To improve the accuracy of the geographic positioning of a single aerial remote sensing image, the height information of a building in the image must be considered. Oblique remote sensing images are essentially two-dimensional images and produce a large positioning error if a traditional positioning algorithm is used to locate the building directly. To address this problem, this study uses a convolutional neural network to automatically detect the location of buildings in remote sensing images. Moreover, it optimizes an automatic building recognition algorithm for oblique aerial remote sensing images based on You Only Look Once V4 (YOLO V4). This study also proposes a positioning algorithm for the building target, which uses the imaging angle to estimate the height of a building, and combines the spatial coordinate transformation matrix to calculate high-accuracy geo-location of target buildings. Simulation analysis shows that the traditional positioning algorithm inevitably leads to large errors in the positioning of building targets. When the target height is 50 m and the imaging angle is 70°, the positioning error is 114.89 m. Flight tests show that the algorithm established in this study can improve the positioning accuracy of building targets by approximately 20%–50% depending on the difference in target height.

High-precision, real-time, and long-range target geo-location is crucial to UAV reconnaissance and target strikes. Traditional geo-location methods are highly dependent on the accuracies of GPS/INS and the target elevation, which restricts the target geo-location accuracy for LRORS. Moreover, due to the limitations of laser range and the common, real time methods of improving the accuracy, such as laser range finders, DEM and geographic reference data are inappropriate for long-range UAVs. To address the above problems, a set of work patterns and a novel geo-location method are proposed in this paper. The proposed method is not restricted by conditions such as the accuracy of GPS/INS, target elevation, and range finding instrumentation. Specifically, three steps are given, to perform as follows: First, calculate the rough geo-location of the target using the traditional method. Then, according to the rough geo-location, reimage the target. Due to errors in GPS/INS and target elevation, there will be a re-projection error between the actual points of the target and the calculated projection ones. Third, a weighted filtering algorithm is proposed to obtain the optimized target geo-location by processing the reprojection error. Repeat the above process until the target geo-location estimation converges on the true value. The geo-location accuracy is improved by the work pattern and the optimization algorithm. The proposed method was verified by simulation and a flight experiment. The results showed that the proposed method can improve the geo-location accuracy by 38.8 times and 22.5 times compared with traditional methods and DEM methods, respectively. The results indicate that our method is efficient and robust, and can achieve high-precision target geo-location, with an easy implementation.

Object detection is a basic and important task in the field of aerial image processing and has gained much attention in computer vision. However, previous aerial image object-detection approaches have insufficient use of scene semantic information between different regions of large-scale aerial images. In addition, complex background and scale changes make it difficult to improve detection accuracy. To address these issues, we propose a relationship representation network for object detection in aerial images (RelationRS): (1) Firstly, multi-scale features are fused and enhanced by a dual relationship module (DRM) with conditional convolution. The dual relationship module learns the potential relationship between features of different scales and learns the relationship between different scenes from different patches in a same iteration. In addition, the dual relationship module dynamically generates parameters to guide the fusion of multi-scale features. (2) Secondly, the bridging visual representations module (BVR) is introduced into the field of aerial images to improve the object detection effect in images with complex backgrounds. Experiments with a publicly available object detection dataset for aerial images demonstrate that the proposed RelationRS achieves a state-of-the-art detection performance.