Research On Algorithm Of Laser Lidar Image Mosaic Based On SIFT

Streak tube imaging lidar (STIL) is an active imaging technology of flash laser.The STIL can obtain the three-dimensional geometric range image of high-rangeresolution and the intensity image of high-spatial resolution, simultaneously it canachieve passive image for glimmer. The STIL can achieve “A machine for a variety ofimage”. The STIL can provide rich target information. Due to the limitations in the fieldof view, often the image mosaic technology can make multiple images of the singlefield mosaic a complete laser lidar image of the wide field of view, and it can providethe reliable image information for the subsequent target extraction and recognition. Thescale invariant feature transform (SIFT) algorithm is a fast and effective image featureextraction algorithm in the image mosaic technology, and the distinctiveness andmatching of the feature point extracted by SIFT algorithm are very well.This paper focuses on the principle of SIFT algorithm through analyzes the basictheory of image mosaicking, and it makes the specific process of SIFT feature pointsextraction. This paper analysis the performance of SIFT algorithm through experiment,includes the illumination invariance, the scaling invariance and the rotation invariance.The experiment shows that the performance of SIFT algorithm is well. This paperimproves the image mosaicking algorithm.(1)The paper improves algorithm for solvingthe transformation matrix. The algorithm can solve the parameters of transformationmatrix by using the method of generalized inverse matrix. The improved algorithm isbetter on the error matching rate.(2)The operating efficiency of algorithm is improved.The algorithm reduces the image dimensionality by using down sampling. Theexperiment results show that the improved algorithm can improve the operatingefficiency.The original algorithm and the improved algorithm of SIFT respectively processthe streak tube laser lidar intensity image, range image and glimmer passive image. Forthe intensity image, the experimental results show that the accuracy of improvedalgorithm is better than the original algorithm. For the range image, the method ofsimplified transform matrix can achieve the range image mosaicking and obtains thematching degree of high-precision. For the glimmer passive image, the method ofsharpening pretreatment and the simplified transform matrix can achieve the mosaickingimage of high-precision through, and it can obtain the glimmer passive image of widefield.This paper focuses on the research of SIFT algorithm, and it considers of theproblems about image mosaicking, simultaneously it improves the solving method of transform matrix and the algorithm efficiency. According to the imaging characteristicof streak tube laser lidar, this paper achieves the high-precision image mosaicking ofintensity image, range image and glimmer passive image, and it indicates theeffectiveness of algorithm and the well application prospect.