Evaluation of ORB-SLAM based Stereo Vision for the Aircraft Landing Status Detection

Peng, Chao-Chung; He, Rong; Chuang, Chin-Sheng

doi:10.1109/iecon49645.2022.9969111

Cited by 1 publication

(5 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Note that for Path B, which simulates heavy landings (datasets B and B_city), the results during the last part are invalid since the camera was too close to the ground and could not capture any image features. Next, we further used a reference path designed in [17] as a benchmark to judge whether the aircraft's height was safe and aligned with the center of the runway, such that the pilot would receive an alarm when the aircraft's status was at risk. As illustrated in Figure 2b, once the aircraft started to descend, a virtual landing tunnel popped up, representing the aforementioned reference path.…”

Section: Resultsmentioning

confidence: 99%

“…For real-time systems, a high update rate is also required. These simulations were performed using MATLAB, as mentioned in [17], and the system's operation time per frame is summarized in Table 8. Next, we further used a reference path designed in [17] as a benchmark to judge whether the aircraft's height was safe and aligned with the center of the runway, such that the pilot would receive an alarm when the aircraft's status was at risk.…”

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

“…Previous research [17] evaluated the performance of ORB-SLAM2 in scenes with different image feature complexities, showing that while ORB-SLAM2 offers 6DoF localization for aircraft, the feature maps generated under complex scenes include a significant number of non-ground points, increasing the difficulty of estimating the ground slope. Additionally, the accuracy of localization shows a decline in monotonous scenes, as discussed in the previous section.…”

Section: Gmm-based Landing Area Extractionmentioning

confidence: 99%

“…In this section, the Unreal Engine dataset [17] is utilized to evaluate the performance of MLESAC and IWPF. The dataset is created from two scenes with different complexities (referred to as simple and complex) and three aircraft landing paths (referred to as Ref., A, and B), resulting in a total of six scenarios.…”

Section: Comparison Between Iwpf and Mlesacmentioning

confidence: 99%

See 4 more Smart Citations

Development of an Uneven Terrain Decision-Aid Landing System for Fixed-Wing Aircraft Based on Computer Vision

Chuang,

Peng

2024

Electronics

Self Cite

View full text Add to dashboard Cite

This paper presents a computer vision-based standalone decision-aid landing system for light fixed-wing aircraft, aiming to enhance safety during emergency landings. Current landing assistance systems in airports, such as Instrument Landing Systems (ILSs) and Precision Approach Path Indicators (PAPIs), often rely on costly and location-specific ground equipment, limiting their utility for low-payload light aircraft. Especially in emergency conditions, the pilot may be forced to land on an arbitrary runway where the road flatness and glide angle cannot be ensured. To address these issues, a stereo vision-based auxiliary landing system is proposed, which is capable of estimating an appropriate glide slope based on the terrain, to assist pilots in safe landing decision-making. Moreover, in real-world scenarios, challenges with visual-based methods arise when attempting emergency landings on complex terrains with diverse objects, such as roads and buildings. This study solves this problem by employing the Gaussian Mixture Model (GMM) to segment the color image and extract ground points, while the iterative weighted plane fitting (IWPF) algorithm is introduced to mitigate the interference of outlier feature points, reaching a highly robust plane normal estimation. With the aid of the proposed system, the pilot is able to evaluate the landing glide angle/speed with respect to the uneven terrain. Simulation results demonstrate that the proposed system can successfully achieve landing guidance in unknown environments by providing glide angle estimations with an average error of less than 1 degree.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

Section: Gmm-based Landing Area Extractionmentioning

confidence: 99%

Section: Comparison Between Iwpf and Mlesacmentioning

confidence: 99%

See 3 more Smart Citations