To accomplish reliable pedestrian detection using unmanned aerial vehicles (UAVs) under night-time conditions, an image enhancement method is developed to improve the low-illumination image quality. Firstly, the image brightness is mapped to a desirable level by a hyperbolic tangent curve. Secondly, the block-matching and 3-D filtering methods are developed for an unsharp filter in YCbCr colour space for image denoising and sharpening. Finally, pedestrian detection is performed using a convolutional neural network model to complete the surveillance task. Experimental results show that the Minkowski distance measurement index of enhanced images is increased to 0.975, and the detection accuracies, in F-measure and confidence coefficient, reach 0.907 and 0.840 respectively, which are the highest as compared with other image enhancement methods. This developed method has potential values for nighttime UAV visual monitoring in smart city applications.

Machine vision based object detection techniques have been widely used in product inspection, defect detection and dimension measurement. These techniques have largely improved the efficiency of industrial production and increased the level of production autonomy. However, demands on advance hardware design and image processing algorithms are needed for the quality inspection of a largecomplex- surface. In order to solve this problem, a hybrid surface defect detection method is developed. An image of the product surface is first divided into two areas: background with similar features and special pattern area, such as product trademarks. For the background area, defects have significant differences in gray intensity from the normal area. Fault detection is conducted using a gradient threshold segmentation method that can limit segmentation errors arising from uneven illuminations. For the special pattern area, image registration and image difference are adopted to detect defects, which are adaptive to irregular image contents with discontinuous shapes and appearances. Experimental results indicate that the proposed method achieves about 1.21 times and 2.94 times higher accuracy, in F-measure, for large-complex-surface defect detection than traditional methods of gradient threshold segmentation and template matching respectively. The proposed image processing technique can be applied in product quality inspections.

On-line images of wear debris contain important information for real-time condition monitoring, and a dynamic imaging technique can eliminate particle overlaps commonly found in static images, for instance, acquired using ferrography. However, dynamic wear debris images captured in a running machine are unavoidably blurred because the particles in lubricant are in motion. Hence, it is difficult to acquire reliable images of wear debris with an adequate resolution for particle feature extraction. In order to obtain sharp wear particle images, an image processing approach is proposed. Blurred particles were firstly separated from the static background by utilizing a background subtraction method. Second, the point spread function was estimated using power cepstrum to determine the blur direction and length. Then, the Wiener filter algorithm was adopted to perform image restoration to improve the image quality. Finally, experiments were conducted with a large number of dynamic particle images to validate the effectiveness of the proposed method and the performance of the approach was also evaluated. This study provides a new practical approach to acquire clear images for on-line wear monitoring.

This paper proposes a position estimation method for a planar switched reluctance motor (PSRM). In the method, a second-order sliding mode observer (SMO) is used to achieve sensorless control of a PSRM for the first time. A sensorless closed-loop control strategy based on the SMO without a position sensor for the PSRM is constructed. The SMO mainly consists of a flux linkage estimation, an adaptive current estimation, an observing error calculation, and a position estimation section. An adaptive current observer is applied in the current estimation section to minimize the error between the measured and estimated currents and to increase the accuracy of the position estimation. The flux linkage is estimated by the voltage equation of the PSRM, and the estimated flux linkage is then used to estimate the phase current in the adaptive current observer. To calculate the observing error of the SMO using the measured and estimated phase currents, the observing error of the thrust force is introduced to replace the immeasurable state error of the position and speed of the mover. The sliding surface is designed based on the error of the thrust force, and stability analysis is given. Once the sliding surface is reached, the mover position is then estimated accurately. Finally, the effectiveness of the proposed method for the PSRM is verified experimentally.