Flexible and tunable devices have, for decades, been the subject of many research areas, including, integrated systems, flexible electronics and photonics, which are calling for advanced properties of materials. Metasurfaces with designable optical properties have now been intensively studied for flexible photonic devices. Enabled by microelectromechanical systems (MEMS) and microfluidic technologies, reconfigurable metasurfaces base on mechanical actuations offer new approaches for flexible and tunable devices. In this view, we introduce metasurfaces based on MEMS actuations, droplets arrays, flexible substrates and microfluidic systems together with their applications on flexible and tunable devices.

The trap-state density in perovskite films largely determines the photovoltaic performance of perovskite solar cells (PSCs). Increasing the crystal grain size in perovskite films is an effective method to reduce the trap-state density. Here, we have added NH4SCN into perovskite precursor solution to obtain perovskite films with an increased crystal grain size. The perovskite with increased crystal grain size shows a much lower trap-state density compared with reference perovskite films, resulting in an improved photovoltaic performance in PSCs. The champion photovoltaic device has achieved a power conversion efficiency of 19.36%. The proposed method may also impact other optoelectronic devices based on perovskite films.

All inorganic CsPbI3-xBrx perovskites have been widely used in photodetectors due to their excellent optoelectronic properties and simple preparation processes. Here, high-performance flexible photodetectors based on inorganic CsPbI3-xBrx perovskites are demonstrated, which are achieved by a modified solution-processed method. When biased at a low voltage of 10 mV, the device yielded fast response speeds (90 μs /110 μs for CsPbI2Br PDs and 100 μs/140 μs for CsPbIBr2 PDs), a high on/off ratio of 104, and a high detectivity about 1012 Jones. Meanwhile, the devices showed outstanding environmental stability and mechanical flexibility. The periodic I-t curves had negligible fluctuation (< 5%) after storing in air atmosphere for 30 days or bending for 100 times. The results indicate that CsPbI3-xBrx perovskites have great potential in photodetection areas and pave the way to achieve high-performance flexible PDs.

Motion blur, which results from time-averaging an image over the camera’s exposure time, is a common problem in microscopy of moving samples. Here, we demonstrate linear motion deblurring using temporally coded illumination in an LED array microscope. By illuminating moving objects with a well-designed temporal coded sequence that varies during each single camera exposure, the resulting motion blur is invertible and can be computationally removed. This scheme is implemented in an existing LED array microscope, providing benefits of being grayscale, fast, and adaptive, which leads to high-quality deblur performance and a flexible implementation with no moving parts. The proposed method is demonstrated experimentally for fast moving targets in a microfluidic environment.

The misalignment between RGB and thermal images significantly impairs RGB-Thermal semantic segmentation accuracy. Current non-end-to-end methods treat RGB-Thermal registration independently of semantic segmentation, resulting in fusion errors, redundant computations, and poor real-time performance. Semantic segmentation accuracy directly correlates with registration precision: better registration yields more accurate segmentation. Moreover, regions with identical semantic labels, indicating the same object, tend to share similar registration offsets. Based on these correlations, we propose an end-to-end multimodal registration and segmentation method using flexible deformation fields. Our method utilizes a shared encoder for registration and semantic segmentation to reduce redundancy. Unlike traditional non-end-to-end approaches, it directly registers high-level perceptual features, thereby optimizing computational efficiency and real-time performance. Additionally, we employ a flexible deformation field to register RGB-Thermal data, addressing limitations of traditional affine transformations in handling non-coplanar and non-rigid registrations. However, the increased flexibility of deformation fields compared to affine transformations, and the sacrificing of geometric feature preservation, pose training challenges. To overcome this, we introduce a semantic alignment loss function to train the alignment module. This function calculates the semantic segmentation loss between the predictions from registered thermal features and RGB semantic labels. It shortens the gradient backpropagation path, aligning the objectives of registration and segmentation. We validate our end-to-end approach through extensive experiments, achieving significant performance enhancements. On the IR SEG dataset, our end-to-end method achieves state-of-the-art results with a mean Intersection over Union (mIoU) of 61.1% and a mean accuracy (mAcc) of 76.0%.

Violent swing of the pantograph deteriorates contact status of the pantograph and catenary system (PCS) and seriously threatens the operation security of high-speed trains. Currently, video surveillance is commonly used for pantograph pose monitoring. Due to complex weather conditions of illumination change, occlusion, and background interference on the roof, achieving effective real-time high-precision attitude measurement remains challenging. Therefore, this study proposes a deep learning-based three dimensional (3D) measurement method. First, an accurate end-to-end pantograph keypoint location network (PKLNet) is proposed to position image coordinates of the keypoints. In PKLNet, local and global contexts are integrated by an improved transformer encoder to solve the common interference and occlusion. Differentiable numerical regression (DNR) is performed to realize an accurate coordinate regression of the keypoints. Second, the keypoint positioning is used to overcome the interference of the complex background to extract the edge, and an edge shape algorithm is employed to optimize the pose. If the edge extraction fails, Perspective-n-Point (PnP) of the keypoints is carried out for pose calculation. Evaluation of practical image sequences indicates that the average precision at PCK _p @ 0.03 is 96%. An attitude accuracy experiment further confirms that the optimized attitude RMSE (Root Mean Squared Error) is less than 0.35° and average speed is 36.2 FPS (Frames Per Second), fulfilling the strict requirement of pantograph attitude measurement.