The plenoptic camera is a new camera structure which can record the intensity, color and the direction of the light by adding a microlens array in front of the image sensor. Based on some basic concepts of the focused plenoptic camera, we first analyze the imaging characteristics. Then, the design method was given. Furthermore, we establish the depth resolution model after detailed derivation. The imaging characteristics of the focused plenoptic camera with a novel structure of four types focal lengths microlens array was analyzed. The simulation results show that our designed camera has the advantages of large depth of field and high depth resolution. The method proposed in this paper can provide reference for designing plenoptic cameras for specific application scenarios.
Cloud is always the weak and even uninformative area inevitably existing in the remote sensing images, and greatly limits the development of remote sensing applications. Accurate and automatic detection of clouds in satellite scenes is a key problem for the application of remote sensing images. Most of the previous methods use the low-level feature of the cloud, which often generate error results especially with thin cloud or in complex scenes. In this paper, we propose a novel cloud detection method based on deep learning framework for remote sensing images. The designed deep Convolution Neural Network (CNN) which can mine the deep features of cloud consists of three convolution layers and three fully-connected layers. Using the designed network model, we can predict the probability of each image that belongs to cloud region, and then generate the cloud probability map of the image. To demonstrate the effectiveness of the method, we test it on Landsat-8 satellite images. The overall accuracy of our proposed method for cloud detection is higher than 95%. Experimental results indicate that both thin and thick cloud can be well detected with higher accuracy and robustness using our method.
Complex morphology target, which is size-varying and shape-varying, is a great challenge for infrared surveillance system. In this paper, temporal low-rank and sparse decomposition model and spatial low-rank and sparse decomposition model are designed respectively. Subsequently, a joint spatial-temporal detection method of complex morphology target is presented. Firstly, initial background subspace is obtained based on training sequence which does not contain infrared target. Secondly, temporal target image is recovered by l1 minimization after projecting orthogonal to background subspace. Thirdly, original image is decomposed into background image and spatial target image using inexact augmented Lagrange multipliers approach. Fourthly, by fusing the two target images, the possible small targets can be extracted well. Finally, background subspace is updated based on incremental singular value decomposition algorithm. The experimental results show that our method is effective and robust to detect complex morphology infrared targets. In particular, the proposed method can extract targets accurately, which is important for target recognition.