Search results

The purpose of this paper is to solve the problem that the location of the initiation point cannot be measured accurately in the shallow underground space, this paper proposes a method, which is based on fusion of multidimensional vibration sensor information, to locate single shallow underground sources.

First, in this paper, using the characteristics of low multipath interference and good P-wave polarization in the near field, the adaptive covariance matrix algorithm is used to extract the polarization angle information of the P-wave and the short term averaging/long term averaging algorithm is used to extract the first break travel time information. Second, a hybrid positioning model based on travel time and polarization angle is constructed. Third, the positioning model is taken as the particle update fitness function of quantum-behaved particle swarm optimization and calculation is performed in the hybrid positioning model. Finally, the experiment verification is carried out in the field.

The experimental results show that, with root mean square error, spherical error probable and fitness value as evaluation indicators, the positioning performance of this method is better than that without speed prediction. And the positioning accuracy of this method has been improved by nearly 30%, giving all of the three tests a positioning error within 0.5 m and a fitness less than 1.

This method provides a new idea for high-precision positioning of shallow underground single source. It has a certain engineering application value in the fields of directional demolition of engineering blasting, water inrush and burst mud prediction, fuze position measurement, underground initiation point positioning of ammunition, mine blasting monitoring and so on.

This paper proposes a solution for recognizing human faces under mask-wearing. The lower part of human face is occluded and could not be used in the learning process of face recognition. So, the proposed solution is developed to recognize human faces on any available facial components which could be varied depending on wearing or not wearing a mask.

The proposed solution is developed based on the FaceNet framework, aiming to modify the existing facial recognition model to improve the performance of both scenarios of mask-wearing and without mask-wearing. Then, simulated masked-face images are computed on top of the original face images, to be used in the learning process of face recognition. In addition, feature heatmaps are also drawn out to visualize majority of parts of facial images that are significant in recognizing faces under mask-wearing.

The proposed method is validated using several scenarios of experiments. The result shows an outstanding accuracy of 99.2% on a scenario of mask-wearing faces. The feature heatmaps also show that non-occluded components including eyes and nose become more significant for recognizing human faces, when compared with the lower part of human faces which could be occluded under masks.

The convolutional neural network based solution is tuned up for recognizing human faces under a scenario of mask-wearing. The simulated masks on original face images are augmented for training the face recognition model. The heatmaps are then computed to prove that features generated from the top half of face images are correctly chosen for the face recognition.