Research On Image Measurement Technology Of Small Displacement Based On Optical Flow Field Algorithm

The occurrence of fire is more common in daily life.The collapse of the building caused by fire is the major factors that threaten the safety of the rescuers.Buildings often produce enough deformation before they collapse,which is caused by high temperatures.At present,there are no technologies that are specifically designed to monitor the collapse of buildings in the event of a fire,and related technologies(such as techniques for monitoring the displacement of dams and mountains)are not easily portable and are unstable to the specific environments(rescue environments).The above reasons make it difficult to meet the special needs of building displacement monitoring during a fire.Therefore,a micro-displacement telemetry technology based on the optical flow field is proposed in this paper.It aims at real-time monitoring the building expansion,which is caused by high temperatures,and gives an early warning signs of building collapse.The research work made in this thesis mainly includes the following points:1)Analyze social needs and existing related monitoring technologies to complete the layout of the functions and the overall design concept of the monitoring system.2)Combining the requirements of the application scenario,the disadvantages of the dense optical flow field,the sparse optical flow field,and several common corner points are compared to determine the algorithm that should be used in this experiment.Then some improvements in the sparse optical flow field are introduced in this thesis.And finally the feasibility of optical flow method for monitoring building displacement caused by the thermal expansion has been analyzed in this thesis.3)In view of the problems of interference monitoring and calculation that may appear in practical application,an inquiry-type experimental study is conducted.This thesis mainly deals with the problems of image preprocessing and uneven distribution of corner points.The solutions to the problem of distinguishing the displacements of interferences from the monitored objects and the decline of calculation accuracy caused by frequent "rounding" are explored and analyzed.The experimental basis is provided for the algorithm design of subsequent experiments by the above results.4)A feasible solution has been designed after synthesizing the results of inquiring experiments and theoretical analysis.A series of methods such as use optical flow tomatch corner points to improve the real-time performance of the system and using the characteristics of different motion speeds to eliminate interference,calculate sub-pixel corner point to avoid excessive "rounding" error and so on is proposed,which makes the calculation of the algorithm stable and accurate.5)According to laboratory conditions,experimental system design and construction.Simulate some interferences that may occur under real conditions to verify the feasibility and accuracy of the solution.The experimental data were collated and analyzed,and the results showed that the solution can monitor the small displacement of the building stably,and not only is it unaffected by the above mentioned interference conditions,but it has also achieved a good monitoring effect.