基于OpenCV的棕榈果计数系统的研究与实现

The Research and Implementation of Palm Fruit Counting System Based on OpenCV

在大面积的棕榈树种植园,棕榈果采摘后,主要依靠人工计数统计,存在计数效率低、误差大甚至虚报数量等现象。基于此,提出并实现了一种基于OpenCV的棕榈果自动计数系统,系统首先以拍摄的棕榈果照片为输入,经过图像降维、降分辨率、灰度化和滤波等预处理,以图像的灰度直方图中两峰之间的鞍值作为二值化的阈值,能够较好地自适应图片中光线的影响,提取出识别目标并剔除背景;然后对图像采用形态学展开操作;最后在图像中查找联通区域,并剔除形态失真的区域,统计并输出棕榈果的计数结果。经试验验证,识别效果较好,计数准确,系统具有一定的实用价值。

In a large area of palm plantations,after harvesting palm fruits,it mainly relies on manual counting statistics.There are some phenomena such as low counting efficiency,large errors and even false numbers.Based on this,an automatic counting system of palm fruits based on OpenCV is proposed and implemented.Firstly,the system takes the photograph of palm fruits as input,preprocesses the image by dimension reduction,resolution reduction,gray level and filtering,and takes the saddle value between two peaks in the gray histogram of the image as the threshold of binarization,which can better adapt to the influence of light in the image and extract the recognition eye.The background is marked and removed;then the image is operated by morphological expansion;finally,the connected area is found in the image,and the morphological distortion area is eliminated,and the counting results of palm fruits are counted and output.The experimental results show that the recognition effect is good and the counting is accurate.The system has certain practical value.