AEROPLANE SAFE LANDING SITE DETECTION SYSTEM

Electronics & Telecommunication Engineering

Conventionally, pilots are trained to identify safe landing sites by looking to the ground with raw eyes. But human vision can be significantly affected by weather conditions. Also the crucial decision greatly depends on the pilot’s flight experience. In addition, the pilot will be under the high pressure in such extremely urgent situations. Time is another vital factor of the survival of passengers and the pilot himself. Aero plane safe landing-area detection system is invented for aero plane emergency landing, which is based on images acquired by aero plane-mounted high resolution cameras. Top-four important factors of unplanned landing (emergency landing) are aero plane hijack, running out of fuel, extremely bad weather, medical emergency. So, the proposed system focuses on the detection mechanism of safe landing area. Firstly, terrain images are acquired by aircraft-mounted cameras (Here, goggle map images are used). Each camera looks in a specific direction covering a portion of the region in front of the airplane. Second, the separate images that are acquired at the same time instant are registered and stitched together to form a larger panorama image that covers the full bigger field-of-view (FOV) in front of the airplane. Third, if the images are captured under poor illumination or weather conditions, we make use of the image enhancement method to minimize the effect of environmental factors, and improve the contrast and sharpness of images. The first three modules are necessary for getting high quality images and directly affect the performance of the subsequent module. Specifically, Fourth Module horizon detection is used to identify the ground in the image & separate the sky. The fifth module assesses the roughness of each block of the ground, which is measured by canny edge detector .Then; Sixth Segmentation & Classification module classifies the blocks into a category based on the degree of roughness. Segmentation is used to find out various clusters, performed by K-Mean Method. Principle component analysis is used for feature extraction & Classification of various clusters done by Cross correlation Classifier. Morphological operation such as dilation & sobel operator is used for extracting image components useful in the representation and description of region shape, such as boundaries, skeletons and convex hulls. For the concern of efficiency, isolated tiny spots and narrow branches of merged areas can be removed by applying the morphological operation .The length and width of the detected landing sites are measured in the eighth module. Dimension of detected smooth & large region is measure by using blob Analysis. Then, detected landing sites are sorted in descending order, based on size of area in visualization module. If the dimensions of a candidate region exceed the minimum requirement for safe landing, the potential landing-site is considered a safe candidate and is highlighted on the human machine interface. At the end of the process, the pilot makes the final decision by selecting a landing site from the recommended candidates, also considering external factors such as wind speed and wind direction, etc.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.