Hardware Realizations Of Real-Time Detection Algorithm For Defects Based On FPGA

Since the textile industry has entered the era of automation, the technological progress leads to the rapid raise of production efficiency, but the development of quality detection technology is far behind the progress of textile product industry and becomes a bottleneck which lags the labor efficiency of the textile industry. The undeveloped manual detection system based on naked eyes has not only a high false reject rate but low efficiency, and brings the labors repeated work which easily leads to over tiredness. As a result the automation of defect detection is put on the agenda. With the rapid development of large-scale integrated circuits and detection algorithms, machine vision in the fabric defect detection and other industrial surface detection areas emerges as an indispensable technology in automated inspection..FPGA as one of the chips which has the shortest circuit design cycle, lowest development cost and smallest risk in ASIC, is playing a more important role in various hardware platforms. Its powerful logic operating capabilities, flexible function realization and I / O configurations makes us to choose it as the algorithm realization chip in the hardware platform. Assisted with the high-performance, low-power, low-cost ARM7 processor we implement the construction of the hardware platform of real-time defect detection algorithm. In this thesis, we introduce the construction and the developed process of the FPGA development board with the Altera Stratix FPGA family as the core and the ARM Linux development board with the 32 bit ARM7TDMITM network processor S3C4510BTM as the core. Through comparing two different methods of communicating with two development board we at last choose one which is more suitable for the chosen hardware. Replanted theμC/OS-II operating system into the ARM Linux development board we implement the communication between the PC and the serial port in the board, the management of the extend system bus and the communication between the ARM Linux development board and the FPGA development board. In the same time we set up the NiosII soft-core processor into the Altera Stratix development board to implement the real-time detection algorithm based on developed GLCM( Gray Level Co-occurrence Matrix) which verify the feasibility of the hardware designing scheme. GLCM is one of the most effective methods to get the character of texture. Because of its regularity, textile is propitious to use GLCM to discriminate the defect. Finally the theory of GLCM and the developed method—LCM which is more useful in implementing the real-time algorithm is researched, and the real-time performance and effectiveness of the whole system is checked.