Study On Moving Target Detection Based On Artificial Compound Eyes

Academia and industry pay attention to insect compound eye, because it has a lot of advantages:a wide field, high sensitivity, multi-channel and so on. Scholars around the world have launched the exploration and research on it. At the same time, the study of using bionic compound eye to target is also quite hot. The realization of the bionic compound eye methods are also different. Due to the restriction of image sensor technology, the biggest difficulty lies in the imaging of curved compound eye. And affected by the micro processing technology, its own structure also limits many advantages of compound eyes, such as the spatial resolution. The innovation of this article is converge the ommatidium unit signal through the optical to the same plane, and imaging the compound eye’s signals by the planar CCD. This article combines the 3D printing technology and machinery processing technology to achieve a bionic compound eye system, developed a set of imaging and image processing system and related computer software.In this article, the compound eye ball processing used the 3D printing technology, and processed set of the ommatidium unit through the technology of mechanical processing. The ommatidium unit is installed in the compound eye ball. All the ommatidium unit’s signals converge in the same plane through the optical fiber, constituted a fiber end face array. Imaged optical fiber end surface by CCD image sensor, and the optical fiber end face form a spot array in the image. To determine whether there is a goal by capturing the changes of the spot‘s gray level information. This article designed and realized an imaging and image processing system with CPLD as the center, stm32f4 assisted, which is used to obtain, extraction and process the information of compound eye. In this system, the main components included a CCD, CPLD, AD9840 A, STM32F4, two pieces of SRAM, etc. Every piece of SRAM can store a frame of image data. When the CPLD write one piece of SRAM, the STM32F4 read another piece of SRAM, two pieces of SRAM alternating read and written. The CPLD driving the CCD image sensor and the AD9840 A of CCD signal processing chip, and coordinate the exchange of information between STM32F4 and two pieces of SRAM. Also developed a set of PC software based on MFC. The data is transmitted to the computer via the Internet after processing, it established a TCP connection between imaging and image processing system and computer software. To achieve the three-dimensional display of the target trajectory by transplanted the OPENGL to MFC framework.After the design of imaging and image processing system, the tests were carried out. Then this article made a lot of experiments about target locations in the day and night. And selected 37 representative set of data for analysis. The coordinates of X-axis, Y-axis and Z-axis, respectively range from-250 cm to + 250 cm, + 50 cm to + 400 cm, + 50 cm to + 300 cm in the experiment. It can be found that the maximum error is less than 19 cm by contrasted the detected and measured values for each experimental point’s X-axis, Y-axis and Z-axis coordinates, and the data illustrated that this system can accurately to locate targets within the range of coordinates in space experiments, it also verified the feasibility of this project.