摘要：【中文】本发明公开了一种基于外部知识的情感迁移方法，所述方法包括：步骤1)构建和训练基于情感词典D的词典匹配卷积神经网络；步骤2)对于测试集的每一个句子T，取其不重复的所有词组成T_set，将T、T_set和情感词典D输入训练好的词典匹配卷积神经网络；得到最终的分类结果。相对于手动选择转换词，本发明的方法将外部知识增加至网络中，可以避免人为制定规则带来的干扰；在使用少量的源域标注数据的情况下，本发明的方法能够得到与最好结果相似的正确率。 【EN】The invention discloses an emotion migration method based on external knowledge, which comprises the following steps: step 1) constructing and training a dictionary matching convolutional neural network based on an emotion dictionary D; step 2) for each sentence T of the test set, taking all non-repeated words to form T _ set, and inputting T, T _ set and the sentiment dictionary D into the trained dictionary matching convolutional neural network; and obtaining a final classification result. Compared with the manual selection of the conversion words, the method adds external knowledge to the network, and can avoid the interference caused by artificial rule making; in the case of using a small amount of source domain labeled data, the method of the invention can obtain the accuracy similar to the best result.

摘要：【中文】本发明公开了一种基于混响对称谱的空时自适应处理的水下目标检测方法，所述方法包括：在距离单元内，获取阵元接收的信号，进行时域采样，构成空时快拍数据，输入目标多普勒对应的降维多普勒滤波器组，输出一个降维空时快拍数据，计算目标多普勒对应的降维多普勒滤波器组对应的最优STAP权向量，将最优STAP权向量与降维空时快拍数据相乘，得到混响抑制后该距离单元的输出数据，之后将各距离单元的输出数据输入目标检测器，得到目标信号。本发明的方法通过提高小样本条件下STAP的混响抑制性能来提升目标的检测性能，使水下目标检测过程在由复杂海洋环境导致的辅助样本数不足时仍能稳定进行。 【EN】The invention discloses an underwater target detection method based on space-time adaptive processing of a reverberation symmetric spectrum, which comprises the following steps: in the distance unit, signals received by the array elements are obtained, time domain sampling is carried out, space-time snapshot data are formed, dimension-reduced Doppler filter banks corresponding to target Doppler are input, dimension-reduced space-time snapshot data are output, optimal STAP weight vectors corresponding to the dimension-reduced Doppler filter banks corresponding to the target Doppler are calculated, the optimal STAP weight vectors are multiplied with the dimension-reduced space-time snapshot data, output data of the distance unit after reverberation suppression are obtained, and then the output data of each distance unit are input to a target detector, and target signals are obtained. The method provided by the invention improves the detection performance of the target by improving the reverberation suppression performance of the STAP under the condition of small samples, so that the underwater target detection process can be stably carried out when the number of auxiliary samples is insufficient due to a complex marine environment.

摘要：【中文】本发明提出了基于卷积循环实体网络的端对端对话状态跟踪方法和系统，包括：步骤1)将对话表示成多个句子矩阵集合D＝{S₁,...S_t}，S_i,1≤i≤t为第i个由若干词向量组成的句子矩阵；步骤2)该矩阵集合D经过一个可训练的卷积神经网络模块，最大池化后得到固定长度的句子向量；步骤3)使用动态记忆将各个固定长度的句子向量进行编码，使用动态记忆的最后一个隐层h_t表示整个对话；步骤4)对于每个预先定义的语义槽，建立一层从h_t到该语义槽所有可能值的全连接的神经网络，得到各个语义槽在各个值上的概率分布；步骤5)取最大概率所在的值作为该语义槽的预测结果，得到该对话当前的对话状态。本发明可自动学习和语义槽相关的文本表示，提高了对话状态跟踪的性能。 【EN】The invention provides an end-to-end conversation state tracking method and system based on a convolution cycle entity network, comprising the following steps: step 1) representing a dialog as a plurality of sentence matrix sets D ═ S₁,...S_t}，S_iI is more than or equal to 1 and less than or equal to t is the ith sentence matrix consisting of a plurality of word vectors; step 2) the matrix set D passes through a trainable convolutional neural network module, and sentence vectors with fixed length are obtained after maximal pooling; step 3) using dynamic memory to encode each sentence vector with fixed length, using the last hidden layer h of dynamic memory_tRepresenting the entire conversation; step 4) establishing a layer of slave h for each predefined semantic slot_tObtaining probability distribution of each semantic groove on each value by a fully-connected neural network of all possible values of the semantic groove;and 5) taking the value of the maximum probability as the prediction result of the semantic slot to obtain the current conversation state of the conversation. The invention can automatically learn the text representation related to the semantic slot, and improves the performance of dialog state tracking.

摘要：【中文】本发明公开了一种基于阵列信号的多目标定位识别方法，所述方法包括：一种基于阵列信号的多目标定位识别方法，所述方法包括：步骤1)利用阵列盲信号处理将N个独立信源从混叠信号中分离出来；步骤2)对多通道的阵列信号进行频带分解，分离出M个不同频带的信号；将空间平面分为P*Q个网格，P为网格的行总数，Q为网格的列总数；步骤3)利用阵列信号处理的声源定位算法分别计算M个频带在每个网格位置上输出的功率；步骤4)基于步骤3)的每个网格的M个频带的功率，将每一个网格的信号恢复到时域，获得每个网格上的时域信号；步骤5)将步骤1)的每个独立信源分别与P*Q个时域信号进行匹配，匹配到的网格位置为独立信源的位置，共匹配P*Q*N次。 【EN】The invention discloses a multi-target positioning identification method based on array signals, which comprises the following steps: a multi-target positioning identification method based on array signals comprises the following steps: step 1) separating N independent information sources from aliasing signals by array blind signal processing; step 2) carrying out frequency band decomposition on the multi-channel array signal to separate out M signals with different frequency bands; dividing the space plane into P × Q grids, wherein P is the total number of rows of the grids, and Q is the total number of columns of the grids; step 3) respectively calculating the output power of the M frequency bands at each grid position by using a sound source positioning algorithm of array signal processing; step 4) recovering the signal of each grid to the time domain based on the power of the M frequency bands of each grid in the step 3), and obtaining a time domain signal on each grid; and 5) respectively matching each independent signal source in the step 1) with P x Q time domain signals, wherein the matched grid position is the position of the independent signal source and is matched for P x Q x N times.

摘要：【中文】本发明公开了一种基于“T”型阵的掩埋物合成孔径三维成像方法，所述方法包括：步骤1)根据设定的工作参数计算“T”型阵的声纳回波数字信号；步骤2)根据掩埋目标所在沉积层深度以及沉积层声速，以及水体深度和水体声速剖面计算平均声速，由此计算深度方向的时延参数；步骤3)对所述声纳回波数字信号进行沿航迹方向合成孔径处理，基于步骤2)的深度方向的时延参数在深度向进行脉冲压缩处理，获得目标三维成像数据。本发明的方法有效利用每一帧和每一条轨迹的数据形成二维面阵实现高分辨率的成像结果；在此基础上利用低频信号的穿透性，能够实现对掩埋物的三维成像处理，获取掩埋物的深度及方位信息。 【EN】The invention discloses a buried object synthetic aperture three-dimensional imaging method based on a T-shaped array, which comprises the following steps: step 1) calculating sonar echo digital signals of a T-shaped array according to set working parameters; step 2) calculating an average sound velocity according to the depth of a sedimentary layer where the buried target is located, the sound velocity of the sedimentary layer, the depth of the water body and the sound velocity profile of the water body, and calculating a time delay parameter in the depth direction; and 3) carrying out synthetic aperture processing on the sonar echo digital signals along the track direction, and carrying out pulse compression processing in the depth direction based on the time delay parameters in the depth direction in the step 2) to obtain target three-dimensional imaging data. The method of the invention effectively utilizes the data of each frame and each track to form a two-dimensional area array to realize the imaging result with high resolution; on the basis, the three-dimensional imaging processing of the buried object can be realized by utilizing the penetrability of the low-frequency signal, and the depth and the azimuth information of the buried object can be acquired.

摘要：【中文】本发明公开了一种近场的快速下视合成孔径三维成像方法，该方法包括：步骤1)根据声纳系统工作参数计算声纳回波数字信号，对声纳回波数字信号进行深度向脉冲压缩处理，获得柱坐标系下的深度向压缩后信号；步骤2)对深度向压缩后信号进行跨航向距离徙动矫正，并根据斜距进行数据分块，对每个数据分块进行跨航向距离矫正，获得跨航向压缩后信号；步骤3)对跨航向压缩后信号进行沿航向距离徙动矫正，由此完成沿航向合成孔径成像处理，形成最后的三维图像。本发明提出的快速成像方法，可分解成斜距向、沿航迹向和俯仰向三个方向分别成像处理，与精确时延成像算法相比，很大程度的降低了计算量，能够有效的应用于下视合成孔径成像声纳系统。 【EN】The invention discloses a near-field rapid downward-looking synthetic aperture three-dimensional imaging method, which comprises the following steps: step 1) calculating sonar echo digital signals according to working parameters of a sonar system, and performing depth-wise pulse compression processing on the sonar echo digital signals to obtain depth-wise compressed signals under a cylindrical coordinate system; step 2) performing course-crossing distance migration correction on the deeply compressed signal, performing data blocking according to the skew distance, and performing course-crossing distance correction on each data block to obtain a course-crossing compressed signal; and 3) carrying out migration correction along the course distance on the signal after the cross-course compression, thereby completing the imaging processing along the course synthetic aperture and forming a final three-dimensional image. The rapid imaging method provided by the invention can be decomposed into imaging processing in three directions of an oblique distance direction, a track direction and a pitching direction respectively, compared with an accurate time delay imaging algorithm, the calculation amount is reduced to a great extent, and the method can be effectively applied to a downward-looking synthetic aperture imaging sonar system.

摘要：【中文】本发明提供了一种用于超声断层扫描设备声束切片厚度检测的仿组织体模，所述仿组织体模呈圆台状，包括：体模外壳、声窗(3)、支护板(16)、背景仿组织材料(15)和斜面靶标；所述体模外壳由上面板(2)、下面板(8)和支撑板(6)固定连接而成，所述体模外壳与圆台侧面粘贴的声窗(3)形成一个密闭空间，密闭空间内部灌充背景仿组织材料(15)，其内嵌埋具有漫散射层的斜面靶标，所述支撑板(6)上开有灌充背景仿组织材料(15)的入口，其上贴有封堵橡皮；在所述支撑板(6)外固定支护板(16)。本发明的仿组织体模能够用于超声断层扫描设备声束切片厚度的检测和评价。 【EN】The invention provides a tissue phantom for detecting the thickness of an acoustic beam slice of an ultrasonic tomography device, which is in a round table shape and comprises: the device comprises a phantom shell, an acoustic window (3), a support guard plate (16), a background tissue imitating material (15) and an inclined plane target; the phantom comprises a phantom shell, a circular truncated cone and a circular truncated cone, wherein the phantom shell is formed by fixedly connecting an upper panel (2), a lower panel (8) and a support plate (6), the phantom shell and a sound window (3) adhered to the side surface of the circular truncated cone form a closed space, a background tissue-imitating material (15) is filled in the closed space, an inclined plane target with a diffuse scattering layer is embedded in the closed space, an inlet filled with the background tissue-imitating material (15) is formed in the support plate (6), and a plugging rubber is adhered to the inlet; and a supporting and protecting plate (16) is fixed outside the supporting plate (6). The tissue-imitated body model can be used for detecting and evaluating the thickness of an acoustic beam slice of an ultrasonic tomography device.

摘要：【中文】本发明属于海洋探测设备技术领域，具体涉及一种深海声学发射潜标，其包括依次顺序连接的主浮体(1)、压力测量仪(2)、水密电子舱(3)、深海发射换能器(4)、多个浮球(5)、深海并联声学释放器(6)、锚系重块(7)；其中，多个浮球(5)串联连接；锚系重块(7)固定在海底；主浮体(1)、压力测量仪(2)、水密电子舱(3)、深海发射换能器(4)、多个浮球(5)、深海并联声学释放器(6)、锚系重块(7)之间通过凯夫拉绳(8)连接。 【EN】The invention belongs to the technical field of ocean detection equipment, and particularly relates to a deep sea acoustic emission submerged buoy which comprises a main floating body (1), a pressure measuring instrument (2), a watertight electronic cabin (3), a deep sea emission transducer (4), a plurality of floating balls (5), a deep sea parallel acoustic releaser (6) and an anchor system heavy block (7) which are sequentially connected; wherein, a plurality of floating balls (5) are connected in series; the anchor system weight block (7) is fixed on the seabed; the main floating body (1), the pressure measuring instrument (2), the watertight electronic cabin (3), the deep sea transmitting transducer (4), the plurality of floating balls (5), the deep sea parallel acoustic releaser (6) and the anchor system weight block (7) are connected through a Kevlar rope (8).

摘要：【中文】本发明属于海洋环境和声学观测仪器技术领域，具体涉及一种深海声学接收潜标；其包括：水密电子舱和湿端配套设备；所述水密电子舱安装在湿端配套设备上，依据对应的采集时间表，所述湿端配套设备同步采集多通道的模拟水声信号，并将其传输至干端水密电子舱，水密电子舱将各通道的模拟水声信号进行数模转换，并将转换后的各通道的数字水声信号进行记录和存储；同时，湿端配套设备还包括海洋仪器设备，依据对应的采集时间表，所述海洋仪器设备同步记录海洋环境参数，并将其传输至干端水密电子舱，并将海洋环境参数进行记录和存储。 【EN】The invention belongs to the technical field of marine environments and acoustic observation instruments, and particularly relates to a deep sea acoustic receiving submerged buoy; it includes: watertight electronic cabin and wet end corollary equipment; the watertight electronic cabin is arranged on the wet end corollary equipment, the wet end corollary equipment synchronously collects multichannel analog underwater sound signals according to a corresponding collection time table and transmits the multichannel analog underwater sound signals to the dry end watertight electronic cabin, and the watertight electronic cabin performs digital-to-analog conversion on the analog underwater sound signals of each channel and records and stores the converted digital underwater sound signals of each channel; and meanwhile, the wet end corollary equipment also comprises marine instrument equipment, and according to the corresponding acquisition time table, the marine instrument equipment synchronously records marine environment parameters, transmits the marine environment parameters to the dry end watertight electronic cabin, and records and stores the marine environment parameters.

摘要：【中文】本发明公开了一种用于低温环境中的自泵式喷油内循环散热电机，所述电机包括：电机壳体(8)、定子绕组(2)、定子铁芯(3)、转子(4)和转子轴(6)，还包括：油泵(9)、喷油嘴(5)和冷凝油液(7)；所述转子轴(6)为部分中空结构，一端是动力输出端以驱动电机负载，通过轴承的轴向定位与电机壳体(8)连接，另一端与油泵(9)通过花键结构(1)连接；所述转子轴(6)上设置喷油嘴(5)；所述油泵(9)的出油口沿轴向设置，进油口连通电机底部；所述冷凝油液(7)存储在电机底部。本发明减少了电机油冷对系统的外部需求，又获得了喷油冷却的好处，功率损失小、冷却效果佳，能够适应要求电机重量小、体积小的场合。 【EN】The invention discloses a self-pumping oil injection internal circulation heat dissipation motor used in low temperature environment, which comprises: motor casing (8), stator winding (2), stator core (3), rotor (4) and rotor shaft (6), still include: the oil pump (9), the oil nozzle (5) and the condensed oil (7); the rotor shaft (6) is of a partial hollow structure, one end of the rotor shaft is a power output end to drive a motor load, the rotor shaft is connected with a motor shell (8) through axial positioning of a bearing, and the other end of the rotor shaft is connected with an oil pump (9) through a spline structure (1); an oil nozzle (5) is arranged on the rotor shaft (6); an oil outlet of the oil pump (9) is axially arranged, and an oil inlet is communicated with the bottom of the motor; the condensed oil (7) is stored at the bottom of the motor. The invention reduces the external requirement of motor oil cooling on the system, obtains the advantage of oil injection cooling, has small power loss and good cooling effect, and can be suitable for occasions requiring small weight and small volume of the motor.