ass日本风韵熟妇pics男人扒开女人屁屁桶到爽|扒开胸露出奶头亲吻视频|邻居少妇的诱惑|人人妻在线播放|日日摸夜夜摸狠狠摸婷婷|制服 丝袜 人妻|激情熟妇中文字幕|看黄色欧美特一级|日本av人妻系列|高潮对白av,丰满岳妇乱熟妇之荡,日本丰满熟妇乱又伦,日韩欧美一区二区三区在线

自然環(huán)境下柑橘采摘機(jī)器人識別定位系統(tǒng)研究
CSTR:
作者:
作者單位:

作者簡介:

通訊作者:

中圖分類號:

基金項目:

重慶市重點產(chǎn)業(yè)共性關(guān)鍵技術(shù)創(chuàng)新專項(cstc2015zdcy-ztzx70003)和重慶市基礎(chǔ)科學(xué)與前沿技術(shù)研究一般項目(cstc2016jcyjA0444)


Research and Experiment on Recognition and Location System for Citrus Picking Robot in Natural Environment
Author:
Affiliation:

Fund Project:

  • 摘要
  • |
  • 圖/表
  • |
  • 訪問統(tǒng)計
  • |
  • 參考文獻(xiàn)
  • |
  • 相似文獻(xiàn)
  • |
  • 引證文獻(xiàn)
  • |
  • 資源附件
  • |
  • 文章評論
    摘要:

    為了準(zhǔn)確理解柑橘采摘機(jī)器人在自然環(huán)境下的作業(yè)場景,,獲取采摘目標(biāo)及周圍障礙物的位置信息,,構(gòu)建了基于卷積神經(jīng)網(wǎng)絡(luò)和Kinect V2相機(jī)的識別定位系統(tǒng),。首先,,對采摘場景中的果樹提出5類目標(biāo)物分類準(zhǔn)則,,包含1類可采摘果實和4類障礙物目標(biāo),;然后,,在YOLO V3(You only look once)卷積層模塊中添加3層最大池化層,,對預(yù)測候選框進(jìn)行K-means聚類分析,增強(qiáng)模型對枝葉類物體特征的提取能力,,實現(xiàn)采摘場景的準(zhǔn)確理解,;最后,采用Kinect V2相機(jī)的深度圖映射得到采摘目標(biāo)和障礙物的三維信息,,并在自然環(huán)境下進(jìn)行了避障采摘作業(yè),。實驗結(jié)果表明,構(gòu)建的識別定位系統(tǒng)對障礙物和可采摘果實的識別綜合評價指數(shù)分別為83.6%和91.9%,,定位誤差為5.9mm,,單幀圖像的處理時間為0.4s,,采摘成功率和避障成功率分別達(dá)到80.51%和75.79%,。

    Abstract:

    For citrus picking robot in natural environment, the accurate recognition and location vision system is one of the key factors ensuring the efficiency and safety of picking operations. In order to make the robot not only acquire the location information of the picking target accurately but also the surrounding obstacles, a novel obstacle recognition and location system based on Kinect V2 and improved you only look once (YOLO V3) algorithm was proposed. Firstly, five classification principles of citrus tree in natural orchard were defined, including one class that the fruit can be picked directly and four obstacle classes. Secondly, three maximum pooling layers were added to the convolution module of the YOLO V3 structure and K-means clustering analysis was conducted on anchor box to enhance the feature extraction performance of branches and leaves of the convolution neural network. Finally, threedimensional coordinates of the classification targets were obtained by using the Kinect V2 depth mapping to guide obstacleavoiding picking operation. The experimental results showed that the F-scores of obstacles and normal fruits were 83.6% and 91.9%, respectively, the positioning error was 5.9mm and the processing time of each frame was 0.4s, the success picking rate was 80.51% and success rate of obstacle avoidance was 75.79%. The research results provided a basis and guide for the picking path planning and obstacle avoidance of robotic harvesting task in natural scene.

    參考文獻(xiàn)
    相似文獻(xiàn)
    引證文獻(xiàn)
引用本文

楊長輝,劉艷平,王毅,熊龍燁,許洪斌,趙萬華.自然環(huán)境下柑橘采摘機(jī)器人識別定位系統(tǒng)研究[J].農(nóng)業(yè)機(jī)械學(xué)報,2019,50(12):14-22. YANG Changhui, LIU Yanping, WANG Yi, XIONG Longye, XU Hongbin, ZHAO Wanhua. Research and Experiment on Recognition and Location System for Citrus Picking Robot in Natural Environment[J]. Transactions of the Chinese Society for Agricultural Machinery,2019,50(12):14-22.

復(fù)制
分享
文章指標(biāo)
  • 點擊次數(shù):
  • 下載次數(shù):
  • HTML閱讀次數(shù):
  • 引用次數(shù):
歷史
  • 收稿日期:2019-04-21
  • 最后修改日期:
  • 錄用日期:
  • 在線發(fā)布日期: 2019-12-10
  • 出版日期: 2019-12-10
文章二維碼