Farmland Boundary Extraction Technology Based on UAV Images

With the development of 5G and Internet of Things, more and more intelligent agricultural products are put into use, such as solar insecticidal lamps, agricultural robots, intelligent irrigation systems. The efficient operation of these intelligent products deployed in farmland depends on the improvement of ridge information. However, in the existing map system, the gap of ridge information has caused great difficulties in laying these nodes. Because it will be an extremely arduous task to use the traditional method to extract the aerial ridge map, so this paper uses the image processing method to extract the ridge information. To solve this problem, We use the edge detection method to extract the ridge. The main steps include image enhancement, image smoothing and edge detection. In this paper, many experiments have been carried out on these three parts, and the best method has been selected. However, because the traditional edge detection operator can not extract the ridge with high precision, this paper optimizes the edge detection results step by step, so as to obtain an algorithm more suitable for ridge extraction and improve the extraction accuracy. At the same time, in view of the large area of cultivated land, unmanned aerial vehicle (UAV) can not get all the images of cultivated land in the case of improving the definition, so it is necessary to stitch the images. Therefore, this paper mainly expounds the basic principles of image stitching based on feature points and image stitching based on regions, and selects an algorithm suitable for stitching farmland images through experimental verification. Finally, through the research of this paper, the vacancy of ridge information can be made up to some extent, and the workload of manual measurement can be reduced.