The paper presented the concept of KSAC(Kernel Spectral Angel Cosine) to address the restrictions that the classification precision of hyperspectral imagery is very sensitive to segmentation threshold based on SAC(Spectral Angel Cosine).First,we defined the representation of KSAC,and then analyzed the effect of polynomial kernel-function parameter on KSAC,at last,we presented the method of spatial neighboring clustering based on KSAC.The experiments on the hyperspectral imagery of Shenzhen Red-forest field indicate that the threshold area coverage of the spatial neighboring clustering based on KSAC is extended up to nine times of that based on SAC.
Due to the influence of the complex background of airports and damaged areas of the runway, the existing runway extraction methods do not perform well. Furthermore, the accurate crater extraction of airport runways plays a vital role in the military fields, but there are few related studies on this topic. To solve these problems, this paper proposes an effective method for the crater extraction of runways, which mainly consists of two stages: airport runway extraction and runway crater extraction. For the previous stage, we first apply corner detection and screening strategies to runway extraction based on multiple features of the runway, such as high brightness, regional texture similarity, and shape of the runway to improve the completeness of runway extraction. In addition, the proposed method can automatically realize the complete extraction of runways with different degrees of damage. For the latter stage, the craters of the runway can be extracted by calculating the edge gradient amplitude and grayscale distribution standard deviation of the candidate areas within the runway extraction results. In four typical remote-sensing images and four post-damage remote-sensing images, the average integrity of the runway extraction reaches more than 90%. The comparative experiment results show that the extraction effect and running speed of our method are both better than those of state-of-the-art methods. In addition, the final experimental results of crater extraction show that the proposed method can effectively extract craters of airport runways, and the extraction precision and recall both reach more than 80%. Overall, our research is of great significance to the damage assessment of airport runways based on remote-sensing images in the military fields.
Existing vehicle detection methods in remote sensing images encounter challenges when detecting vehicles with large aspect ratios. Due to the big scale gap between the long edge and the short edge, large aspect ratio vehicles are hard to extract fine features. In addition, large aspect ratio results in strong orientation information and the inconsistency between regression task and classification task is even more severe. To address these issues, this paper proposes a Large Aspect Ratio Vehicles Detector (LARDet). Aiming at the difficulty of feature extraction for objects with large aspect ratios, we adopt more data augmentation and introduce PAN structure to pass through the short edge feature from shallow layer to deep layer, so as to extract more discriminative features. A lightweight Boxes Quality Predication Module (BQPM) is designed to alleviate the inconsistency between classification score and location accuracy. To alleviate the feature inconsistency between regression and classification, we further design the Align Classification Module (ACM), change the regression branch and classification branch from parallel to serial, then apply AlignConv to extract rotation-invariance feature for classification. A Large Aspect Ratio Vehicles Dataset (LAR1024) is proposed to evaluate our method. Compared with other SOTA methods, LARDet gains 5.0% AP on LAR1024 with the fastest speed of 23.9 FPS, which achieves a better speed-accuracy trade-off in the detection of large aspect ratio vehicles.
The gun intelligent aiming system can calculate the aiming point real-timely for the stationary target, but for the moving target, the shooter still needs to estimate the preaiming position by himself. In order to realize the unmanned intelligent gun system, a smart aiming method based on target trajectory prediction was proposed in this paper. Firstly, we measure the target through sensors and video images, obtain the target motion trajectory and preprocess it to obtain the target space-time trajectory sequence. Secondly, we build an ARIMA model to predict the target path, decompose the trajectory into simple sub-trajectory, which are classified by moving state and stationary state. Finally, we use the prediction algorithm based on Kalman filter to predict the trajectory position of the target at the future moment in the sub-trajectory segment combined with the target motion trend estimation. We carried out simulation experiments to simulate different motion modes of the target and constructed time-space trajectory sequences for prediction. Experimental results demonstrated that the method proposed in this paper has high prediction accuracy.
Existing the detected ellipse coefficients-driven circle-pose measurement algorithms suffer from low measurement accuracy as well as measurement failure in the presence of partial occlusions. This paper proposes a high-accuracy circle pose measurement algorithm that combing the shape prior constraint with the detected elliptical arc data and accurately obtain the image contour of the spatial circle to measure circle pose. This algorithm selects the spatial circle projection contour as the shape prior item; the algebraic detected elliptical arc distance is defined as the sum of points-to-contour distance between detected elliptical arc points and the spatial circle projection contour, which is elliptical arc data item. It defines the likelihood function based on these items so that the circle pose parameters can be obtained by minimizing the algebraic detected elliptical arc distance. The proposed algorithm avoids performing ellipse fitting; it utilizes the shape prior constraint of the spatial circle and the algebraic point-to-contour distance to effectively reduce the impacts of partial occlusions on the pose measurement results. Experiment results demonstrate that the algorithm can reliably and accurately measure the circle pose in presence of occlusions.
In order to detect ground vehicles from UAV in real time, a fast detection method using region feature gradient (RFG) was proposed. The RFG, which included the gradient mean amplitude and main directions, was used to find the potential position of target and its rotations rapidly. The gradient mean amplitude was first used to find the potential position of target, and the gradient integral image would make this process be computed very quickly. And the main directions of the target and potential position, which was computed by gradient histogram, were compared to find the rotation of the target. Finally, traditional recognition method (such as HOG) was adopted to accurately detect the target. Experimental results obtained against the VIVID database demonstrate that the proposed method is ten times faster than the improved method of HOG, and maintains a high recognition rate at the same time.
Detecting small objects and objects with large scale variants are always challenging for deep learning based object detection approaches. Many efforts have been made to solve these problems such as adopting more effective network structures, image features, loss functions, etc. However, for both small objects detection and detecting objects with various scale in single image, the first thing should be solve is the matching mechanism between anchor boxes and ground-truths. In this paper, an approach based on multi-scale balanced sampling(MB-RPN) is proposed for the difficult matching of small objects and detecting multi-scale objects. According to the scale of the anchor boxes, different positive and negative sample IOU discriminate thresholds are adopted to improve the probability of matching the small object area with the anchor boxes so that more small object samples are included in the training process. Moreover, the balanced sampling method is proposed for the collected samples, the samples are further divided and uniform sampling to ensure the diversity of samples in training process. Several datasets are adopted to evaluate the MB-RPN, the experimental results show that compare with the similar approach, MB-RPN improves detection performances effectively.
In order to improve the accuracy of object detection based on deep learning for military vehicles., an object detection algorithm based on image style transfer and domain adversarial learning is proposed. Aiming at the problem of the small number of military vehicle images., a military vehicle image dataset is constructed by using games., miniature models and spider technology. However., due to the visual difference between these images., the accuracy of the trained detection model is poor when it is directly used to detect the actual collected images. CycleGAN is used for image style transfer to obtain images with a similar style to military vehicle images at the data level. Domain adversarial learning is used to optimize the one-stage object detection algorithm to make the network learn domain invariant features and reduce the domain discrepancy at the feature level. The proposed algorithm is implemented with YOLOv5s as an example. The test results show that the proposed algorithm improves the average precision (AP) by 5.7 % without increasing the amount of inference computation., compared with the YOLOv5s algorithm.
Aiming at the vibration frequency measurement of REPDS (Reduced Pressure Drop Stick), a measurement method based on vision is proposed. Firstly, the parameters of ellipse features of top disk and top disk mark in each image of REPDS are extracted. Next, these parameters are used to measure the position and attitude of the corresponding space circle. Then, the normal vector of the REPDS at a current time is calculated by using the measured coordinates of the center of the space circle. And the attitude angle of the REPDS is calculated by the normal vector. Finally, the vibration frequency can be obtained by Fourier transform and spectrum analysis of the attitude angle sequence obtained from multiple image processing. It is proved that the method is feasible by using simulation image data. The measurement error is less than 3.3%. The method is optimized with global fixed threshold and image subsampling. Real-time measurement of vibration frequency based on the method is realized on RK3588 embedded platform, and the measurement frequency is greater than or equal to 1Hz.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)