Research On UAV Image Mosaic Based On Reducing Data Redundancy

Usually the overlap degree of UAV images of a region is too high,the quantity of the UAV images of a region is too large,so image mosaic takes a lot of time.It's seriously affect the work efficiency.Therefore,in order to meet the needs of production,improve the efficiency of image mosaic is very important.Based on the confirmation of land right in Gansu province and the data of the unmanned aerial vehicle(UAV)images,the reduction methods of overlap degree and quantity of UAV images are analyzed and researched.The specific contents and results are as follows:1.LFirst of all,the overlap degree statistics of UAV images of the study area has be made,combined with UAV optimal overlap degree and specification requirements,we found that UAV images are full of redundancy.For the characteristics of UAV images with high overlap,the images are diluted with different intervals.The experimental results show that the larger image shrinkage interval,the less time to mosaic the image,the more error in the plane position.4-image interval make image overlap degree reach the specification threshold,mosaic time of these images is less than half of the mosaic time of all non-thinning images.2.Through the construction of a sample library containing 138 UAV images,the author uses the knowledge base method to put forward the comprehensive evaluation value of image by using the weighted average standard deviation,the average gradient and the information entropy on the basis of the 10-point system.One image-thinning method is established based on the comprehensive evaluation value of image.The time of image mosaic by this method is less than the time of total image mosaic nearly the half time,the error of image plane position is in the good control,the "white spot" phenomenon caused by excessive exposure image is significantly less than before,and image recognition is significantly improved.3.Image cutting also can reduce the image overlap and speed up the UAV image matching.At the same time,image distortion of the UAV image is larger at the edge of the image,after image cutting the edge of the image is basically cut off,which can greatly improve the registration accuracy of the image.Due to the benefits of image cutting,one method combined image cutting with image thinning is put forward to deal with the images,the time of image mosaic is less,Which reduces the influence of image edge distortion on image mosaic.