| Changes in tidal flat topography are one of the key issues related to the sustainability of coastal ecosystems and coastal protection.Its short-term changes will directly affect the expansion and reproduction of salt marsh vegetation in the bare tidal flats,and then affect the protection effect and recovery ability of salt marsh wetlands.Previous researches mainly focus on the changes of tidal flats on large and medium time and space scales(from several years to several decades in large and medium regions).However,the tidal flats change rapidly and are easily disturbed by external factors.Thus,more advanced monitoring methods are needed to finely detect micro-scale(The scales are agreed to be spatial scales in this work,excluding time scales.Micro-scales are agreed as the spatial range of a few centimeters to several hundred meters)characteristics of tidal flats and their changing processes in different time scales(hourly,dayly,monthly and yearly),which can provide theoretical and technical support for coastal protection and management.Recently,researchers have begun to pay attention on the changes at small and micro spatiotemporal scales.Various corresponding monitoring methods and equipments have been gradually applied to the monitoring of tidal flat.But these works are still in the preliminary exploratory stage.Unlike traditional remote sensing methods deployed on satellites and aviation platforms,ground-based remote sensing is closer to the surface of coastal zone,whick makes the information they obtained more detail,abundant and diverse.As two excellent ground-based remote sensing methods,terrestrial LIDAR(TLS: Terrestrial Laser Scanner)and Coastal Video System have been increasingly used in fine coastal detection.However,these previous studies mainly focus on the monitoring of sandy coastal zones,and there are few works contribute to monitoring of muddy flats(tidal flats).Barely no study use Video System on monitoring muddy flats.TLS has the advantages of wide coverage,high precision and high spatial resolution,with one single scan takes from ten minutes to dozens of minutes.The Video System can realize long-term monitoring of fixed areas at small and micro scales.The temporal resolution of the acquired video data can reach more than one frame per second.It has many advantages such as low cost,unattended operation and ignoring weather conditions.These two ways have great complementarity in space and time scale.Thus,studying the key technology of fine detecting tidal flats by using TLS and Video System is the premise to further study the micro-scale change of tidal flats.This paper intends to focus on key issues such as the change process of tidal flats based on the micro-scale features extracted from the video data and point cloud data.Following are four main achievements from this work:1.Considering the characteristics of muddy flats,algorithms and methods for extracting micro-features from video images and point cloud data are proposed:Algorithms and methods for extracting micro-features of tidal flats from video images are proposed,including micro traces of tidal flats,micro topographic units and waterlines.Based on the video data from video system in this work,the above features are extracted and evaluated by confusion matrix.The results show that the extraction accuracy of micro traces and micro topographic units has reached more than 95%,while the average deviation of the extracted waterline is only 2 pixels.The algorithms perform well and provide reliable features for further analysis.Meanwhile,algorithms for extracting micro-scale features from point cloud data is proposed,including vegetation extraction and micro traces extraction.Then the effect of these algorithms is analyzed.The results show that the accuracy of vegetation extraction results in both the horizontal and vertical directions exceeds97%.However,this algorithm requires the tidal flat scene to have both bare tidal flats vegetation.For the area completely covered by vegetation is not within the consideration in this work.Meanwhile,the extraction accuracy of micro traces is also above 97%.These algorithms are reliable enough and can provide support for further change analysis.2.The quantitative analysis methods of micro-scale features are proposed,and the fusion experiments between video image and TLS point cloud are carried out and discussed:Firstly,the specific quantification methods of micro-scale features after geographic registration are proposed.Then the possible fusion applications are further explored,including improving the registration quality of point cloud data,using point cloud to perform geo-registration of video images based on curve matching,and 3D reconstruction of the waterlines using multi-view geometry.The results show that the point cloud registration quality has been significantly improved,its average error has been reduced by about 5 times,and all site errors have dropped below 4 cm.Secondly,the error of the point pairs extracted from the curve matching is less than the pixel resolution within 8cm.Finally,the horizontal and vertical errors of the 3D reconstructed waterlines are large;with the vertical accuracy within 110 meters from the video equipment can only analyze the seasonal topagraphic changes.In the further research,higher pixel resolution can be considered to reduce errors of 3D reconstructed waterlines.3.Short-term(minutes to daily)changes and of the micro-scale features are analysisd by using video images:Based on the quantification and change results of the micro-scale features extracted from the video images,the short-term change pattern is studied,and is found on both two short-term timescales ranging from minutes to daily.First,from the changes in the sediment transport process on the tidal flat surface under the influence of strong wind,it is found that the sediments are transported in a blocky manner in the micro traces.And the edges of micro traces change significantly in the direction of the wind.At the same time,it is found that the change process presents the characteristics of nonlinear change,and its boundary is constantly oscillating in the direction of the wind direction.Further monitoring of one ROI shows that the most drastic changes are concentrated in the convex region in the wind direction.The possible reason is that during the sediment transport process,part of the sediment concentrated in the convex area and continuously filled the bottom of micro traces.Finally,after the underlying area is filled to a certain extent,it is quickly exposed and becomes part of the micro-scale units.By studying the daily change process of micro-scale units,it is found that there is a smooth trend on the surface of tidal flats.Its change process is always in the periodic process from breaking into more small units,to smoothing and breaking again.4.Monthly and interannual variation of the micro-scale features are carried out and analyzed by using TLS:First,the inter-promoting relationship between vegetation and depositional processes is confirmed from the changes of pioneer vegetation at micro scale.The deposition rate in the monitored area has increased by 2 to 4 times compared to the rate before 2016.The areas with higher rate are mainly located in the vegetation area.A highly correlated spatial relationship between tussocks and seedlings is also found.Seedlings are densely concentrated within a certain distance(20 m in this work)from tussocks.And with the expansion of tussocks,the distance between them will be compressed.This confirms that previous studies have observed the seedlings decrease with the distance from tussocks,which leads to the tendency of seedling to be close to tussocks.From the analysis of the interannual state of the micro traces,it is found that the left and right sides of the intensity curve with the peak position as the dividing point are monotonic.The peak position essentially represents a critical value,and the change of the position represents the spatial variation of the two factors: the softness of the tidal flat surface and the hydrodynamic force in the cross-shore direction.In the correlation analysis,it is further found that the strength is very strongly correlated with the distance from the seawall,which proves that micro traces related to the tidal dynamics. |
PDF Full Text Request