The Development Of Embedded System Of Ranging Based On Image

In recent years, images ranging has been aroused widely based on computer vision. The principle of this technology is to imitate human eyes through cameras. In order to get two different images from different angle, it usually uses two cameras to observe a specific target. Then through the parallax, it calculates range of this target. In the meantime, embedded system also widely boomed. With the emergence of various embedded processor system structure, embedded System On Chip(SOC) will replace traditional CPU. Moreover, embedded operating system has small kernel, strong specificity, streamlined system and etc. Hence, with the development of this technology, achieve image measurement by embedded system has become a trend.This project aims to develop a set of real time image measurement of embedded system platform, and it mainly supports two cameras input, LCD display and Ethernet function. The system design is based on the idea of software and hardware co-design and functional requirements. The main work is to accomplish the following points.1.The design of hardware system.32 bit ARM processor S3C2440 of Samsung Corp for this system is the controlling center. The method using two pieces of 16 bit 32 M SDRAM is to achieve 32 bit of total capacity of 64 M memory, and it configures 2M Nor Flash and 256 M Nand Flash for external memory. The use of highly integrated monolithic fast Ethernet MAC controller DM9000 is to realize the Ethernet interface system; Using TVP5150 video decoder chip which is launched by TI can realize video input unit of two CVBS signal. At the same time, this design completes LCD display mold and USB-Hub mold. Then, the design of PCB board is completed based on the high speed circuit design principles.2.The design of software system.Build up Linux software development environment; complete the transplantation of bootloader system; customize linux operating system; Make the root file system. Meanwhile, the driver transplantation of the video decoder chip, network card chip, LCD display and other device is completed. In the video decoding driver, not only the read-write pattern but also memory mapping can access to image data. Memory mapping assists to save the time of reading image data, improve the efficiency of program and then enhance the rate of range.3.Complete the study and design of Ranging Algorithm. Through the calibration of the camera, we obtained image correction parameters, researched on the filter of image, extracted clarification of objective, and determined the coordinate information of the target object in the image. Calculate the object range according to coordinate information and triangle algorithm.The results show that each function module of this system operates well. The system can read the image information from the two cameras, and use the image ranging algorithm to calculate the distance of the objective. The system takes less than 500 ms to measure the distance. Meanwhile, the effective range of the target can achieve 2m, and the relative error is less than 3%.