李立君,阳涵疆.基于改进凸壳理论的遮挡油茶果定位检测算法[J].农业机械学报,2016,47(12):285-292,346.
Li Lijun,Yang Hanjiang.Revised Detection and Localization Algorithm for Camellia oleifera Fruits Based on Convex Hull Theory[J].Transactions of the Chinese Society for Agricultural Machinery,2016,47(12):285-292,346.
摘要点击次数: 2234
全文下载次数: 887
基于改进凸壳理论的遮挡油茶果定位检测算法   [下载全文]
Revised Detection and Localization Algorithm for Camellia oleifera Fruits Based on Convex Hull Theory   [Download Pdf][in English]
投稿时间:2016-02-29  
DOI:10.6041/j.issn.1000-1298.2016.12.035
中文关键词:  凸壳理论  油茶果  图像分割  轮廓重建  遮挡果实  凹点
基金项目:国家林业公益性项目(201104090)、湖南省高校科技创新团队支持计划项目(2014207)、湖南省研究生科研创新项目(CX2016B326)和中南林业科技大学研究生科技创新基金项目(CX2016B12)
作者单位
李立君 中南林业科技大学 
阳涵疆 中南林业科技大学 
中文摘要:针对传统凸壳理论进行遮挡果实定位检测时由于过多剔除有效轮廓,造成目标果实定位误差较大,甚至无法识别目标果实的问题,提出了一种基于改进凸壳理论的遮挡油茶果定位检测算法。首先利用基于颜色特征的阈值分割法对油茶果遮挡图像进行目标分割,并通过预处理操作剔除图像中的背景噪声,获得目标果实的二值图像;然后采用凹点搜寻算法检测重叠目标的凹点,并根据凹点对重叠目标进行分离,获得相互独立的目标图像;再构建各独立目标的凸包,并提取凸壳,利用轮廓提取算法确定各独立目标凸壳上的有效轮廓;最后根据提取的有效轮廓求解目标果实形心坐标和半径,完成遮挡果实的定位检测。试验结果表明,改进算法平均耗时为0.491s,比传统凸壳方法增加了24.07%,但其仅占油茶果采摘机器人单个果实采摘周期的2.46%,对于图像中的遮挡油茶果目标,改进方法的识别率达到93.21%,相比传统凸壳方法提升了7.47个百分点,改进算法的平均定位检测误差和平均重合度分别为5.53%和93.43%,比传统凸壳算法平均定位误差降低了6.22个百分点,平均重合度提高了6.79个百分点,表明文中所提出的方法能够较好地识别和定位自然环境中的遮挡油茶果。
Li Lijun  Yang Hanjiang
Central South University of Forestry and Technology and Central South University of Forestry and Technology
Key Words:convex hull  Camellia oleifera fruits  image segmentation  contour reconstruction  occluded fruits  concave points
Abstract:The existing method based on convex hull theory has low detecting ratio and large locating error because of failing to extract effective contour of the concave regions for occluded fruits. In order to improve recognition accuracy and reduce error of the current method, a kind of improved algorithm for detecting and locating occluded Camellia oleifera fruits was proposed. Firstly, in order to get a grayscale image of occluded Camellia oleifera fruits, different color spaces of the original image were compared and then R—B chromatic aberration characteristic was chosen. The Otsu method was used to segment the grayscale image and the morphological operation was employed to remove residual noise, thus the regions of targets and backgrounds can be successfully separated by the algorithm. A kind of algorithm was used to extracte convex closure of each occluded regions and then the concave regions were obtained by subtracting the binary image from its convex closure image. The regions with pixels less than half of the biggest one in concave image were removed and the intersection points or concave points of occluded Camellia oleifera fruits were detected by a kind of concave point detection algorithm, then the occluded targets were separated by using Bresenham line drawing algorithm according to the intersection points. Convex closure of each separated regions was built and convex hull was extracted from it, after that a kind of ineffective contour removing algorithm was used to extracte effective contour that used to reconstruct the target contour from each convex hull. Contour reconstruction algorithm was used to rebuild the target contour of the occluded Camellia oleifera fruits based on the points of each corresponding effective contour, and then the reconstruction contour was merged that the distance between their centers was below the threshold value. In order to validate the performance of the improved algorithm, a comparative test was conducted, and the positioning errors were calculated. The test results showed that it needed 0.491 s to finish the recognition and location process in average by the proposed method, which accounted for only 2.46% of the total time-consuming for a single Camellia oleifera fruit by harvesting robot. Average recognition success rate of occluded Camellia oleifera fruits by the proposed method was 93.21%, which was 7.47 percentage points higher than that of the original method. Average segmentation error of the proposed method was 5.53%, which was reduced by 6.22 percentage points compared with that of the original method. Average overlap ratio of the proposed method was 93.43%, which was 6.79 percentage points less than the that of the traditional method. The test results indicated that the proposed method was feasible and effective to recognize and locate occluded Camellia oleifera fruits.

Transactions of the Chinese Society for Agriculture Machinery (CSAM), in charged of China Association for Science and Technology (CAST), sponsored by CSAM and Chinese Academy of Agricultural Mechanization Science(CAAMS), started publication in 1957. It is the earliest interdisciplinary journal in Chinese which combines agricultural and engineering. It always closely grasps the development direction of agriculture engineering disciplines and the published papers represent the highest academic level of agriculture engineering in China. Currently, nearly 8,000 papers have been already published. There are around 3,000 papers contributed to the journal each year, but only around 600 of them will be accepted. Transactions of CSAM focuses on a wide range of agricultural machinery, irrigation, electronics, robotics, agro-products engineering, biological energy, agricultural structures and environment and more. Subjects in Transactions of the CSAM have been embodied by many internationally well-known index systems, such as: EI Compendex, CA, CSA, etc.

   下载PDF阅读器