摘要：【中文】本发明提供一种基于微流控的相移器件及相移方法，该相移器件包括微型混合器和移相室，所述移相室为一透明的盒体，所述微型混合器为具有多处弯折的管体，所述微型混合器的其中一端部设有两进液口，另一端连接于移相室上；该相移方法为调整两进液口中分别通入液体的流速比，改变移相室内液体的折射率从而使经过移相室的入射光发生相移。本发明一种基于微流控的相移器件及方法的有益效果为：所述相移器件结构简单紧凑、易于集成；且结合所述相移方法可使入射光产生稳定的相移，操作简单，并能灵活、实时地调节相移大小，相移范围可从0‑2π逐渐变化。 【EN】The invention provides a microfluidic-based phase shift device and a phase shift method, wherein the phase shift device comprises a micro mixer and a phase shift chamber, the phase shift chamber is a transparent box body, the micro mixer is a pipe body with a plurality of bent parts, one end part of the micro mixer is provided with two liquid inlets, and the other end of the micro mixer is connected to the phase shift chamber; the phase shift method is to adjust the flow rate ratio of the liquid respectively introduced into the two liquid inlets, and change the refractive index of the liquid in the phase shift chamber so as to shift the phase of the incident light passing through the phase shift chamber. The phase shift device and the method based on microfluidics have the beneficial effects that: the phase shift device has a simple and compact structure and is easy to integrate; and the phase shift method is combined to generate stable phase shift for the incident light, the operation is simple, the phase shift can be flexibly adjusted in real time, and the phase shift range can be gradually changed from 0-2 pi.

摘要：【中文】本发明提供了一种表格提取方法、装置和电子设备，其中，该方法包括：提取预先获取的待处理图纸中的垂线段和水平线段；根据垂线段和水平线段的坐标值，确定垂线段和水平线段的线段交点；通过线段交点，提取多个单元格；对多个单元格进行归类，根据归类结果确定多个单元格对应的表格。该方式能够自动从待处理图纸中确定表格，无需人工干预，从而提高了表格提取的通用性。 【EN】The invention provides a form extraction method, a form extraction device and electronic equipment, wherein the method comprises the following steps: extracting vertical line segments and horizontal line segments in a pre-acquired drawing to be processed; determining a line segment intersection point of the vertical line segment and the horizontal line segment according to the coordinate values of the vertical line segment and the horizontal line segment; extracting a plurality of cells through the intersection points of the line segments; and classifying the plurality of cells, and determining a table corresponding to the plurality of cells according to the classification result. The method can automatically determine the form from the drawing to be processed without manual intervention, thereby improving the universality of form extraction.

摘要：【中文】本申请实施例公开了一种图像的排版方法、装置、介质及电子设备。该方法包括：获取构成目标图像的图像数据，并根据所述图像数据确定至少一个排版图层；根据所述排版图层的初始位置拼接得到初始的图像状态，输入至强化学习模型，得到目标图像的目标图层以及目标图层的移动步长；将所述目标图层按照目标图层的移动步长进行移动，并将移动后的图像状态输入至评分网络模型；若评分网络模型的输出评分结果符合预设标准，则将当前图像状态确定为图像的排版结果。通过运行本申请所提供的技术方案，可以通过机器学习的手段对图像进行自动排版，以达到节省对图像排版和审核过程中节省人力成本的效果。 【EN】The embodiment of the application discloses a method, a device, a medium and electronic equipment for typesetting images. The method comprises the following steps: acquiring image data forming a target image, and determining at least one typesetting layer according to the image data; splicing according to the initial position of the typesetting layer to obtain an initial image state, inputting the initial image state to a reinforcement learning model, and obtaining a target layer of a target image and a moving step length of the target layer; moving the target layer according to the moving step length of the target layer, and inputting the moved image state to a scoring network model; and if the output grading result of the grading network model meets the preset standard, determining the current image state as the typesetting result of the image. By operating the technical scheme provided by the application, the images can be automatically typeset by means of machine learning, so that the effect of saving labor cost in the image typesetting and auditing process is achieved.

摘要：【中文】本公开实施例公开了一种信息过滤方法、装置、电子设备及存储介质，该方法包括：将包含待过滤信息的原版图像输入至预先训练好的语义分割网络，得到仅包括所述待过滤信息的掩码图；将所述掩码图与所述原版图像进行对应位置的像素比对处理，以对所述掩码图中的所述待过滤信息的坐标位置进行修正，得到修正后的掩码图；基于所述修正后的掩码图，将所述原版图像中所包含的所述待过滤信息进行删除，得到目标图像。本公开实施例的技术方案，实现了对目标信息进行过滤的目的。 【EN】The embodiment of the disclosure discloses an information filtering method, an information filtering device, electronic equipment and a storage medium, wherein the method comprises the following steps: inputting an original image containing information to be filtered into a pre-trained semantic segmentation network to obtain a mask image only comprising the information to be filtered; performing pixel comparison processing on corresponding positions of the mask image and the original edition image to correct the coordinate position of the information to be filtered in the mask image to obtain a corrected mask image; and deleting the information to be filtered contained in the original image based on the corrected mask image to obtain a target image. According to the technical scheme of the embodiment of the disclosure, the purpose of filtering the target information is achieved.

摘要：【中文】本发明提供了一种图元识别的方法、装置及服务器，涉及图元识别的技术领域，在提取待识别图元对应的特征信息后，根据特征信息确定与待识别图元相匹配的标准图元；判断与待识别图元相匹配的标准图元的个数是否大于预设个数阈值；如果是，根据预构建的位置拓扑关系对待识别图元进行识别；利用图元识别的方法、装置及服务器能够对CAD图纸上的各图元进行识别，有效缓解了因人工审核存在的审图效率低以及人为失误等问题。 【EN】The invention provides a method, a device and a server for identifying a primitive, which relate to the technical field of primitive identification, and are characterized in that after characteristic information corresponding to a primitive to be identified is extracted, a standard primitive matched with the primitive to be identified is determined according to the characteristic information; judging whether the number of standard primitives matched with the primitives to be identified is greater than a preset number threshold value or not; if so, identifying the primitive to be identified according to the pre-constructed position topological relation; the method, the device and the server for identifying the primitives can be used for identifying each primitive on the CAD drawing, so that the problems of low drawing examination efficiency, human errors and the like caused by manual examination are effectively solved.

摘要：【中文】本发明公开了一种图像风格转换方法、装置、电子设备及存储介质，其中，方法包括：提取待转换图像的全局特征；根据全局特征得到待转换图像的背景特征和前景特征；根据背景特征、前景特征和目标风格转换图像对待转换图像进行风格转换，得到风格转换图像。由此，解决了未区分前景和背景，容易对图像中的内容造成过度的破坏，导致转换图像失去原有含义，风格转换的实用性较低、使用体验较低的问题。 【EN】The invention discloses an image style conversion method, an image style conversion device, electronic equipment and a storage medium, wherein the method comprises the following steps: extracting global features of an image to be converted; obtaining background features and foreground features of the image to be converted according to the global features; and performing style conversion on the image to be converted according to the background feature, the foreground feature and the target style conversion image to obtain a style conversion image. Therefore, the problems that the original meaning of the converted image is lost, the practicability of style conversion is low, and the use experience is low due to the fact that the foreground and the background are not distinguished, and the content in the image is easily damaged excessively are solved.

摘要：【中文】本发明涉及一种图纸图像定位与校对方法、装置及电子设备。所述方法包括：利用图像分割的深度学习模型计算得到第一图纸图像的目标元素的质心及第二图纸图像的目标元素的质心；根据第一图纸图像的目标元素的质心及第二图纸图像的目标元素的质心采用ransac算法计算得到第一图纸图像与第二图纸图像之间的变换矩阵；及通过变换矩阵将第一图纸图像与第二图纸图像进行对齐，再将对齐后的第一图纸图像及第二图纸图像进行相减运算获得第一图纸图像及第二图纸图像之间的差异项。本发明根据基于深度学习模型得到两张图纸图像的变换矩阵进行图纸间的定位与校对，提高了图纸图像间定位与校对的效率，避免人工校对效率低下集校对遗漏的风险的问题。 【EN】The invention relates to a drawing image positioning and correcting method, a drawing image positioning and correcting device and electronic equipment. The method comprises the following steps: calculating by using a depth learning model of image segmentation to obtain the centroid of the target element of the first drawing image and the centroid of the target element of the second drawing image; calculating by adopting a ransac algorithm according to the centroid of the target element of the first drawing image and the centroid of the target element of the second drawing image to obtain a transformation matrix between the first drawing image and the second drawing image; and aligning the first drawing image and the second drawing image through the transformation matrix, and then carrying out subtraction operation on the aligned first drawing image and the aligned second drawing image to obtain a difference item between the first drawing image and the second drawing image. According to the method, the positioning and the proofreading between the drawings are carried out according to the transformation matrix of the two drawing images obtained based on the deep learning model, so that the efficiency of the positioning and the proofreading between the drawing images is improved, and the problems of low manual proofreading efficiency and risk of proofreading omission are avoided.

摘要：【中文】本发明公开了一种海报模板的生成方法、装置、电子设备及存储介质，其中，方法包括：获取多张海报的每个图层，并且确定每个图层的命名；提取每个图层的外轮廓的位置信息，在预设图片上描绘出每个图层外轮廓，并结合每个图层的命名生成多个海报模板；利用多个海报模板训练海报生成模型，将待制作的海报模板输入至训练好的海报生成模型，生成新的海报模板。由此，解决了人工制作海报模板，导致制作过程耗时费力、效率较低，具有一定局限性，且模板布局风格单一，无法适应多个海报尺寸等问题。 【EN】The invention discloses a poster template generation method, a poster template generation device, electronic equipment and a storage medium, wherein the method comprises the following steps: acquiring each layer of a plurality of posters, and determining the name of each layer; extracting position information of the outer contour of each layer, drawing the outer contour of each layer on a preset picture, and generating a plurality of poster templates by combining the name of each layer; and training the poster generation model by utilizing the plurality of poster templates, inputting the poster template to be manufactured into the trained poster generation model, and generating a new poster template. From this, solved artifical preparation poster template, lead to that the manufacture process is consuming time hard, efficiency is lower, has certain limitation, and template layout style is single, can't adapt to a plurality of poster size scheduling problems.

摘要：【中文】本发明公开一种基于地面直连试验的冲压发动机及其燃烧室性能评估方法，包括：获取超燃冲压发动机在地面直连试验工况下试验台架推力‑时间曲线、流量参数与压强参数；基于上述参数得到发动机燃烧室入口截面与出口截面的气流参数；基于发动机燃烧室入口截面与出口截面的气流参数得到发动机燃烧室的性能评估参数；基于发动机燃烧室入口截面与出口截面的气流参数前后模拟延伸发动机的进气道和尾喷管，并对发动机性能参数进行评估。只通过单次开展地面直连试验，通过实时地测量试验台架推力与流量参数与压强参数，可以直接得到实时的发动机及其燃烧室的性能参数，为发动机性能评估提供一种快速、实时、有效的评估方法。 【EN】The invention discloses a ramjet based on a ground direct connection test and a method for evaluating the performance of a combustion chamber of the ramjet, which comprises the following steps: acquiring a thrust-time curve, a flow parameter and a pressure parameter of a test bench of the scramjet engine under the ground direct connection test condition; obtaining airflow parameters of the inlet section and the outlet section of the engine combustion chamber based on the parameters; obtaining performance evaluation parameters of the engine combustion chamber based on airflow parameters of an inlet section and an outlet section of the engine combustion chamber; and simulating and extending an air inlet channel and a tail nozzle of the engine based on the air flow parameters of the inlet section and the outlet section of the combustion chamber of the engine, and evaluating the performance parameters of the engine. The real-time performance parameters of the engine and the combustion chamber of the engine can be directly obtained by carrying out a ground direct connection test only once and measuring the thrust, flow and pressure parameters of the test bed in real time, so that a quick, real-time and effective evaluation method is provided for the performance evaluation of the engine.

摘要：【中文】本发明提出一种文本纠错方法、装置及存储介质，该方法包括对待纠错文本进行分词，得到多个分词；采用纠错模型对各分词进行逐词纠错，从而得到目标文本，其中，纠错模型是预先采用强化学习方法训练得到的。通过本发明能够有效地检测出文本中涉及上下文的错误，有效提升文本纠错效率，提升文本纠错效果。 【EN】The invention provides a text error correction method, a text error correction device and a storage medium, wherein the method comprises the steps of performing word segmentation on a text to be error corrected to obtain a plurality of word segments; and performing word-by-word error correction on each participle by adopting an error correction model so as to obtain a target text, wherein the error correction model is obtained by training in advance by adopting a reinforcement learning method. The invention can effectively detect the errors related to the context in the text, effectively improve the text error correction efficiency and improve the text error correction effect.