Search for: All records

When performing remote sensing image segmentation, practitioners often encounter various challenges, such as a strong imbalance in the foreground–background, the presence of tiny objects, high object density, intra-class heterogeneity, and inter-class homogeneity. To overcome these challenges, this paper introduces AerialFormer, a hybrid model that strategically combines the strengths of Transformers and Convolutional Neural Networks (CNNs). AerialFormer features a CNN Stem module integrated to preserve low-level and high-resolution features, enhancing the model’s capability to process details of aerial imagery. The proposed AerialFormer is designed with a hierarchical structure, in which a Transformer encoder generates multi-scale features and a multi-dilated CNN (MDC) decoder aggregates the information from the multi-scale inputs. As a result, information is taken into account in both local and global contexts, so that powerful representations and high-resolution segmentation can be achieved. The proposed AerialFormer was benchmarked on three benchmark datasets, including iSAID, LoveDA, and Potsdam. Comprehensive experiments and extensive ablation studies show that the proposed AerialFormer remarkably outperforms state-of-the-art methods. 

This manuscript presents airborne jet propulsion by audio sounds and ultrasounds through orifices formed by bulk-micromachining of a silicon wafer. The propeller is integrated with a small, printed circuit board (PCB) with a DC/DC converter, an oscillator, and a power amplifier, all powered by a 100F lithium-ion capacitor to make the propeller operable wirelessly. The peak propulsion force of the wireless propeller is measured to be 63.1 mg (or 618 mN) while the packaged wireless propeller’s weight is 10.6 g, including the drive electronics and adapter) when driven by 2.5kHz sinusoidal voltage with 21.4Vpp. A wired propeller (with 563 mg weight without adapter) is shown to high jump, long jump, wobbly fly, and propel objects. Also, the propeller is shown to work when driven by ultrasounds with a maximum propulsion force of 8.4 mg (82 mN) when driven by 20kHz, 20Vpp sinusoidal signal. Varying the frequency gradient of the applied sinusoidal pulses is shown to move the propeller to the left or right on demand to reach a specific location. 

This paper describes a wrist-wearable non-resonant vibrational energy harvester (1.4 cc in volume and 3.2 gram in weight, with two arrays of wound copper coils adjacent to a movable array of magnets suspended by ferrofluid bearing) for generating power from a human's walking motion. Thousand-turn coils are wound with a customized coil winding machine, and two sets of such coils are mounted on the top and bottom of a movable magnet array to obtain 20% improvement (compared to the earlier version based on an electroplated coil array) on the figure of merit (FOM) defined to be the power (delivered to a matched load) divided by the device's volume for a given acceleration of 1 g at 2 Hz.