixed and relatively low altitude to obtain the same resolution of ground images. A forward-looking camera is mounted on the upside of the aircraft's nose. In combination with the skyline detection algorithm, this helps the aircraft to maintain a stable flight pattern. Experimental results show that the proposed system enables UAVs to obtain terrain images at constant resolution, and to detect the relative altitude along the flight path.

No MeSH data available.