Search

Highly transparent material classification can play an important role in the field of computer vision to classify glass or plastics for recycling and for home service robots to recognize transparent material. In these areas, there is a need to classify materials that are more than 73% transparent, but current transparent material classification methods cannot classify materials with full transparency levels. This paper proposes a highly transparent material classification method based on the refractive index, reflectivity, and transmissivity features from an imaging model of a time-of-flight (ToF) camera as the classification feature. First, we use the ToF camera to collect the depth and light intensity of the transparent material, as well as the scene information. The acquisition depth is distorted owing to the material characteristics of transparent materials. Second, we estimate the refractive index, reflectance, and transmittance from the depth distortion and IR (infrared rays) image. Finally, we choose a classifier that conforms to the nonlinear characteristics of the data to achieve transparent material classification. The method’s classification accuracy reached 94.1% in an experiment, indicating that our method considers the unique phenomenon of highly transparent materials reflecting against the background, incorporates this phenomenon into the ToF distance model, it can extract material features that express the characteristics of highly transparent materials, making it applicable to the classification of transparent materials at all levels of transparency. [ABSTRACT FROM AUTHOR]

The classification of materials is a research hotspot. These methods generally focus on the classification of flat materials and do not consider the influence of polishing and convex surfaces. We develop a classification algorithm of polishing and convex surface objects, and derive the photon accumulation point spread function (PAPSF) of material from the imaging model of a binocular pulsed time-of-flight (ToF) camera as the classification feature, which consists of depth distortion, the indirect reflection photon cumulant and the indirect reflection photon cumulant. We design a one-versus-all support vector machine (SVM) classifier to classify materials of polishing and convex surfaces objects. We conduct classification experiments on four plastics and four metal materials with a similar appearance. Our method in flat and raw material classification has the same classification accuracy as the latest method based on a continuous-wave- modulation ToF camera, but also our method achieved accuracies of 91.0% in flat and polishing material classification, 93.0% in different convex surface and fixed polishing material classification, 91.5% in fixed convex surface and different polishing material classification and 90.2% in polishing and convex surface material classification. [ABSTRACT FROM AUTHOR]

Identifying the locations of ice surface and bottom boundary in the radar imagery enables the calculation of ice sheet thickness, which is one of important inputs for ice-sheet modelling and global climate research. Therefore, accurate predictions of the boundaries can contribute to improve the accuracy of global climate analysis and sea level prediction. However, an accurate boundary detection in radar sounder data collected from the polar ice sheet has still been a challenge because the boundaries of the ice layer are usually very weak and noisy, and subglacial topography is highly variable. In recent years, the deep learning methods have surpassed the performances of traditional technology and helped to overcome a series of problems, including image boundary segmentation and target detection. This paper proposes a multi-scale feature fusion network (MFFN) for boundary detection of ice sheet radar echograms, where the ground truth supervises the output of the network at different stages, rather than the output of the last layer of the network. Also, a multi-scale convolution module (MCM) is introduced to learn the rich multi-scale representation of each network stage from shallow to deep, which uses convolution with different dilation rates to obtain multi-scale features. Furthermore, an improved loss function makes the proposed MFFN more effective to solve the sample imbalance problem of boundary detection, and further improves the accuracy of boundary detection. The proposed method is verified experimentally using the radar echograms from 2009 provided by the Center of Remote Sensing of Ice Sheets (CReSIS) that are used as training and test data. In the experiments, the proposed MFFN not only achieves state-of-the-art boundary detection accuracy on the test set but also improves the visual effect by generating fine boundaries. [ABSTRACT FROM AUTHOR]

We propose an innovative method of material classification based on the imaging model of pulsed time-of-flight (ToF) camera integrated with the unique signature that describes physical properties of each material named reflection point spread function (RPSF). First, the optimization method reduces the effect of material surface interreflection, which would affect RPSF and lead to decreased accuracy in classification, by alternating direction method of multipliers (ADMM). A method named feature vector normalization is proposed to extract material RPSF features. Second, according to the nonlinearity of the feature vectors, the structure of hidden layer neurons of radial basis function (RBF) neural network is optimized based on singular value decomposition (SVD) to improve generalization. Finally, the similar appearance of plastics and metals are classified on turntable-based measurement system by own design. The average classification accuracy reaches 93.3%, and the highest classification accuracy reaches 94.6%. [ABSTRACT FROM AUTHOR]