遥感相机的扫描镜通常位于光学系统的最前端，容易受到在轨外热流的影响产生非规则的面型变化，进而影响相机的成像质量。针对该问题设计了基于扫描镜温度值，以神经网络算法为基本，通过温度场变化反演扫描镜面型变化的扫描镜热变形测试方法。以该方法得出的扫描镜的面型变化可以作为后续成像系统面型修正的依据。该方法能够有效地提升遥感相机在轨的温度适用性，延长遥感相机的可用时间。通过该方法计算出的面型变化与理论面型变化的差值优于RMS 12.6 nm，具有较高的面型还原精度，验证了通过扫描镜温度值变化反演其面型变化的可行性。

In this study a commercial particle analyzer was used to image and help sorting microplastic particles (MPs) dispersed in filtrated and de-aerated tap water. The device provides a relatively easy and fast procedure for obtaining ultra-high-definition imaging, allowing the determination of shape, size, and number of 2D-projections of solid particles. The image analysis revealed clear differences among the studied different MPs originating from the grinding of five common grades of plastic sheets as they affect the image rendering differently, principally due to the light scattering either at the surface or in the volume of the microplastics. The high-quality imaging of the device also allows the discrimination of the microplastics from air bubbles with well-defined spherical shapes as well as to obtain an estimate of the size of MPs in a snapshot. We associate the differences among the shapes of the identified MPs in this study depending on the plastic type with known physical properties, such as brittleness, crystallinity, or softness. Furthermore, as a novel method we exploit a parameter based on the light intensity map from moving particles in cuvette flow to sort MPs from other particles, such as, wood fiber, human hair, and air bubbles. Using the light intensity map, which is related to the plastic-water refractive index ratio, the presence of microplastics in water can be revealed among other particles, but not their specific plastic type.

We investigate a possibility of producing the quantitative optical metrics to characterize the evolution of gastric tissue from healthy conditions via inflammation to cancer by using Mueller microscopy of gastric biopsies, regression model and statistical analysis of the predicted images. For this purpose the unstained sections of human gastric tissue biopsies at different pathological conditions were measured with the custom-built Mueller microscope. Polynomial regression model was built using the maps of transmitted intensity, retardance, dichroism and depolarization to generate the predicted images. The statistical analysis of predicted images of gastric tissue sections with multi-curve fit suggests that Mueller microscopy combined with data regression and statistical analysis is an effective approach for quantitative assessment of the degree of inflammation in gastric tissue biopsies with a high potential in clinical applications.We investigate a possibility of producing the quantitative optical metrics to characterize the evolution of gastric tissue from healthy conditions via inflammation to cancer by using Mueller microscopy of gastric biopsies, regression model and statistical analysis of the predicted images. For this purpose the unstained sections of human gastric tissue biopsies at different pathological conditions were measured with the custom-built Mueller microscope. Polynomial regression model was built using the maps of transmitted intensity, retardance, dichroism and depolarization to generate the predicted images. The statistical analysis of predicted images of gastric tissue sections with multi-curve fit suggests that Mueller microscopy combined with data regression and statistical analysis is an effective approach for quantitative assessment of the degree of inflammation in gastric tissue biopsies with a high potential in clinical applications.

针对数字液滴聚合酶链式反应（droplet digital Polymerase Chain Reaction， ddPCR）核酸检测中数字聚合酶链式反应由液滴数量多、尺寸小、排列紧密、荧光强度不均匀而导致的液滴难以计数问题，提出一种图像处理方法，基于灰度遍历法采集图像的信息，通过微分分析法对液滴进行分类和计数，可准确采集ddPCR实验的图像信息。通过卷积算法去除图像中的噪声，使用灰度分布均衡化法增强图像的对比度。以灰度遍历的方式将图像在逐个阈值下二值化，并以几何条件为限制统计液滴数量。通过微分分析法对数据进行分析，得出荧光液滴与全部液滴的计数结果。在以人类gDNA（genomic DNA）为检测样本的ddPCR实验中，该算法的平均检测准确率为99.36%，与商用仪器算法和同类算法相比分别提高了2.24%，2.53%。该方法为ddPCR实验提供了可靠的检测结果，可更好地适用于ddPCR实验。

针对目前热电池内部装配缺陷检测效率低、准确度不高的问题，研究了一种可精准分割内部电池堆图像并能够准确识别缺陷种类的方法。首先采用水平、垂直积分投影法对目标电池堆边缘特征进行提取，利用局部自适应对比度增强算法对局部不清晰部分进行细节纹理增强；然后研究了缺陷结构的灰度特性，计算提取出缺陷特征参数；最后使用BP（back propagation）神经网络和CART（classification and regression tree）决策树对特征参数分类识别，并根据分类准确度进行权重分配，将加权融合后的结果作为检测的最终判据。实验结果表明：该方法对2 000个样本的检测准确度达98.9%，为热电池的X射线缺陷检测提供了有效的途径。