In this paper, we propose a confocal microscopy based on dual blur depth measurement (DBCM). The first blur is defocus blur, and the second blur is artificial convolutional blur. First, the DBCM blurs the defocus image using a known Gaussian kernel and calculates the edge gradient ratio between it and the re-blurred image. Then, the axial measurement of edge positions is based on a calibration measurement curve. Finally, depth information is inferred from the edges using the original image. Experiments show that the DBCM can achieve depth measurement in a single image. In a 10×/0.25 objective, the error measured for a step sample of 4.7397 µm is 0.23 µm. The relative error rate is 4.8%. © 2023 OSA - The Optical Society. All rights reserved.

Aiming at the problems of poor anti-interference of existing pixel-level fusion rules and low efficiency of transform domain fusion rules, this study proposes a confocal microscopic multi-focus image fusion method (IGCM) based on differential confocal axial information guidance. Unlike traditional multi-focus image fusion (MFIF) methods, IGCM uses height information rather than grayscale or frequency to determine clear areas. First, the differential confocal axial measurement curve is calibrated to determine the suitable scan step u. Second, the image set required for fusion is constructed by performing a hierarchical scan of the measurement samples. Then, multiple differential image pairs are constructed using the step size u and the set of images, and the extraction area of the current reference image is decided based on the height obtained from the differential image. Finally, the regions determined by each reference image are extracted and the duplicated pixels are averaged to obtain the MFIF image. The results were that IGCM improves the interference immunity based on pixel-level image fusion compared to the maximum peak fusion method. Compared with other MFIFs, IGCM has excellent fusion efficiency while ensuring fusion clarity, which can meet the application scenario of real-time fusion and offers a new approach to panoramic depth images for confocal devices. © 2023 Optica Publishing Group.

The existing differential confocal axial three-dimensional (3D) measurement method cannot determine whether the surface height of the sample in the field of view is within its effective measurement range. Therefore, in this paper, we propose a differential confocal over-range determination method (IT-ORDM) based on an information theory to determine whether the surface height information of the sample to be examined is within the effective measurement range of the differential confocal axial measurement. First, the IT-ORDM finds the boundary posi-tion of the axial effective measurement range by the differential confocal axial light intensity response curve. Then the effective intensity measurement ranges of the pre-focus axial response curve (ARC) and the post-focus ARC are determined by the correspondence between the boundary position and the ARC. Finally, the intersection opera-tion of the pre-focus image of effective measurement and the post-focus image of effective measurement is used to realize the extraction of the effective measurement area of the differential confocal image. The experimental results show that the IT-ORDM can effectively determine and restore the 3D shape of the measured sample surface at the reference plane position in the multi sample experiments. (c) 2023 Optica Publishing Group

针对传统聚焦评价方法难以适应显微样本复杂纹理表面及反射率不均的问题，本文提出了一种抗光照强度、抗反射率不均、亚微米精度的基于双模糊理论的显微图像聚焦评价方法（Double blur micro-images focusing evaluation method,DB-FEM），并研究了人工模糊有效标准差的概念，通过理论和实验找寻了较优σ的取值。DB-FEM主要通过分析采集图像和其人为模糊图像的特征差异程度实现聚焦判断。首先，对采集的层扫图像进行人为模糊处理。然后，通过局部方差计算图像和其模糊图像在空域边缘信息和图像Haar小波频域信息的差异度。差异度包括空域边缘、低频纹理和高频边缘。最后，将所有差异度相乘得到基于差异度的聚焦评价曲线。实验结果表明，DB-FEM在不同光照幅值、不同表面形貌复杂度条件下，均具有优良的聚焦评价性能，优于现有的空域边缘和频域模态评价方法。在单焦面检测中，具有更窄的半高宽。在双焦面检测中，具有更好的分辨能力。在20x/0.65物镜条件下，轴向分辨力优于0.3μm。

In this letter, a novel chromatic differential con-focal matrix (CDCM) sensor is proposed and demonstrated for accurate and fast three-dimensional (3D) profiling without mechanical scanning. In contrast to conventional chromatic con-focal microscopy (CCM) methods that require spectral expansion and a peak-intensity-wavelength detection algorithm for axial positioning, our CDCM employs the intensity difference between a pair of spectral confocal images to accurately discriminate surface height. Due to the elimination of spectral expansion required by a conventional CCM, the CDCM has an at least one order of magnitude higher axial sensitivity and surface profiling efficiency than a conventional CCM. In the CDCM, axial scanning is achieved by residual chromatic aberration of an achromatic objective and by constructing a difference intensity signal between two spectral confocal images. Spectral confocal images are acquired in a parallel manner by using a digital micromirror device as an array of illumination pinholes and digitally applying the virtual detection pinhole concept. Quantitative evaluation experiments with the proof-of-principle sensor system indicate that with a 10xNA 0.25 objective, the CDCM can achieve better than 0.1 mu m axial accuracy with an axial measurement range of approximately 28 mu m, a lateral resolution of 0.87 mu m

为了解决视觉检测中由光照不均引起的金属线形被测物表面亮度差异变化过大而导致的灰度图像不能正确显示的问题，提出一种用于空间高曲率金属线形被测物的均匀化照明光源优化分析方法及解决方案。首先，基于光度学理论建立照明空间内任意点的理想照明模型。之后，推导该点表面照度、反射率、空间坐标信息与图像灰度之间的数学关系模型，利用MATLAB软件进行仿真模拟，对数据进行分析获取影响被测物均匀化照明成像的主要因素，在此基础上采用穹顶光源对照明光源进行改进，改进后金属线形被测物表面均匀度达92.85%。实验结果表明，该方案可以有效地提高金属线形被测物表面均匀度，较改进前提高34.77%，可以有效解决该类型被测物视觉检测中的成像问题。

The purpose of this study was to evaluate and compare three kinds of scan plane manipulation tools for real-time magnetic resonance imaging (MRI). The first one was a Line-Projection tool which was clinically used on a General Electric's MRI system (Signa((TM))) in its real-time imaging mode. The second one was called rtViewer, which was a custom designed tri-planner graphical user interface for volumetric image visualization. The third kind, using a Plane Navigator (PN) as a representative, was an arm-like mechanical armature with coordinated 6-degree-of-freedom (DoF) input capability. Two usability experiments were conducted to compare their effects in a simulated interactive MRI setting. Twelve cardiac practitioners were recruited to prescribe diagnostic views of a familiar object, i.e., a beating heart, and to prescribe an unfamiliar object, i.e., a custom designed phantom. Results indicated that the coordinated 6-DoF Plane Navigator outperformed both the Line-Projection tool and rtViewer in terms of task completion time and accuracy for prescribing both simple and complex views, though rtViewer performed best for through plane translation. The Plane Navigator saved about 50% of the time needed by the Line-Projection tool or rtViewer for prescribing unfamiliar double-oblique phantom views. Thanks to its desirable features including coordinated 6-DoF input, rich kinesthetic feedback, visual cue, and stay-put due to its static balance, the Plane Navigator tool may be used to manipulate the scan plane for fast visualization of a dynamic beating heart during real-time cardiac MRI and for the visualization of a surgical tool during an interventional MRI-guided minimum invasive surgery, or for interactive visualization of large-sized volumetric dataset including pre-acquired medical volumetric images.

为了解决视觉检测中由光照不均引起的金属线形被测物表面亮度差异变化过大而导致的灰度图像未能正确显示的问题，本文提出一种用于空间高曲率金属线形被测物的均匀化照明光源优化分析方法及解决方案。首先基于光度学理论建立照明空间内任意点的理想照明模型，之后推导该点表面照度、反射率、空间坐标信息与图像灰度之间的数学关系模型，利用MATLAB软件进行仿真模拟，对数据进行分析获取影响被测物均匀化照明成像的主要因素，在此基础上采用穹顶光源对照明光源进行改进，改进后金属线形被测物表面均匀度达92.85%。最后，实验证明了该方案可以有效地提高金属线形被测物表面均匀度，较改进前提高了34.77%，验证了所提方案可以有效解决该类型被测物的视觉检测中的成像问题。

针对应用场景中存在的运动物体会降低视觉同步定位与地图构建(SLAM)系统的定位精度和鲁棒性的问题,提出一种基于语义信息的动态环境下的视觉SLAM算法。首先,将传统视觉SLAM前端与YOLOv4目标检测算法相结合,在对输入图像进行ORB特征提取

SLAM一直是机器人领域的研究热点,近年来取得了万众瞩目的进步,但很少有SLAM算法考虑到动态场景的处理。针对视觉SLAM场景中动态目标的处理,提出一种在动态场景下的图像处理方法。将基于深度学习的语义分割算法引入到ORBSLAM2方法中,对输入图像进行分类处理的同时剔除人身上的特征点。基于已经剔除特征点的图像进行位姿估计。在TUM数据集上与ORBSLAM2进行对比,在动态场景下的绝对轨迹误差和相对路径误差精度提高了90%以上。在保证地图精度的前提下,改善了地图的适用性。