Journals >Laser & Optoelectronics Progress
ing at the existing problems in the line detection for standard Hough transform, a line detection algorithm based on improved random Hough transform is proposed. The pixels of edge images are clustered and grouped by 8-neighborhood search. The concept of the pixel gradient direction difference is proposed, and the gradient direction difference between adjacent pixel is calculated in each edge group, thus the line pre-detection is carried out to exclude the edge groups without line features. Based on the theory of the random sample consensus algorithm, the improved random Hough transformation algorithm with a linear parameter pre-test model is proposed. The research results show that the proposed algorithm effectively solves the problem in standard Hough transform and improves the error detection rate in the process of line detection. The proposed algorithm has the advantages of fast detection and high detection accuracy.
.ing at the problem of object detection in remote sensing images, the Faster-Rcnn network based on the convolutional neural network models is used to extract the features of the object area. An object detection dataset containing three kinds of common targets in remote sensing images is made to train this network. In addition, in order to solve the problem of large rotation angle of remote sensing images, a target detection model with a rotation invariance self-learning ability is proposed, which integrates the spatial transformation network into the Faster R-CNN framework. By the analysis and comparison with the traditional object detection methods, the true effects of object detection in remote sensing images by different methods are explored. The features extracted by the convolutional neural networks based on the spatial transformation networks possess stronger orientation robustness than those by the traditional methods, which makes it possible to obtain a high detection precision.
.ing at the problem that the current approaches of semantic segmentation cannot meet the simultaneous demands on accuracy and efficiency in scene parsing in the intelligent vehicles, an accurate and efficient algorithm for semantic segmentation is proposed. Based on the proposed separable residual module and the down-sampling module, a real-time and accurate semantic segmentation network is designed. With the Cityscapes dataset, the segmentation accuracy can reach 67.86% on the basis of the 12 frame/s efficiency. The research results demonstrate that the proposed method can achieve a good performance both in accuracy and efficiency, and makes a balance between accuracy and efficiency.
.ing at the difficulty of binocular synchronous stereo matching under high-speed camera conditions, a fast measurement algorithm for the spatial linear motion trajectory based on binocular camera is proposed. The camera space coordinate system is established ,and the imaging model is constructed based on the principle of small hole imaging and the mapping relationship between the target point and the corresponding point. The principle of the algorithm is verified experimentally, and the method of reducing the error by traversing the input points and finding the average value of the result is obtained. The feasibility of the principle and the effectiveness of the algorithm are verified.
.