A Deep Learning-Based Satellite Target Recognition Method Using Radar Data.

A novel satellite target recognition method based on radar data partition and deep learning techniques is proposed in this paper. For the radar satellite recognition task, orbital altitude is introduced as a distinct and accessible feature to divide radar data. On this basis, we design a new distance metric for HRRPs called normalized angular distance divided by correlation coefficient (NADDCC), and a hierarchical clustering method based on this distance metric is applied to segment the radar observation angular domain. Using the above technology, the radar data partition is completed and multiple HRRP data clusters are obtained. To further mine the essential features in HRRPs, a GRU-SVM model is designed and firstly applied for radar HRRP target recognition. It consists of a multi-layer GRU neural network as a deep feature extractor and linear SVM as a classifier. By training, GRU neural network successfully extracts effective and highly distinguishable features of HRRPs, and feature visualization technology shows its advantages. Furthermore, the performance testing and comparison experiments also demonstrate that GRU neural network possesses better comprehensive performance for HRRP target recognition than LSTM neural network and conventional RNN, and the recognition performance of our method is almost better than that of other several common feature extraction methods or no data partition.

An automatic versatile system integrating solid-phase extraction with ultra-high performance liquid chromatography-tandem mass spectrometry using a dual-dilution strategy for direct analysis of auxins in plant extracts.

An automatic versatile system which integrated solid phase extraction (SPE) with ultra-high performance liquid chromatography-tandem mass spectrometry (UHPLC-MS/MS) was developed. Diverse commercial SPE columns can be used under an ambient pressure in this online system realized by a dual-dilution strategy. The first dilution enabled the direct injection of complex samples with minimal pretreatment, and the second dilution realized direct introduction of large volume of strong eluent into the UHPLC column without causing peak broadening or distortion. In addition, a post-column compensation mode was also designed for the matrix-effects evaluation. The features of the online system were systematically investigated, including the dilution effect, the capture of desorption solution, the column-head stacking effect and the system recovery. Compared with the offline UHPLC system, this online system showed significant advantages such as larger injection volume, higher sensitivity, shorter analysis time and better repeatability. The feasibility of the system was demonstrated by the direct analysis of three auxins from different plant tissues, including leaves of Dracaena sanderiana, buds and petals of Bauhinia. Under the optimized conditions, the whole analysis procedure took only 7min. All the correlation coefficients were greater than 0.9987, the limits of detection and the limits of quantitation were in the range of 0.560-0.800ng/g and 1.80-2.60ng/g, respectively. The recoveries of the real samples ranged from 61.0 to 117%. Finally, the post-column compensation mode was applied and no matrix-effects were observed under the analysis conditions. The automatic versatile system was rapid, sensitive and reliable. We expect this system could be extended to other target analytes in complex samples utilizing diverse SPE columns.

Magnetic solid-phase extraction using magnetic hypercrosslinked polymer for rapid determination of illegal drugs in urine.

A novel magnetic material Fe(3)O(4)/SiO(2)/P(MAA-co-VBC-co-DVB) was prepared via the hypercrosslinking of its precursor which was produced via precipitation polymerization of methacrylic acid (MAA), vinylbenzyl chloride (VBC), and divinylbenzene (DVB) in the presence of Fe(3)O(4)/SiO(2) submicrospheres with the surface containing abundant reactive double bonds. The resultant sorbent was characterized by scan electron microscopy, N(2) adsorption, and Fourier transform infrared spectroscopy. It was found that this material had remarkable features such as large surface area (500 m(2)/g) and pore volume (0.32 cm(3)/g), as well as desirable chemical composition (including hydrophobic and ion-exchange moieties). Taking advantages of the Fe(3)O(4)/SiO(2)/P(MAA-co-VBC-co-DVB), a magnetic SPE (MSPE) coupled with capillary electrophoresis (CE) method was developed for the determination of illegal drugs in urine samples. The extraction time could be clearly shortened up to 3 min. The recoveries of these drug compounds were in the range of 84.0-123% with relative standard deviations ranging between 1.7 and 10.5%; the limit of detection was in the range of 4.0-6.0 µg/L. The proposed method is simple, effective, and low-cost, and provides an accurate and sensitive detection platform for abused drug analysis.