散装仓粮食数量识别关键——矩形标尺图像识别

Key of bulk warehouse grain quantity recognition——Rectangular benchmark image recognition

根据散装粮仓粮食数量识别要求,以场景视频图为识别对象,对通过边缘检测得到的差分结果进行迭代分析,得到对象边界.基于与识别目标相吻合的梯度算子的区域迭代阈值进行图像特征二次提取,利用模糊识别隶属度函数进行矩形标尺判断,有效地提高了识别算法的抗干扰性、鲁棒性、识别精度和效率,并用Visual C++实现了该识别算法.该方法为杜绝粮食储备管理中的弊端,提供了一种粮食储备的智能稽核方法.

According to the requests of bulk warehouse grain quantity recognition, the scene video was taken as the identified object to obtain the object＇s boundary from the result of edge detection difference iterative analysis, The region iterative threshold value of the gradient operator fitting closely with the identified target was used to execute the picture characteristic second-extraction, then rectangular benchmark judgment using the membership functions of fuzzy recognition was carried out, and finally this recognition algorithm was realized by adopting Visual C＋＋. Experimental results showed that this recognition algorithm effectively enhances the anti-jamming capability, robustness, recognition precision and efficiency. This work provides a grain reserving intelligent audit method for eliminating the shortcomings in grain reserving management.