Shi-Wei Lo

Cyber surveillance for flood disasters.

Regional heavy rainfall is usually caused by the influence of extreme weather conditions. Instant heavy rainfall often results in the flooding of rivers and the neighboring low-lying areas, which is responsible for a large number of casualties and considerable property loss. The existing precipitation forecast systems mostly focus on the analysis and forecast of large-scale areas but do not provide precise instant automatic monitoring and alert feedback for individual river areas and sections. Therefore, in this paper, we propose an easy method to automatically monitor the flood object of a specific area, based on the currently widely used remote cyber surveillance systems and image processing methods, in order to obtain instant flooding and waterlogging event feedback. The intrusion detection mode of these surveillance systems is used in this study, wherein a flood is considered a possible invasion object. Through the detection and verification of flood objects, automatic flood risk-level monitoring of specific individual river segments, as well as the automatic urban inundation detection, has become possible. The proposed method can better meet the practical needs of disaster prevention than the method of large-area forecasting. It also has several other advantages, such as flexibility in location selection, no requirement of a standard water-level ruler, and a relatively large field of view, when compared with the traditional water-level measurements using video screens. The results can offer prompt reference for appropriate disaster warning actions in small areas, making them more accurate and effective.

Visual Sensing for Urban Flood Monitoring

With the increasing climatic extremes, the frequency and severity of urban flood events have intensified worldwide. In this study, image-based automated monitoring of flood formation and analyses of water level fluctuation were proposed as value-added intelligent sensing applications to turn a passive monitoring camera into a visual sensor. Combined with the proposed visual sensing method, traditional hydrological monitoring cameras have the ability to sense and analyze the local situation of flood events. This can solve the current problem that image-based flood monitoring heavily relies on continuous manned monitoring. Conventional sensing networks can only offer one-dimensional physical parameters measured by gauge sensors, whereas visual sensors can acquire dynamic image information of monitored sites and provide disaster prevention agencies with actual field information for decision-making to relieve flood hazards. The visual sensing method established in this study provides spatiotemporal information that can be used for automated remote analysis for monitoring urban floods. This paper focuses on the determination of flood formation based on image-processing techniques. The experimental results suggest that the visual sensing approach may be a reliable way for determining the water fluctuation and measuring its elevation and flood intrusion with respect to real-world coordinates. The performance of the proposed method has been confirmed; it has the capability to monitor and analyze the flood status, and therefore, it can serve as an active flood warning system.

Fluvial monitoring and flood response

This paper proposes a framework for an Image-based Flood Alarm (IFA) that includes bi-seeded region-based image segmentation for the extraction of a water region of interest from an image, as well as an alarm classifier for identifying the degree of flood risk. When the risk reaches the predetermined threshold, a flood response message reports to the main EWS for end-user decision support.

Flood Tracking in Severe Weather

Severe weather conditions greatly impair the performance of outdoor imaging. In this study, two region-based image segmentation methods, Grow Cut and Region Growing (RegGro), were applied to rain scenes. This study demonstrates that segmentation accuracy depends on fog and rain stains. In severe rainfall periods, heavy rain and fog reduced the overall image quality, and both methods yielded segmentation failure. The results show that both region-based methods are effective for segmenting objects in images captured under poor weather conditions. Both methods have unique advantages and disadvantages for fog and stain conditions. The segmentation accuracy yielded by the Grow Cut and RegGrow methods was 75% and 85%, respectively.

Video Query Using Temporal Signature and Similarity Matching

Abstract. Large amount of video data is stored and distributed in wide variety of application. Due to the fast video material increases, manage and query of video become more and more important. In this paper, we address a temporal signature representation and similarity model to retrieval the similar video within database by video query. Experimental results on real date are presented. The experimental results show that the statistical approach permits accurate query of video clip, in particular, the performance of the approach was found extremely satisfactory with determine all similar video in database.

Uncertainty Comparison of Visual Sensing in Adverse Weather Conditions.

This paper focuses on flood-region detection using monitoring images. However, adverse weather affects the outcome of image segmentation methods. In this paper, we present an experimental comparison of an outdoor visual sensing system using region-growing methods with two different growing rules—namely, GrowCut and RegGro. For each growing rule, several tests on adverse weather and lens-stained scenes were performed, taking into account and analyzing different weather conditions with the outdoor visual sensing system. The influence of several weather conditions was analyzed, highlighting their effect on the outdoor visual sensing system with different growing rules. Furthermore, experimental errors and uncertainties obtained with the growing rules were compared. The segmentation accuracy of flood regions yielded by the GrowCut, RegGro, and hybrid methods was 75%, 85%, and 87.7%, respectively.

SSIM for video representing and matching

This paper propose a compact video representation using SSIM assessment for matching video sequences. As opposed to using the complex features extracting methods to analysis color and motion information through the sequence. We use the similarity index of successive frames to produce a representation of videos. We employ the SSIM assessment to represent a video for matching the same video within database. The experimental results show that the proposed approach permits accurate for video matching. The performance was found extremely satisfactory with determine all correct video in database.

Video Matching by One-Dimensional PSNR Profile

This paper addresses a compact framework to matching video sequences through a PSNR-based profile. This simplify video profile is suitable to matching process when apply in disordered undersea videos. As opposed to using color and motion feature across the video sequence, we use the image quality of successive frames to be a feature of videos. We employ the PSNR quality feature to be a video profile rather than the complex contend-based analysis. The experimental results show that the proposed approach permits accurate of matching video. The performance is satisfactory on determine correct video from undersea dataset.