Liquid crystal display (LCD) defects detection on module level is increasingly important for flat-panel displays (FPD) industry to increase the production capacity via machine vision technology. However, it is an overwhelmingly challenging issue due to various difficulties. This article discloses a practical automatic optical inspection (AOI) system consisting of hardware structure and software algorithm to detect module-level defects. The AOI system is the core component to build a distributed integrated inspection system with the help of the Internet of Things (IoT). Starting from the analysis of the challenges encountered in module-level defects inspection, a delicate photograph scheme is proposed to reveal different kinds of defects. In order to robustly work on the module-level defects detection with complex situations, a novel framework based on YOLOV3 detection unit is proposed in this article, including the preprocessing module, detection module, defects definition module, and interferences elimination module. To the best of our knowledge, this is the first work that designs a practical AOI system for module-level defects detection. In order to demonstrate the effectiveness of the proposed method, extensive experiments have been conducted on the manufacturing lines. The evaluation of the detection performance of the AOI system in comparison with a manual scheme indicates that the proposed system is practical for module-level defects detection. Currently, the proposed system has been deployed in a real-world LCD manufacturing line from a major player in the world.

Field vehicle type recognition plays an essential role in border protection tasks. Acoustic and seismic sensors can effectively collect the signal of field vehicle targets in real time. Most vehicle temporal signal classification algorithms are based on extracting and identifying handcrafted features. These algorithms focus on the signal’s frequency domain characteristics and despise the signal’s time domain characteristics. In order to extract appropriate features, this paper proposes a Long-Term Correlation Feature Network (LTCFN) to perform field vehicle acoustic and seismic signal classification. The model includes AlexNet-type feature extractor and an overall classifier implemented by a Long-Short Term Memory(LSTM) network. We presents an intra-frame network and fusion method for extracting feature vector from signals. Meanwhile, an inter-frame classifier is proposed firstly for analysing the time-correlation of the feature map and overall classification. The experiments illustrate that the LTCFN has excellent recognition performance and anti-noise ability. The classification accuracy of the LTCFN can be increased to 96%. This paper also provides a new idea for ground target classification through inter-frame feature measurement.

Noncontact heart rate monitoring that enables the detection of cardiac activity cycles with a normal camera without cutaneous contact has broad application prospects. However, the state-of-theart approaches lack robustness to the lower illuminance and the varying illuminance. Objective: This paper builds an optical noncontact heart rate monitoring framework (lower/vary illumiance PPG, LiPPG) with the aim of improving the robustness of these illuminance situations. Approach: In LiPPG, we use the YIQ color space to represent the PPG signal. By applying the Savitzky-Golay filter, we propose a mathematical model, named multiscale variable-weight Savitzky Golay combination (MVSGC), to extract signals in the time domain and reduce the disturbance from noise. Main results: We test the LiPPG on three scenarios: stationary, facial motion and varying illuminance, with eighteen video clips from six participants. Significance: The experimental results show that the noncontact heart rate estimation using the LiPPG reduced the errors compared with the gold-standard method of recording from a finger using a pulse oximeter and that it is superior to existing methods.

This paper proposes a lightweight model combined with data augmentation for vehicle detection in an intelligent sensor system. Vehicle detection can be considered as a binary classification problem, vehicle or non-vehicle. Deep neural networks have shown high accuracy in audio classification, and convolution neural networks are widely used for audio feature extraction and audio classification. However, the performance of deep neural networks is highly dependent on the availability of large quantities of training data. Recordings such as tracked vehicles are limited, and data augmentation techniques can be applied to improve the overall detection accuracy. In our case, spectrogram augmentation is applied on the mel spectrogram before extracting the Mel-scale Frequency Cepstral Coefficients (MFCC) features to improve the robustness of the system. Then depthwise separable convolution is applied to the CNN network for model compression and migrated to the hardware platform of the intelligent sensor system. The proposed approach is evaluated on a dataset recorded in the field using intelligent sensor systems with microphones. The final frame-level accuracy achieved was 94.64% for the test recordings and 34% of the parameters were reduced after compression.

In this paper, we introduce a sub-band direction-of-arrival (DOA) estimation method suitable for employment within an automatic bearing tracking system. Inspired by the magnitude-squared coherence (MSC), we extend the MSC to the sub-band and propose the sub-band magnitude-squared coherence (SMSC) to measure the coherence between the frequency sub-bands of wideband signals. Then, we design a sub-band DOA estimation method which chooses a sub-band from the wideband signals by SMSC for the bearing tracking system. The simulations demonstrate that the sub-band method has a good tradeoff between the wideband methods and narrowband methods in terms of the estimation accuracy, spatial resolution, and computational cost. The proposed method was also tested in the field environment with the bearing tracking system, which also showed a good performance.

Acoustic sensors are used as an important sensor component in wireless sensor networks (WSNs) for wild environmental monitoring, because of its small size, light weight and low power consumption. To further utilize the superiorities of acoustic sensor, this paper aims to design a single-mode wild area surveillance sensor based on acoustic sensors, which has the advantages of practicality, low hardware power consumption and low software complexity. The proposed algorithm can effectively suppress wind noise and improve the performance of target detection, classification and direction of arrival (DOA) estimation. The power consumption test results show that the designed micro-sensor node has low power consumption of about 13.8mW in on-duty mode and long-term continuous monitoring capability of about 33 days. The field experiment results reveal that the node has favorable performance in detection (detection rate as high as 96% with false alarm rate under 5%), tracking (error is less than 7.6 degree) and target classification (accuracy is greater than 92.6%).