An Automatic Method Of Parsing Airport Region In High Resolution Remote Sensing Image

The use of high resolution remote sensing images is becoming more and more widespread and the remote sensing image processing software technique is a key research in the world. Moreover, it is significant to recognize target in remote sensing images. As an important traffic and military target, airport detection attracts a lot of attention, so recognizing airport target exactly and highly effective has a great value.For this case, the research designes the remote sensing image processing software. This paper mainly has completed the data management library and presents an automatic method to parse the airport region by analyzing the Multi-character in high resolution remote sensing images by using the function of the software. The main research work is as follows:Design the data management library. All the RAPS source codes are programmed based on Visual C++ 2005. After analyzing the main problems in the process of remote sensing image data, the paper designes the software that has features on style of hierarchical layers and style of data abstraction. Uses forward the tile and hierarchical structure to read and write the data. To accelerate the massive image processing, uses Pyramid algorithm in the original images.By using the function of the software, this paper presents an automatic method to parse the airport region by analyzing the Multi-character in high resolution remote sensing images. Utilizing multi-resolution method, the arithmetic increases the detecting speed. First, in order to improve the detecting speed, we extract airport ROI based on contextual in low resolution image by down-sampling. Then, detecting skeleton lines of the runway nets in the ROI and extracting the straight lines. In the end, mark lines are detected in corresponding high resolution image to prove the existence of airport. This method can extract the complete runway networks fairly good, provise a priory knowledge for subsequent manipulation of detecting planes in the runway networks and improve the defects in traditional methods than can only detect the straight runway.