摘要：【中文】本发明公开了一种图像分割结果优化方法、装置、智能终端及存储介质，所述方法包括：获取待分割图像以及预设的标准图像，并将所述待分割图像以及所述标准图像输入至预设的第一神经网络模型进行训练，得到初步分割结果；对第一神经网络模型进行优化，并对初步分割结果输入至预设的第二神经网络模型进行训练，得到第二分割结果；将所述标准图像、第二分割结果与所述第三分割结果输入值预设的第四神经网络模型进行训练，得到第四分割结果，所述第二分割结果、第三分割结果和第四分割结果的结合分割结果的精度高于所述初步分割结果。本发明通过图像组合训练的方式，缓解大模型被压缩后精度下降的问题，提高了压缩网络所得的分割精度。 【EN】The invention discloses an image segmentation result optimization method, an image segmentation result optimization device, an intelligent terminal and a storage medium, wherein the method comprises the following steps: acquiring an image to be segmented and a preset standard image, and inputting the image to be segmented and the standard image into a preset first neural network model for training to obtain a primary segmentation result; optimizing the first neural network model, inputting the preliminary segmentation result into a preset second neural network model for training to obtain a second segmentation result; and training a fourth neural network model preset by the standard image, the second segmentation result and the third segmentation result input value to obtain a fourth segmentation result, wherein the precision of the combined segmentation result of the second segmentation result, the third segmentation result and the fourth segmentation result is higher than that of the preliminary segmentation result. According to the invention, the problem of accuracy reduction of the compressed large model is solved by means of image combination training, and the segmentation accuracy of the compressed network is improved.

摘要：【中文】一种加速管运输保护装置，保护装置套接在加速管内部并与加速管顶端的吸极筒、加速管内部的蝶形内电极接触，包括底座和内撑结构，内撑结构相对于底座垂直设置且内撑结构的外侧壁与加速管内部的蝶形内电极相抵接，内撑结构的底端通过设置第一连接件与底座固定连接，内撑结构的顶端通过设置第二连接件嵌入加速管顶端的吸极筒内以将加速管套接固定在底座上。本发明旨在解决加速管在长途运输中因颠簸振动引起蝶形内电极发生较大移位的问题，保持在运输过程中蝶形内电极的位移量在允许范围内。 【EN】The utility model provides a protective device is transported to accelerating tube, protective device cup joint inside accelerating tube and with the utmost point section of thick bamboo of inhaling on accelerating tube top, the inside butterfly inner electrode contact of accelerating tube, including base and internal stay structure, internal stay structure sets up and the lateral wall of internal stay structure and the inside butterfly inner electrode looks butt of accelerating tube for the base is perpendicular, and the bottom of internal stay structure is through setting up first connecting piece and base fixed connection, and the top of internal stay structure is fixed on the base through setting up in the utmost point section of thick bamboo of inhaling on second connecting piece embedding accelerating tube top with higher speed tube box joint. The invention aims to solve the problem that the butterfly-shaped inner electrode is greatly displaced due to bumping and vibration in the long-distance transportation of the accelerating tube, and the displacement of the butterfly-shaped inner electrode is kept within an allowable range in the transportation process.

摘要：【中文】本发明涉及一种提高聚酯帘线与覆胶橡胶黏着力的方法，包括如下步骤：首先，对聚酯帘线进行电子束辐照；其次，对经过电子束辐照的聚酯帘线进行浸胶、覆胶，得到聚酯帘子布。该方法在二次浸胶法制备聚酯帘布工艺之前增设了电子束辐照改性工艺，可采用在线嵌入或离线辐照，对聚酯帘线进行改性，提高了聚酯帘线的活性，降低了浸胶加工成本，优化了生产工艺，提高帘线与覆胶的黏着力，使帘布层质量稳定。环境友好，工艺简单，操作方便。 【EN】The invention relates to a method for improving the adhesion of a polyester cord and rubber coating, which comprises the following steps: firstly, performing electron beam irradiation on the polyester cord; and secondly, performing gum dipping and gum covering on the polyester cord thread subjected to electron beam irradiation to obtain the polyester cord fabric. According to the method, an electron beam irradiation modification process is added before a process for preparing the polyester cord fabric by a secondary dipping method, and the polyester cord fabric can be modified by adopting online embedding or offline irradiation, so that the activity of the polyester cord fabric is improved, the dipping processing cost is reduced, the production process is optimized, the adhesion force of the cord fabric and the coating is improved, and the quality of the cord fabric is stable. Environment-friendly, simple process and convenient operation.

摘要：【中文】本发明公开了一种基于目标感知相关滤波的目标跟踪方法，其包括以下步骤：步骤1、构建深度网络特征提取模型；步骤2、确定跟踪目标及标签；步骤3、提取目标深度特征，构建通道选择相关滤波器；步骤4、滤波器逆变换，取实部并全局池化求绝对值，选择L个较大值对应的深度特征构建目标感知相关滤波器模型；步骤5、确定目标搜索区域，提取L个通道特征；步骤6、应用目标感知相关滤波器确定跟踪目标位置及尺度；步骤7、应用线性迭代更新通道选择相关滤波器；步骤8、应用线性迭代更新目标感知相关滤波器；步骤9、循环执行，实现目标连续跟踪。本发明可实现跟踪目标的自动感知，通道权重的在线自适应更新，有效增强跟踪算法的鲁棒性。 【EN】The invention discloses a target tracking method based on target perception correlation filtering, which comprises the following steps: step 1, constructing a deep network feature extraction model; step 2, determining a tracking target and a label; step 3, extracting target depth features and constructing a channel selection correlation filter; step 4, inverse transformation of the filter, namely taking a real part and solving an absolute value through global pooling, and selecting depth features corresponding to L larger values to construct a target perception correlation filter model; step 5, determining a target search area and extracting L channel characteristics; step 6, determining the position and the scale of a tracked target by applying a target perception correlation filter; step 7, updating a channel selection correlation filter by applying linear iteration; step 8, updating the target perception correlation filter by applying linear iteration; and 9, circularly executing to realize continuous target tracking. The invention can realize the automatic perception of the tracking target and the online self-adaptive updating of the channel weight, and effectively enhance the robustness of the tracking algorithm.

摘要：【中文】本发明公开了一种基于双层深度特征感知的目标跟踪方法，其包括以下步骤：步骤1、确定跟踪目标及Padding窗；步骤2、提取目标区域深度特征并加窗处理，进行第一层特征感知；步骤3、依据第一层感知特征创建特征感知相关滤波器；步骤4、计算特征感知相关滤波器的梯度图均值，选择L个较大值通道特征，完成第二层特征感知；步骤5、依据上一帧目标位置确定搜索区域尺度图像集合，提取第二层感知特征；步骤6、应用相关滤波确定目标位置及尺度；步骤7、应用跟踪结果的第一层感知特征更新特征感知相关滤波器；步骤8、循环执行步骤4、5、6、7。本发明可以解决基于像素点误差的目标感知效率较低的问题。 【EN】The invention discloses a target tracking method based on double-layer depth feature perception, which comprises the following steps: step 1, determining a tracking target and a Padding window; step 2, extracting depth features of the target area, performing windowing processing, and performing first-layer feature perception; step 3, establishing a feature perception correlation filter according to the first layer perception feature; step 4, calculating the average value of the gradient image of the feature perception correlation filter, selecting L channel features with larger values, and finishing second-layer feature perception; step 5, determining a search area scale image set according to the position of the previous frame of target, and extracting second-layer perception features; step 6, determining the position and the scale of the target by applying relevant filtering; step 7, updating the feature perception correlation filter by applying the first layer perception feature of the tracking result; and 8, circularly executing the steps 4, 5, 6 and 7. The invention can solve the problem of low target perception efficiency based on pixel point errors.

摘要：【中文】本发明公开了一种低能直流高压型电子加速器，包括钢筒组件，与钢筒组件结合设置的灯丝电源、电子枪、高压组件和加速管组件，灯丝电源为电子枪供能，高压组件具有与加速管组件的分压级分别连接的分层倍压级，加速管组件通过真空管件与扫描盒连接，扫描盒侧部设置真空泵组件构建高真空环境，以供电子束出束，扫描盒下半部设置有聚拢线圈能使发散的部分束流聚拢后加以利用。所述低能直流高压型电子加速器的体积小、成本低，电能转换率高、工作稳定可靠、检修方便。 【EN】The invention discloses a low-energy direct-current high-voltage type electron accelerator which comprises a steel cylinder assembly, and a filament power supply, an electron gun, a high-voltage assembly and an accelerating tube assembly which are combined with the steel cylinder assembly, wherein the filament power supply supplies power to the electron gun, the high-voltage assembly is provided with layered voltage doubling stages which are respectively connected with voltage division stages of the accelerating tube assembly, the accelerating tube assembly is connected with a scanning box through a vacuum tube, the vacuum pump assembly is arranged on the side part of the scanning box to construct a high vacuum environment so as to enable an electron beam to be emitted, and a gathering coil is arranged on the lower half part of the scanning box to enable scattered partial beam to be gathered. The low-energy direct-current high-voltage electron accelerator is small in size, low in cost, high in electric energy conversion rate, stable and reliable in work and convenient to overhaul.

摘要：【中文】本发明公开了一种基于多颜色空间主元分析描述的煤矸石识别方法，其包括以下步骤：步骤1、选择不同光照条件下的矸石样本图像制作样本集合；步骤2、构建多颜色空间模型；步骤3、应用主元分析方法，在多颜色空间中构建能够描述矸石的主元；步骤4、将待检测样本图像在主元方向上进行投影，构建投影图像；步骤5、对投影结果图像进行滤波、腐蚀膨胀等处理；步骤6、确定处理结果图像中的连通区域即为矸石图像区域。本发明可以实现矸石区域的有效检测，通过实验验证了该方法能够在煤与矸石的混合图像中有效的识别出矸石区域，识别精度较高，能够为煤与矸石的在线分选提供有效的图像处理技术支撑。 【EN】The invention discloses a coal gangue identification method based on multi-color space principal component analysis description, which comprises the following steps: step 1, selecting waste rock sample images under different illumination conditions to make a sample set; step 2, constructing a multi-color space model; step 3, constructing a principal component capable of describing the gangue in the multi-color space by applying a principal component analysis method; step 4, projecting the sample image to be detected in the principal component direction to construct a projected image; step 5, filtering, corroding and expanding the projection result image; and 6, determining a connected region in the processing result image as a gangue image region. The method can realize effective detection of the gangue region, and experiments prove that the method can effectively identify the gangue region in the mixed image of the coal and the gangue, has high identification precision, and can provide effective image processing technical support for online separation of the coal and the gangue.

摘要：【中文】本发明公开了一种基于CN(Color Names)和FHOG(Fused Histogram Oriented Gradient)的森林火灾在线识别方法及装置，包括以下步骤：S1、构建火焰样本图像集合；S2、将样本图像在CN颜色空间投影，并依据主元分析方法确定投影颜色空间主元；S3、构建火焰的FHOG特征，应用金字塔池化方法进行均值池化，构建FHOG特征集合；S4、采集图像，并进行主元颜色空间投影，在投影图像中应用阈值处理确定疑似火点区域；S5、计算疑似火点区域的FHOG特征并进行金字塔池化；S6、计算疑似火点区域金字塔池化FHOG特征与FHOG特征集合中样本特征的相似度，依据阈值最终确定火灾发生情况；S7、如果发生火灾，给出报警信息。本发明还提供了基于CN+FHOG特征的在线火灾识别装置。本发明精度高，能够有效的进行火灾在线识别。 【EN】The invention discloses a forest fire online identification method and a device based on CN (color names) and FHOG (fused Histogram organized gradient), comprising the following steps: s1, constructing a flame sample image set; s2, projecting the sample image in a CN color space, and determining a projection color space pivot according to a pivot analysis method; s3, constructing FHOG characteristics of the flame, performing mean pooling by applying a pyramid pooling method, and constructing an FHOG characteristic set; s4, collecting images, performing principal component color space projection, and determining a suspected fire point area by applying threshold processing in the projection images; s5, calculating FHOG characteristics of the suspected fire point area and carrying out pyramid pooling; s6, calculating the similarity between the pyramid pooling FHOG characteristics of the suspected fire point area and the sample characteristics in the FHOG characteristic set, and finally determining the fire occurrence condition according to a threshold value; and S7, if a fire disaster happens, alarming information is given. The invention also provides an online fire identification device based on the CN + FHOG characteristic. The invention has high precision and can effectively carry out on-line fire identification.

摘要：【中文】本发明涉及一种用于EB辐照印刷固化的联动控制方法，该联动控制方法的步骤为，S1、同一按钮同时开启印刷设备与EB设备；S2、印刷设备与EB设备进入准备工作状态；S3、在EB设备采集印刷设备的启动准备信号，若未启动完成，则EB设备继续采集直至印刷的启动准备完成；S4、EB设备进行高压加载并传输至印刷设备；S5、印刷设备开始印刷并将线速度信号传输至EB设备；S6、EB设备进行束流加载并辐照固化，将辐照固化信号传输至印刷设备；S7、启动后续处理及收卷工序。上述控制方法实现了信号共享、束流跟随、故障连锁等功能，保证印刷产品固化剂量的均匀性，设备运行的安全性，简便操作。 【EN】The invention relates to a linkage control method for EB irradiation printing and curing, which comprises the steps of S1, simultaneously starting a printing device and an EB device by the same button; s2, the printing equipment and the EB equipment enter a preparation working state; s3, collecting a starting preparation signal of the printing equipment on the EB equipment, if the starting is not finished, continuing to collect the signal by the EB equipment until the starting preparation of the printing is finished; s4, carrying out high-voltage loading by EB equipment and transmitting to printing equipment; s5, the printing equipment starts printing and transmits the linear velocity signal to EB equipment; s6, carrying out beam loading and irradiation curing by EB equipment, and transmitting an irradiation curing signal to printing equipment; and S7, starting subsequent treatment and winding processes. The control method realizes the functions of signal sharing, beam following, fault linkage and the like, ensures the uniformity of the curing dose of the printed product, and has the advantages of safe equipment operation and simple and convenient operation.

摘要：【中文】本发明公开了一种基于迭代法构建相关滤波响应的目标跟踪方法，包括以下步骤：S1、确定跟踪目标参数，提取融合特征并加窗处理；S2、通过频域隐式插值构建多分辨率特征图；S3、PCA方法构建特征降维矩阵；S4、应用共轭梯度法构建特征滤波器；S5、提取候选区域多尺度的融合特征并降维加窗；S6、频域隐式插值构建多分辨率特征图；S7、滤波处理获得各个通道的响应矩阵；S8、构建响应矩阵，确定跟踪结果；S9、更新GMM模型，更新滤波器；S10、循环执行S5、S6、S7、S8、S9。本发明可以有效降低特征冗余，限制较大误差的相关滤波响应对最终响应图的影响，提高跟踪方法效率。 【EN】The invention discloses a target tracking method for constructing a relevant filtering response based on an iteration method, which comprises the following steps: s1, determining tracking target parameters, extracting fusion characteristics and windowing; s2, constructing a multi-resolution characteristic diagram through frequency domain implicit interpolation; s3, constructing a feature dimension reduction matrix by a PCA method; s4, constructing a characteristic filter by applying a conjugate gradient method; s5, extracting multi-scale fusion characteristics of the candidate region, and performing dimensionality reduction and windowing; s6, constructing a multi-resolution characteristic diagram by frequency domain implicit interpolation; s7, filtering to obtain a response matrix of each channel; s8, constructing a response matrix and determining a tracking result; s9, updating the GMM model and updating the filter; s10, executing S5, S6, S7, S8 and S9 in a circulating mode. The method can effectively reduce the characteristic redundancy, limit the influence of the relevant filtering response with larger error on the final response diagram, and improve the efficiency of the tracking method.