柑橘叶面病害监测颜色空间改进算法研究

Improved Algorithm of Color Space for Citrus Leaf Disease Monitoring

柑橘叶面病害是影响柑橘产量和质量的重要因素,如何对柑橘叶面进行病害识别是后期病害检测的一个关键步骤。为了提高柑橘在各种病害环境下病斑图像识别的准确率,提出了一种针对柑橘病害监测的HSV颜色直方图空间的图像检索改进算法。此算法用颜色空间来展现柑橘病害所导致的颜色变动,根据柑橘叶面病变产生的异常颜色信息,结合传统直方图图像检索方法,对柑橘叶面图像颜色进行非均匀量化处理,且将非均匀化量化处理后的归一化颜色矩阵作为因子来进行图像检索。使用者可以经过使用这种形式来检测柑橘叶面图像,及早发现病害。与传统的方案相比,此算法在柑橘叶面病害监测方面的检索查准率和查全率均有显著提高,从而验证了本文算法的有效性。

Citrus leaf diseases are important factor affecting the yield and quality of citrus. How to find the disease of citrus leaves is a key step in the late disease detection. In order to improve the accuracy of image recognition in orange leaf spots in various disease conditions, a new citrus leaf diseases image retrieval method based on HSV(Hue, Saturation, Value)Histograms is proposed, which by using color space to show the citrus diseases lead color changes, according to the abnormal color of citrus leaf disease information, combined with image retrieval method which based on the traditional histogram, non-uniform quantization of the citrus leaf color image is used in this method, and the heterogeneity of normalized color matrix quantization processed as a factor for citrus leaf image retrieval. The user can use the method to find the Citrus leaf diseases by retrieving images of citrus leaf spots. Compared with the traditional image retrieval methods, the method can not only improve retrieval speed, but also have better results.