任意形状林业目标物的像素级自动标注算法

doi:10.3778/j.issn.1002-8331.2205-0152

摘要/Abstract

摘要： 在林业智能应用问题中，经常需要对遥感图像中的非刚性目标如“火”“烟”“云”等进行识别。由于此类目标物受限于颜色不确定（或渐变）以及无固定形状，导致现有图像标注方法的性能不佳甚至失效。因此，针对如何准确地选择目标物的兴趣区域（region of interest，ROI），进而提高像素标记的速度以及准确率进行了研究。针对图像目标物形状的非凸性，提出了任意形状ROI的像素级自动标注算法，该算法可将任意形状的兴趣区域转化为多个凸区域问题，针对分解后的每个凸壳内的像素进行提取、训练、标记。基于无人机拍摄的高清图像集，通过将其与边界框法、凸壳法、图像分割以及图像抠图方法进行实验对比，证明该方法存在易操作，符合人类视觉，求解速度快，可批量操作，且能够实现非刚性目标的像素级自动标注等特点。

关键词: 图像标注, 兴趣区域, 凸壳, 任意形状

Abstract: In forestry intelligence applications, it is necessary to identify non-rigid objects, such as fire, smoke or cloud in remote sensing image. Limit to object particularity in shapelessness or color-uncertainty, these lead to poor performance or even failure of existing image annotation methods. Therefore, how to accurately select the region of interest（ROI） of the target object is studied to improve the speed and accuracy of pixel annotation. Aiming at the non-convexity of the image object shape, a pixel-level automatic labeling algorithm for ROI of any shape is proposed. Pixels are extracted, trained, and labeled. Based on high-definition images captured by a drone, comparing with the bounding box, the convex hull, image segmentation and image matting methods, the results show the superiorities in easy operation, rapidity and pixel-level precision, conforms to human vision, and can achieve automatic pixel-level annotation of non-rigid objects.

Key words: image annotation, regions of interest, convex hull, arbitrary-shaped

陈瑞园, 杨绪兵, 范习健, 张礼, 业巧林. 任意形状林业目标物的像素级自动标注算法[J]. 计算机工程与应用, 2023, 59(18): 172-178.

CHEN Ruiyuan, YANG Xubing, FAN Xijian, ZHANG Li, YE Qiaolin. Pixel-Level Automatic Annotation Algorithm for Forestry Targets of Arbitrary-Shaped[J]. Computer Engineering and Applications, 2023, 59(18): 172-178.

参考文献

[1] 崔海鸥，刘珉.我国第九次森林资源清查中的资源动态研究[J].西部林业科学，2020，49（5）：90-95.
CUI H O，LIU M.Analysis on the results of the 9th national forest inventory[J].Journal of West China Forestry Science，2020，49（5）：90-95.
[2] 樊仲谋，周成军，周新年，等.无人机航测技术在森林资源调查中的应用[J].森林与环境学报，2018，38（3）：297-301.
FAN Z M，ZHOU C J，ZHOU X N，et al.Application of unmanned aerial vehicle aerial survey technology in forest inventory[J].Journal of Forest and Environment，2018，38（3）：297-301.
[3] GAIDEL A V.Method of automatic ROI selection on lung CT images[J].Procedia Engineering，2017，201：258-264.
[4] KIADTIKORNTHAWEEYOT W，TATNALL A R L.Region of interest detection based on histogram segmentation for satellite image[J].International Archives of the Photogrammetry，Remote Sensing & Spatial Information Sciences，2016，41：12-19.
[5] JAN M M，ZAINAL N，JAMALUDIN S.Region of interest-based image retrieval techniques：a review[J].IAES International Journal of Artificial Intelligence，2020，9（3）：520-528.
[6] AMIN B，RIAZ M M，GHAFOOR A.Automatic aircraft extraction using video matting and frame registration[J].IET Image Processing，2020，14（8）：1628-1635.
[7] BROOKS J L，ZOUMPOULAKI A，BOWMAN H.Data‐driven region‐of‐interest selection without inflating Type I error rate[J].Psychophysiology，2017，54（1）：100-113.
[8] KANADAM K P，CHEREDDY S R.Mammogram classification using sparse-ROI：a novel representation to arbitrary shaped masses[J].Expert Systems with Applications，2016，57：204-213.
[9] JAYECH K，MAHJOUB M A.Clustering and Bayesian network for image of faces classification[J].arXiv：1204.
1679，2012.
[10] NUGALIYADDE A，WONG K W，PARRY J，et al.RCNN for region of interest detection in whole slide images[C]//International Conference on Neural Information Processing.Cham：Springer，2020：625-632.
[11] 陈科圻，朱志亮，邓小明，等.多尺度目标检测的深度学习研究综述[J].软件学报，2021，32（4）：1201-1227.
CHEN K Q，ZHU Z L，DENG X M，et al.Deep learning for multi-scale object detection：a survey[J].Journal of Software，2021，32（4）：1201-1227.
[12] NICOLáS L F G.Unsupervised generation of polygonal approximations based on the convex hull[J].Pattern Recognition Letters，2020，135：138-145.
[13] 宋怀波，张传栋，潘景朋，等.基于凸壳的重叠苹果目标分割与重建算法[J].农业工程学报，2013，29（3）：163-168.
SONG H B，ZHANG C D，PAN J P，et al.Segmentation and reconstruction of overlapped apple images based on convex hull[J].Transactions of the Chinese Society of Agricultural Engineering，2013，29（3）：163-168.
[14] ZHANG X，WANG C，FAN X.Convex hull-based distance metric learning for image classification[J].Computational and Applied Mathematics，2021，40（4）：1-22.
[15] NEMIRKO A P，DULá J H.Machine learning algorithm based on convex hull analysis[J].Procedia Computer Science，2021，186（8）：381-386.
[16] ANSTREICHER K M，SAMUEL B，PARK K.Convex hull representations for bounded products of variables[J].Journal of Global Optimization，2021，80：757-778.
[17] 杨绪兵，王一雄，陈斌.马氏度量学习中的几个关键问题研究及几何解释[J].南京大学学报（自然科学），2013（2）：133-141.
YANG X B，WANG Y X，CHEN B.Research on several key problems of Mahalanobis metric learning and corresponding geometrical interpretations[J].Journal of Nanjing University（Natural Science），2013，49（2）：133-141.
[18] HUANG N C，LUO Y J，ZHANG Q，et al.Discriminative unimodal feature selection and fusion for RGB-D salient object detection[J].Pattern Recognition，2022，122：108-113.
[19] KUMAR S，ONG S H，RANGANATH S，et al.A rule-based approach for robust clump splitting[J].Pattern Recognition，2006，39（6）：1088-1098.
[20] WANG Z，WANG E，ZHU Y.Image segmentation evaluation：a survey of methods[J].Artificial Intelligence Review，2020，53：5637-5674.
[21] YANG X B，CHEN R，ZHANG F Q，et al.Pixel-level automatic annotation for forest fire image[J].Engineering Applications of Artificial Intelligence，2021，104：104-123.