Research On Visible Light Indoor Positioning Method Based On Image Sensor

As the development of urbanization,with the growth on the volume and area of buildings and the complexity of the structure,indoor positioning has gradually become people's daily needs.Such as large shopping malls reverse car searching,searching for a specific product,locating separated family members and so on,which all require the support of indoor wireless positioning technology.Indoor positioning systems represented by technologies such as WIFI and Bluetooth have large positioning errors due to long distances,weak signals,and other reasons,and the positioning accuracy is less than 5meters.Moreover,the floor recognition of the indoor space is not high,and the signal stability and the confidentiality of information transmission are also difficult to guarantee.Due to the booming development of lighting technology,LED lights are widely used in indoor spaces because of their lower power consumption,longer use time,and stronger environmental protection.The high-speed light and dark flicker signals given off by LED lights can be used to transmit information,and combined with visible light detection technology,densely distributed LEDs can also be used for high-precision indoor space positioning.Visible light positioning technology is an indoor positioning technology which attracts attention and adopts extensive research.In terms of visible light detection,there are mainly two ways that photodiodes receive visible light and cameras such as CCD and CMOS receive visible light.In comparison,the latter can overcome the shortcomings of photosensitive diodes that are susceptible to interference from peripheral light sources.Thanks to the booming development and popularization of mobile smart terminals,as well as the substantial increase in processor performance,the research on image sensor-based visible light positioning technology has important practical significance.In comparison with other visible light positioning technologies,the visible light indoor positioning technology on the basis of image sensors has the good point of high positioning precision,low cost,and strong portability.This project intends to use LED visible light combined with visual technology to carry out indoor positioning technology research.This article primarily illuminates the research background and significance of visible light research,as well as the present status of research in domestic and overseas,introduces the current status of visible light indoor positioning on the basis of image senors,and also investigates the sensors used in this indoor positioning system,the imaging principles of image sensors,and singular values.Decomposition and related theoretical knowledge of related visual image processing technology.Based on the above,the subsequent research contents have be accomplished:(1)Combining the characteristics of the LED light source,the working principle of the image sensor and the coordinate system conversion process from three-dimensional space to two-dimensional space,a light source arrangement model that can quickly identify different LED light source sequences is proposed,and theoretical modeling and analysis are carried out.(2)Aiming at the problem of reduced positioning accuracy caused by the tilt angle when the image sensor captures an image,an improved spatial positioning method fusing IMU and image sensor data is proposed.This method uses both the IMU inertial sensor and the imaging model to estimate the posture of the image sensor,and combines the postures of the two to construct the objective function to minimize the positioning error.The data fusion is transformed into an optimization problem,and a numerical solution algorithm based on non-iterative singular value decomposition is proposed to optimize the solution,accurately estimate the sensor attitude,and improve the positioning accuracy of the visible light indoor space.(3)A scheme based on morphology and connected domain algorithm to accurately identify the center of the spot is proposed.Its spot recognition rate is more accurate than the traditional Hough circle detection method.The pixel coordinates obtained after image processing are brought into the trilateral positioning algorithm to obtain the position coordinates of the imaging sensor.The results of experiment indicates that the positioning error is within 4cm.