近日,信息學(xué)院李國(guó)亮教授團(tuán)隊(duì)在國(guó)際知名期刊Artificial Intelligence in Agriculture上發(fā)表題為“FGPointKAN++ point cloud segmentation and adaptive key plane recognition for cow body size measurement”的研究論文,該研究報(bào)道了融合點(diǎn)云分割與自適應(yīng)平面識(shí)別技術(shù)的奶牛體尺參數(shù)智能測(cè)量新方法。
團(tuán)隊(duì)圍繞現(xiàn)代畜牧業(yè)中奶牛體型精準(zhǔn)測(cè)量這一問題,面對(duì)傳統(tǒng)人工測(cè)量效率低下且易受動(dòng)物姿態(tài)影響的行業(yè)痛點(diǎn),創(chuàng)新性提出基于深度學(xué)習(xí)的像素級(jí)點(diǎn)云分割模型與自適應(yīng)關(guān)鍵切面識(shí)別算法,為實(shí)現(xiàn)規(guī)?;B(yǎng)殖場(chǎng)奶牛生長(zhǎng)監(jiān)測(cè)自動(dòng)化提供了可靠技術(shù)方案。該研究通過構(gòu)建FGPointKNN++分割模型和AKCPR關(guān)鍵切面識(shí)別算法,實(shí)現(xiàn)不同姿態(tài)下奶牛像素級(jí)點(diǎn)云分割(mIoU達(dá)83.24%)與體尺參數(shù)自動(dòng)測(cè)量。
實(shí)驗(yàn)表明,該方法在真實(shí)養(yǎng)殖環(huán)境中針對(duì)不同姿態(tài)牛體高、體寬、胸圍和腹圍等關(guān)鍵表型參數(shù)的測(cè)量平均絕對(duì)百分比誤差分別低至2.07%、3.56%、2.24%和1.42%。同時(shí),本研究使用開源數(shù)據(jù)集對(duì)本方法的有效性進(jìn)行交叉驗(yàn)證,證明該技術(shù)通過三維點(diǎn)云幾何特征解析,突破動(dòng)態(tài)姿勢(shì)下的體尺精準(zhǔn)計(jì)算瓶頸,為智能化畜牧養(yǎng)殖提供了可嵌入自動(dòng)化管理系統(tǒng)的無接觸測(cè)量方案。
華中農(nóng)大信息學(xué)院李嘉位老師為本文通訊作者,2023級(jí)碩士研究生周國(guó)源為第一作者;其他作者包括信息學(xué)院2023級(jí)博士研究生李勝,2024級(jí)碩士研究生趙健、王智文及工學(xué)院2023級(jí)碩士研究生葉文昊等。信息學(xué)院李國(guó)亮教授、動(dòng)科動(dòng)醫(yī)學(xué)院張淑君教授指導(dǎo)了該項(xiàng)工作。該研究得到新疆農(nóng)墾科學(xué)院及其合作牧場(chǎng)大力支持。
據(jù)悉,Artificial Intelligence in Agriculture期刊重點(diǎn)關(guān)注人工智能在農(nóng)業(yè)領(lǐng)域研究和應(yīng)用,最新年度影響因子IF=12.4,為中科院一區(qū)TOP期刊,影響因子位列農(nóng)業(yè)信息領(lǐng)域期刊排名首位。該團(tuán)隊(duì)將以本成果基礎(chǔ),持續(xù)積累原始數(shù)據(jù)、開展后續(xù)研究,加快養(yǎng)殖領(lǐng)域的智能化、無人化轉(zhuǎn)型進(jìn)程。該研究獲得國(guó)家自然科學(xué)基金和中央高校基本科研業(yè)務(wù)費(fèi)專項(xiàng)資金資助。
原文鏈接:https://www.sciencedirect.com/science/article/pii/S2589721725000662
英文摘要:
Accurate and efficient body size measurement is essential for health assessment and production management in modern animal husbandry. To realize the segmentation of the point clouds at the pixel-level and the accurate calculation of body size for the dairy cows in different postures, a segmentation model (FGPointKAN++) and an adaptive key cutting plane recognition (AKCPR) model are developed. FGPointKAN++ introduces FGE module and KAN that enhance local feature extraction and geometric consistency, significantly improving dairy cow part segmentation accuracy. The AKCPR utilizes adaptive plane fitting and dynamic orientation calibration to optimize the key body size measurement. The dairy cow body size parameters are then calculated based on the plane geometry features. The experimental results show that mIoU scores of 82.92 % and 83.24 % for the dairy cow pixel-level point cloud segmentation results. The calculated Mean Absolute Percentage Errors (MAPE) of Wither Height (WH), Body Width (BW), Chest Circumference (CC) and Abdominal Circumference (AC) are 2.07 %, 3.56 %, 2.24 % and 1.42 %, respectively. This method enables precise segmentation and automatic body size measurement of dairy cows in various walking postures, showing considerable potential for practical applications. It provides technical support for unmanned, intelligent, and precision farming, thereby enhancing animal welfare and improving economic efficiency.
日期:2025-08-13
團(tuán)隊(duì)圍繞現(xiàn)代畜牧業(yè)中奶牛體型精準(zhǔn)測(cè)量這一問題,面對(duì)傳統(tǒng)人工測(cè)量效率低下且易受動(dòng)物姿態(tài)影響的行業(yè)痛點(diǎn),創(chuàng)新性提出基于深度學(xué)習(xí)的像素級(jí)點(diǎn)云分割模型與自適應(yīng)關(guān)鍵切面識(shí)別算法,為實(shí)現(xiàn)規(guī)?;B(yǎng)殖場(chǎng)奶牛生長(zhǎng)監(jiān)測(cè)自動(dòng)化提供了可靠技術(shù)方案。該研究通過構(gòu)建FGPointKNN++分割模型和AKCPR關(guān)鍵切面識(shí)別算法,實(shí)現(xiàn)不同姿態(tài)下奶牛像素級(jí)點(diǎn)云分割(mIoU達(dá)83.24%)與體尺參數(shù)自動(dòng)測(cè)量。
實(shí)驗(yàn)表明,該方法在真實(shí)養(yǎng)殖環(huán)境中針對(duì)不同姿態(tài)牛體高、體寬、胸圍和腹圍等關(guān)鍵表型參數(shù)的測(cè)量平均絕對(duì)百分比誤差分別低至2.07%、3.56%、2.24%和1.42%。同時(shí),本研究使用開源數(shù)據(jù)集對(duì)本方法的有效性進(jìn)行交叉驗(yàn)證,證明該技術(shù)通過三維點(diǎn)云幾何特征解析,突破動(dòng)態(tài)姿勢(shì)下的體尺精準(zhǔn)計(jì)算瓶頸,為智能化畜牧養(yǎng)殖提供了可嵌入自動(dòng)化管理系統(tǒng)的無接觸測(cè)量方案。
華中農(nóng)大信息學(xué)院李嘉位老師為本文通訊作者,2023級(jí)碩士研究生周國(guó)源為第一作者;其他作者包括信息學(xué)院2023級(jí)博士研究生李勝,2024級(jí)碩士研究生趙健、王智文及工學(xué)院2023級(jí)碩士研究生葉文昊等。信息學(xué)院李國(guó)亮教授、動(dòng)科動(dòng)醫(yī)學(xué)院張淑君教授指導(dǎo)了該項(xiàng)工作。該研究得到新疆農(nóng)墾科學(xué)院及其合作牧場(chǎng)大力支持。
據(jù)悉,Artificial Intelligence in Agriculture期刊重點(diǎn)關(guān)注人工智能在農(nóng)業(yè)領(lǐng)域研究和應(yīng)用,最新年度影響因子IF=12.4,為中科院一區(qū)TOP期刊,影響因子位列農(nóng)業(yè)信息領(lǐng)域期刊排名首位。該團(tuán)隊(duì)將以本成果基礎(chǔ),持續(xù)積累原始數(shù)據(jù)、開展后續(xù)研究,加快養(yǎng)殖領(lǐng)域的智能化、無人化轉(zhuǎn)型進(jìn)程。該研究獲得國(guó)家自然科學(xué)基金和中央高校基本科研業(yè)務(wù)費(fèi)專項(xiàng)資金資助。
原文鏈接:https://www.sciencedirect.com/science/article/pii/S2589721725000662
英文摘要:
Accurate and efficient body size measurement is essential for health assessment and production management in modern animal husbandry. To realize the segmentation of the point clouds at the pixel-level and the accurate calculation of body size for the dairy cows in different postures, a segmentation model (FGPointKAN++) and an adaptive key cutting plane recognition (AKCPR) model are developed. FGPointKAN++ introduces FGE module and KAN that enhance local feature extraction and geometric consistency, significantly improving dairy cow part segmentation accuracy. The AKCPR utilizes adaptive plane fitting and dynamic orientation calibration to optimize the key body size measurement. The dairy cow body size parameters are then calculated based on the plane geometry features. The experimental results show that mIoU scores of 82.92 % and 83.24 % for the dairy cow pixel-level point cloud segmentation results. The calculated Mean Absolute Percentage Errors (MAPE) of Wither Height (WH), Body Width (BW), Chest Circumference (CC) and Abdominal Circumference (AC) are 2.07 %, 3.56 %, 2.24 % and 1.42 %, respectively. This method enables precise segmentation and automatic body size measurement of dairy cows in various walking postures, showing considerable potential for practical applications. It provides technical support for unmanned, intelligent, and precision farming, thereby enhancing animal welfare and improving economic efficiency.
日期:2025-08-13