Research On Robot Image Sensor Technology

It has been half a century since the birth of robot in 1960s. During this period, with its rapid development, robot has been extensively used and almost pervades every field. No matter the 1st industrial robot-"Unimate" in the world, or the representative of highest level for today's robot technology-"Asimo". It is the effect of various sensors that robot can get the circumstantial information and do some intelligent feedback actions. Therefore, the research on sensor is one of the basic and focus sections of intelligent robotics study, and also the first step to achieve robot's self-control.70% of human perception to the world comes from visual information, so "eye" is very important perception equipment to a robot. Along with the birth and development of CCD and COMS image sensor, robot can both obtain 2D images of 3D features. By using of visual processing algorithms, it also can process, compare, analyse and comprehend one or multiple images and finally get symbol description of relevant features. This provides useful information for specific tasks and induces robot's action. As human's eyes, image sensor to a robot is an important "organ" to perceive regional environment. This kind of sensor has the advantages of wide range detection and rich signal, which makes accurate access to information by vision sensor come to be a priority of intelligent robot research. Taking the image sensing technology as the start of this research,3 experiment devices are built. Both autonomous navigation of Mobile Robot and extraction of target information from the image taken by Aerial Robot are fully achieved in this thesis. In the 1st experiment, image signals are captured and processed in stimulation method by using of line-Array CCD sensor. In the 2nd experiment, integrate sensor based on CMOS image sensor chips is used, and the output PAL video signals will be first separated and decoded, then converted to digital signal. The information needed is gained through comparing and anasysing A/D results. In the 3rd experiment, video signal is converted to digital image by image acquisition card. The connected areas between objectives are recognized, and their number and location are reported after processing single frame image. In these experiments, the image processing carriers based on MCU and PC respectively; and the first two are based on the MCU application system which is simple and of high flexibility. Since their operation capability is limited, they only suits to process graphic information of less data and in a bottom-level information processing manner. In the last experiment, image acquisition card of USB2.0 interface is used to convert image signal to digital image and send the converted result to PC. Image processing program is developed based on Windows operation system in VC++6.0 integrated environment with the SDK of image acquisition card. Target information is captured by PC's powerful operation capability and this process based on system-level information processing.