Real-Time Dynamic Intelligent Image Recognition and Tracking System for Rockfall Disasters.

Taiwan, frequently affected by extreme weather causing phenomena such as earthquakes and typhoons, faces a high incidence of rockfall disasters due to its largely mountainous terrain. These disasters have led to numerous casualties, government compensation cases, and significant transportation safety impacts. According to the National Science and Technology Center for Disaster Reduction records from 2010 to 2022, 421 out of 866 soil and rock disasters occurred in eastern Taiwan, causing traffic disruptions due to rockfalls. Since traditional sensors of disaster detectors only record changes after a rockfall, there is no system in place to detect rockfalls as they occur. To combat this, a rockfall detection and tracking system using deep learning and image processing technology was developed. This system includes a real-time image tracking and recognition system that integrates YOLO and image processing technology. It was trained on a self-collected dataset of 2490 high-resolution RGB images. The system's performance was evaluated on 30 videos featuring various rockfall scenarios. It achieved a mean Average Precision (mAP50) of 0.845 and mAP50-95 of 0.41, with a processing time of 125 ms. Tested on advanced hardware, the system proves effective in quickly tracking and identifying hazardous rockfalls, offering a significant advancement in disaster management and prevention.

Research on nano H2/O2 bubble generating mechanism and characteristics.

In this research, electrolysis water is used to produce hydrogen and oxygen for carrying out the vertical cutting through high-speed water in order that the bubbles will be refined for generating the nano H2/O2 bubble liquid. In the meantime, a Nanobubble Generator is developed to verify the basic characteristics of the produced nano H2/O2 bubbles. Its purpose is to identify the maximum concentration of bubbles in the nano H2/O2 bubble liquid, the bubble production efficiency and bubble electrification characteristics as well as the effect of reducing the pipe flow friction resistance together with the characteristics of nanobubbles containing varied gases. By verifying the nano H2/O2 bubbles, it is hoped that the flowing rate of the hollow electrode can be elevated.

Deep Convolutional Neural Network Optimization for Defect Detection in Fabric Inspection.

This research is aimed to detect defects on the surface of the fabric and deep learning model optimization. Since defect detection cannot effectively solve the fabric with complex background by image processing, this research uses deep learning to identify defects. However, the current network architecture mainly focuses on natural images rather than the defect detection. As a result, the network architecture used for defect detection has more redundant neurons, which reduces the inference speed. In order to solve the above problems, we propose network pruning with the Bayesian optimization algorithm to automatically tune the network pruning parameters, and then retrain the network after pruning. The training and detection process uses the above-mentioned pruning network to predict the defect feature map, and then uses the image processing flow proposed in this research for the final judgment during fabric defect detection. The proposed method is verified in the two self-made datasets and the two public datasets. In the part of the proposed network optimization results, the Intersection over Union (IoU) of four datasets are dropped by 1.26%, 1.13%, 1.21%, and 2.15% compared to the original network model, but the inference time is reduced to 20.84%, 40.52%, 23.02%, and 23.33% of the original network model using Geforce 2080 Ti. Furthermore, the inference time is also reduced to 17.56%, 37.03%, 19.67%, and 22.26% using the embedded system AGX Xavier. After the image processing part, the accuracy of the four datasets can reach 92.75%, 94.87%, 95.6%, and 81.82%, respectively. In this research, Yolov4 is also trained with fabric defects, and the results showed this model are not conducive to detecting long and narrow fabric defects.

Study on the Measurement of Laser Drilling Depth by Combining Digital Image Relationship Measurement in Aluminum.

In this study, laser processing equipment was used to drill aluminum alloy materials and with different auxiliary mechanisms, the deformation around the holes after processing was observed. The experimental results show that, due to the high temperature generated during laser processing, a large thermal gradient causes thermal stress to be introduced into the test piece and outward expansion deformation occurs. In this study, the digital image correlation and residual stress detection methods were applied. Based on the correlation between the drilled hole depth and the hole deformation, the hole depth of the laser processing was estimated. The average coefficient of determination for all auxiliary mechanisms is 0.82. The experimental results confirm that the digital image correlation method can be used to estimate the hole depth of laser processing.

Micro-Drilling of Sapphire using Electro Chemical Discharge Machining.

Electrochemical discharge machining (ECDM) refers to a non-traditional machining method for performing effective material removal on non-conductive hard and brittle materials. To increase the ECDM machining efficiency, traditionally, the method of increasing the machining voltage or increasing the electrolyte concentration is used. These methods can also cause overcut reaming of the drilled holes and a rough surface on the heat affected area. In this study, an innovative combinational machining assisted method was proposed and a self-developed coaxial-jet nozzle was used in order to combine two assisted machining methods, tool electrode rotation and coaxial-jet, simultaneously. Accordingly, the electrolyte of the machining area was maintained at the low liquid level and the electrolyte was renewed at the same time, thereby allowing the spark discharge to be concentrated at the contact surface between the front end of the tool electrode and the machined material. In addition, prior to the machining and micro-drilling, the output of the machining energy assisted mechanism was further controlled and reduced. For the study disclosed in this paper, experiments were conducted to use different voltage parameters to machine sapphire specimens of a 640 µm thickness in KOH electrolyte at a concentration of 5 M.

Characteristics of the Arcing Plasma Formation Effect in Spark-Assisted Chemical Engraving of Glass, Based on Machine Vision.

Spark-assisted chemical engraving (SACE) is a non-traditional machining technology that is used to machine electrically non-conducting materials including glass, ceramics, and quartz. The processing accuracy, machining efficiency, and reproducibility are the key factors in the SACE process. In the present study, a machine vision method is applied to monitor and estimate the status of a SACE-drilled hole in quartz glass. During the machining of quartz glass, the spring-fed tool electrode was pre-pressured on the quartz glass surface to feed the electrode that was in contact with the machining surface of the quartz glass. In situ image acquisition and analysis of the SACE drilling processes were used to analyze the captured image of the state of the spark discharge at the tip and sidewall of the electrode. The results indicated an association between the accumulative size of the SACE-induced spark area and deepness of the hole. The results indicated that the evaluated depths of the SACE-machined holes were a proportional function of the accumulative spark size with a high degree of correlation. The study proposes an innovative computer vision-based method to estimate the deepness and status of SACE-drilled holes in real time.

On-line estimation of laser-drilled hole depth using a machine vision method.

The paper presents a novel method for monitoring and estimating the depth of a laser-drilled hole using machine vision. Through on-line image acquisition and analysis in laser machining processes, we could simultaneously obtain correlations between the machining processes and analyzed images. Based on the machine vision method, the depths of laser-machined holes could be estimated in real time. Therefore, a low cost on-line inspection system is developed to increase productivity. All of the processing work was performed in air under standard atmospheric conditions and gas assist was used. A correlation between the cumulative size of the laser-induced plasma region and the depth of the hole is presented. The result indicates that the estimated depths of the laser-drilled holes were a linear function of the cumulative plasma size, with a high degree of confidence. This research provides a novel machine vision-based method for estimating the depths of laser-drilled holes in real time.