Detecting cracks from optical images plays a crucial role in road maintenance but its good realisation has many challenges. Road cracks exhibit significant diversity and complexity in terms of shape, size and texture and road images may contain various types of noise and interference, such as lighting variations, shadows and different appearances, due to varying perspectives and scales. To address these challenges, a comprehensive road crack dataset called CRCrack has been constructed, which encompasses various crack characteristics. This study proposes a road crack segmentation network called CSegNet, which combines convolutional neural networks (CNNs) and transformers. The network adopts an encoder-decoder framework, namely DeepLabV3+. In the encoder, leveraging the transformers' flexibility in modelling long-term dependencies and the CNNs' ability to capture local contextual information through local receptive fields, weight sharing and spatial subsampling, a ResNeXt-Transformer (ResNeXTR) feature extraction module is designed as the backbone network to enhance the feature extraction capability for road crack images. To reduce the computational cost in the self-attention (SA) computation of the transformer, an average pooling layer is introduced to downsample the dimensions of the encoded features. In the decoder, to focus on the key information of road cracks under diverse environmental conditions and interferences, an efficient channel attention module (ECAM) and a spatial attention module (SAM) are combined to design an efficient convolutional block attention module (ECBAM) to optimise feature representation. Through comparative experiments on the CRCrack dataset, the results demonstrate that the proposed method outperforms classic networks such as U-Net and DeepLabV3+ in terms of intersection over union (IoU), Dice coefficient and area under the receiver operating characteristic (AUROC) curve evaluation metrics. It exhibits good adaptability to ground crack images from different sources, providing a basis for estimating the degree of road damage.

Detecting cracks from images plays a crucial role in road maintenance. Road cracks exhibit significant diversity and complexity in terms of shape, size, texture, and road images may contain various noises and interferences such as lighting variations, shadows, and different appearances due to varying perspectives and scales. To address these challenges, we constructed a comprehensive dataset called the Comprehensive Road Crack Dataset (CRCrack Dataset), which encompasses various crack characteristics. In this study, we propose a road crack segmentation network called CSegNet (Crack Segmentation Network), which combines convolutional neural networks (CNNs) and Transformers. The network adopts an encoder-decoder framework, like DeepLab V3+. In the encoder, leveraging the flexibility of Transformers in modeling long-term dependencies and the ability of CNNs to capture local contextual information through local receptive fields, weight sharing, and spatial subsampling, we design a ResNeXTR (ResNeXt-Transformer) feature extraction module as the backbone network to enhance the feature extraction capability for road crack images. To reduce the computational cost in self-attention computation of transformer, we introduce an average pooling layer to downsample the dimensions of the encoded features. In the decoder, to focus on the key information of road cracks under diverse environmental conditions and interferences, we combine the Efficient Channel Attention Module (ECAM) and the Spatial Attention Module (SAM) to design an Efficient Convolutional Block Attention Module (ECBAM) attention module to further optimize feature representation. Additionally, we employ the ReLU activation function, SGD gradient descent, and a hybrid loss function of Binary Cross Entropy with Logits to accelerate convergence speed and improve segmentation accuracy. Through comparative experiments on the CRCrack dataset, the results demonstrate that our proposed method outperforms classic networks such as U-Net and DeepLab V3 + in terms of IoU, Dice, and AUROC evaluation metrics. It exhibits good adaptability to ground crack images from different sources, providing a basis for estimating the degree of road damage.

Accurate detection and timely treatment of component defects in substations is an important measure to ensure the safe operation of power systems. In this study, taking substation meters as an example, a dataset of common meter defects, such as a fuzzy or damaged dial on the meter and broken meter housing, is constructed from the images of manual inspection in power systems. There are several challenges involved in accurately detecting defects in substation meter images, such as the complex background, different meter sizes and large differences in the shapes of meter defects. Therefore, this paper proposes the PHAM-YOLO (Parallel Hybrid Attention Mechanism You Only Look Once) network for automatic detection of substation meter defects. In order to make the network pay attention to the key areas against the complex background of the meter defect images and the differences between different defect features, a Parallel Hybrid Attention Mechanism (PHAM) module is designed and added to the backbone of YOLOv5. PHAM integration of local and non-local correlation information can highlight these differences while remaining focused on the meter defect features. To improve the expressive ability of the feature map, a Spatial Pyramid Pooling Fast (SPPF) module is introduced, which pools the input feature map using a continuous fixed convolution kernel, fusing the feature maps of different receptive fields. Bounding box regression (BBR) is the key way to determine object positioning performance in defect detection. EIOU (Efficient Intersection over Union) is, therefore, introduced as a boundary loss function to solve the ambiguity of the CIOU (Complete Intersection Over Union) loss function, making the BBR regression more accurate. The experimental results show that the Average Precision Mean (mAP), Precision (P) and Recall (R) of the proposed PHAM-YOLO network in the dataset are 78.3%, 78.3%, and 79.9%, respectively, with mAP being improved by 2.7% compared to the original model and higher than SSD, Fast R-CNN, etc.

The flow field of the environment plays a crucial role in cigarette combustion cone fallout propensity test, with air velocity exhibiting a positive correlation with combustion volume. In order to minimize the impact of the environmental flow field on the test results, it is necessary to control the air speed within the range of 200 ± 30 mm/s in the test area of each tobacco test channel. To address this concern, which used the Realizable model to develop a mathematical model of the testing environment. The uniformity of air speed in each channel and its relationship with structural parameters were then analyzed. Based on these findings, the key structural parameters of the ventilation hood are optimized. After restimulated the optimized model, the results demonstrate a higher level of uniformity in the environmental flow field of the optimized section. To validate the accuracy of the simulation results, measurements indicated that the maximum air speed value at all points is 225.6 mm/s, while the minimum value is 178.44 mm/s. These values fall within the specified range of 200 ± 30 mm/s, thus meeting the design requirements. This study ensures that the cigarette can burn in a steady state during the cigarette combustion fallout propensity test and improves the stability of the cigarette combustion cone drop tendency test results.

In order to further improve accuracy and stability of detection of combustion coal fallout propensity of cigarettes, author of the paper adopted computational fluid dynamics (CFD) technology for a three-dimensional numerical simulation of exhaust system of detection instrument, aiming to study characteristics of flow field near cigarettes. Moreover, a simulation model of eight-channel exhaust enclosure was established, obtaining vector diagram for flow velocity of flow field, velocity contour diagram, and pressure distribution cloud diagram. According to findings, flow field of eight channels is evenly distributed, with slow flow velocity around the instrument but furious inside channels. The wind velocity of cigarette monitoring channel is stable at about 200mm/s specified as per standard. However, there is significant change in pressure and flow velocity at the corners of channels, causing local turbulence. In experiments, average wind velocity of 8 monitoring channels was measured, and simulation results were compared with experiment data. Eventually, a conclusion is drawn that simulation result at cigarette monitoring channels changes consistently with the experimental data, with small errors as a whole. Therefore, the designed exhaust system complies with regulations on wind velocity stipulated by YC/T558-2018 Cigarettes—Determination of Combustion Coal Fallout Propensity of Burning Cigarettes. In a word, this paper is hoped to provide technical support for analogue simulation of exhaust system of cigarette detection instrument, and improve detection accuracy.

In order to guide the production of cigarette products and improve the quality of cigarette products, this paper proposes a classification method for cigarette combustion cones based on deep convolutional neural network model. The method is optimized based on the Inception Resnet V2 model and is innovatively used in the detection of cigarette burning cones. The classification accuracy of combustion cone fallout is characterized by the overall classification accuracy (OA) and the Kappa coefficient (Kappa). The experimental results show that the overall classification accuracy is 97.22%, and the Kappa coefficient is 0.9583. The deep convolutional neural network has better classification effect. Based on the classification method of deep convolutional neural network, the cigarette burning cone can be accurately identified.

Summary The width of cut tobacco strands is an important indicator for physical parameters as well as for the smoking quality. In some countries, cut width helps to distinguish fine-cut tobacco and pipe tobacco and thus differentiates taxation rate. A new method for rapid measurement of the width of cut tobacco strands was developed based on digital image processing, because the method described in ISO 20193, though easy to implement in factories, proved time consuming and generated high testing costs. The essence of this method is to determine the statistic width of incisions. The straight-line segments represent the width of strands of cut tobacco, from which the determination of the width for randomly placed tobacco strands could be achieved. Five kinds of samples (‘ISO collaborative study samples 0.4 mm, 1.0 mm, 1.6 mm and 3.0 mm’ and ‘Guangdong baked 0.9 mm’) were used to study the comparability of the measurement results between the method presented in this work and the current ISO method. Results show that accuracy and repeatability are comparable. In addition, the testing efficiency of the method presented in this work appears to be higher than the current ISO method, and it is thus a promising alternative method for measuring the width of strands of cut tobacco.

In this paper, a classification method of tobacco leaves and impurities is presented, which is based on hyperspectral imaging technology. By combining savitzky-golay smoothing filter, multiplicative scatter correction and random forest classifier, the problem of the classification of tobacco leaves and impurities is overcome. The classification accuracy of the sample was characterized by the overall classification accuracy (OA) and Kappa coefficient, and the OA is 99% and the Kappa coefficient is equal to 0.987. The experimental results indicate that the random forest is an excellent hyperspectral data classifier, and the classification method of tobacco leaves and impurities based on hyperspectral imaging technology can accurately distinguish tobacco leaves and impurities which are in tobacco leaves.

Aiming at the problem of efficiency and autonomy of unknown region exploring,an unknown region exploration method based on boundary search idea and suitable for two-dimensional grid map was proposed.The exit area on the boundary in the built environment map was utilized as the logical node and the function value of the node distance was used as the edge weight,to construct the exploration tree of the node to be explored dynamically through traversing,searching,establishing new node and using simultaneous positioning and mapping technology.The robot would be guided to the next exploration area,to complete the independent exploration of the unknown area.The theoretical analysis and simulation results show that with the same experimental map,moving speed of robot and the maximum effective range of the sensor data,compared with other commonly used unknown region exploration algorithm for robots,the single exploration coverage of the proposed method is much higher,so is its exploration efficiency.Logically,the proposed method is a depth-first search traversal of the minimum spanning tree corresponding to the raster map,where each exploration chooses a clear target area,and there is not multiple coverage and falling into the loop,so it is more efficient and stable.

This paper proposes an experiment method of analyzing the relationship between film wear resistance and material component. The method characterizes film wear resistance from the film image with scratches using the characteristic parameters of gray level co-occurrence matrix(GLCM), and establishes the relationship between film wear resistance and material component. Firstly, the Retinex image enhancement method is used to get rid of the mist; and the image enhancement effect of the proposed method is compared with those of the grayscale transform method and homeostasis filtering method. Secondly, GLCM is applied to analyze the texture information of the film scratches and calculate the characteristic parameters. Thirdly, the film wear resistance characterized by multi-characteristic parameters is taken as the explanatory variable and the material component of the film is taken as the explained variable, and a multivariate regression model is built. At last, the regression model is analyzed to obtain the quantitative description of the relationship between the explanatory variable and the explained variable. In the experiment, a multivariate regression model was built by taking the ratio of ABS04S elastomer in binary copolymer KS359 biaxially oriented polypropylene(BOPP) film and the film wear resistance as the explained variable and the explanatory variable, respectively. Consequently, we obtained that the multiple correlation coefficient of the film wear resistance and the explained variables is 0.992, which implies that the method is feasible.

This paper introduces a kind of intelligent cigarette length measurement apparatus based on digital image processing techniques, and presents the structural method of measurement system as well as the related algorithm of image processing of cigarette ends. In the end the method’s feasibility and superiority is proved through comparing experiment data and analyzing the quality of this method and the traditional methods.

An interferometrical method for Fourier Transform spectral analysis is presented. The system is based on single mode optical fiber Mach-zehnder interferometer. DFB Laser and ASE have been measured, the spectral resolution is up to 0.78 cm-1.

Considering the cigarette paper image with different gray-scale between embossed rib area and non-embossed areas due to the different reflectivity of light, using VFW (Video for Window) technology, feature extraction and other related computer image processing methods, a intensity of press marks on cigarette paper detection system based on ordinary camera and under the Windows Platform was designed. This system realizes the software-only control of the auto detection of the intensity of press marks on cigarette paper. The experiment result indicated the system can provide satisfactory results.

In this weigh-in-motion(WIM)research,we introduce a novel design of WIM system based-on fiber Bragg grating(FBG) sensing technology.When a car run over the elastic body,it will cause the deflection variation of the elastic body.By use of FBG sensors,we can measure the deflection of the elastic body to calculate the car's weight.The method of real-time calibrating primary reference wavelength of FBG was applied in this system to eliminate the influence of the ambient temperature on the measurement result of variation.The experimental results showed that the accuracy of the system is better than ±3%.