Research On Urban Vegetation Extraction Method Of LiDAR Based On Color Information

Laser radar is also called Light Detection And Ranging, short for Li DAR. Li DAR is the product of laser technology And radar technology, is a new kind of fast measurement system, one of the most promising achievements in the area of photogrammetry and remote sensing in recent years. In the research of Li DAR extraction and classification, vegetation and trees as one of the main urban surface information, play an important role in urban planning, environment protection and agroforestry development. Therefore the research of vegetation recognition and extraction has become more and more important.Due to the research of vegetation extraction has already start, research results are short of. And these algorithms have the disadvantages of high time complexity, low efficiency and call for high quality of data. So in this thesis a new vegetation extraction method applied in urban based on color information of Li DAR has been provided. Innovations and research results are as follows:1. Airborne Li DAR technology has been studied. Composition and principle of airborne Li DAR system has been introduced. The whole Li DAR data procession has been given. The standard format of Li DAR point cloud data has been presented. Then follow the preprocessing of Li DAR data, the basic principles and steps of several filter algorithms for Li DAR has been described. The TIN filter method has been used for this thesis.2. A new vegetation extraction method applied in urban based on color information of Li DAR has been researched. The algorithm is supervised, according to the RGB information and classified information, the method constructs a single peak of objective function based on statistics category. And a one-dimensional search algorithm is used to determine the RGB threshold to separate vegetation points and building points roughly. Innovation is the optimization function which is for maximum of the vegetation point number less than the threshold and building point number greater than the threshold. The function which can separate vegetation and buildings as possible, is simple and effective. The segmentation effect is good, which laid a foundation for subsequent detailed extraction.3. Based on the coarse separation results with color segmentation, some mixed discrete vegetation points among buildings are extracted utilizing the method of moving windows. After, use building-edge growth method to remove building-edge points mixed in vegetation points. Merge every step results to extract all the vegetation points. Innovation is the building-edge growth method to remove the vegetation points which are at close range with other building points. Finally use multiple sets of different experimental data for vegetation extraction experiments and analysis the results, results show that extraction rate and accuracy is above 90%.Experiments show that the vegetation extraction and segmentation algorithm based on color information of Li DAR can well extract and separation the urban vegetation information and buildings, vegetation extraction effect is good, low computing complexity and simple and convenient calculation, the extraction rate is greatly increased.