CAJ | 학술논문

以ARM92440处理器为核心，基于BP神经网络的机器人视觉识别和定位，研制了一种苹果采摘机器人及控制系统。首先应用 Photoshop软件将图像分割为640×640像素的图片，然后在RGB颜色空间下，结合数学形态学中值滤波降噪和拉普拉斯算子边缘检测的方法实现果实目标的特征提取，最终完成苹果的定位并在 MatLab环境下使用robotics－toolbox工具箱编程仿真。试验结果表明：该机器人能有效识别果实并完成抓取工作，为后续的深入研究奠定了基础。
이ARM92440처리기위핵심，기우BP신경망락적궤기인시각식별화정위，연제료일충평과채적궤기인급공제계통。수선응용 Photoshop연건장도상분할위640×640상소적도편，연후재RGB안색공간하，결합수학형태학중치려파강조화랍보랍사산자변연검측적방법실현과실목표적특정제취，최종완성평과적정위병재 MatLab배경하사용robotics－toolbox공구상편정방진。시험결과표명：해궤기인능유효식별과실병완성조취공작，위후속적심입연구전정료기출。
A kind of apple-picking robert and the control systems have been developed.The whole system is based on processor ARM92440 , which takes advantage of the Robert ’ s visual recognition and positioning ability based on the Neural Network.Firstly, the processor will split the image into pictures of pixels by using the software of Photoshop. Then under the color space of RGB, combining mathematical morphology median filtering and noise reduction Laplacian edge detection methods, the feature extraction of the target fruit could finally be achieved and the location of the fruit will be completed successfully.The processor ARM 92440 drives the four-dimensions-of-freedom mechanical arm to com-plete the clutching experiment for apples indoors.The experimental results show that the fruit is effectively identified and picked, which laid strong foundation for the subsequent thorough study.