FPGA图像处理系统在棉种分选中的应用实现

Application and Implementation of FPGA Image Processing System in Cottonseeds Sorting

为满足新疆脱绒棉种对外观品质如红种、破碎棉种、开口棉种及瘪籽异型的检测分选要求,以FPGA为核心处理器构建棉种图像处理系统的硬件平台。具体通过建立样品图像的HSV色彩空间模型,分析正常种、红种在色相H、饱和度S特征量的差异,以及破碎粒、开口粒在色度H的差异。同时建立面积-外围周长比因子及相应的不合格品分离度量参数,进一步通过阈值分割算法实现对异色、破碎、开口物料的判定。试验结果表明,该方法对棉种中的不合格品识别率达99%,为进一步完善种子类外观品质检测提供了理论依据。

In order to meet the detecting and sorting requirements for appearance quality of Xinjiang cottonseeds, e. g. , red seeds, broken seeds, openings seeds and profiled shriveled seeds, the hardware platform of the cotton seeds image processing system is constructed with FPGA as the kernel processor. Through establishing the HSV color space model of the seeds image,the difference between normal seeds and red seeds in H and S feature amount; and the difference between broken seeds and openings seeds in H are analyzed; and the ratio factor of area - periphery perimeter is setup,then the separating metrics parameters of the unqualified products are established; and thus the judgment of the different color, broken,and openings materials is implemented through threshold separation algorithm in firrther. The test results show that the recognition rate of the unqualified products in cottonseeds is up to 99% ; the method provides theoretical basis for detecting appearance quality of various seeds.