Research On Target Recognition And Location Technology Of Palletizing Robot Based On Binocular Vision

The vision system is a very important direction in robot technology,and the palletizing robot is gradually applied to all aspects of the logistics industry.Now,as the orders of the tobacco logistics center are moving towards small batches and high frequency,the efficiency of cigarettes palletizing is becoming more and more important for tobacco logistics.Robots that incorporate vision technology can more accurately sense the surrounding environment and make production operations more intelligent.The target recognition and localization technology of the palletizing robot based on binocular vision studied in this thesis has important theoretical and practical significance for the cigarette code.Firstly,according to the requirements of the target recognition and positioning of the palletizing robot,an overall scheme for target recognition and positioning of the palletizing robot is designed.According to this scheme and the actual research needs,the calibration of the camera is completed,and the internal and external parameters are obtained.For the problem that the original image contains a large amount of information and noise interference,the image preprocessing is completed by histogram equalization and median filtering.In the research of target feature extraction,the feature extraction algorithms such as SIFT and ORB have the disadvantages of poor robustness and long detection time.The SURF algorithm is used to extract the feature points of the target object image.This method introduces the integral image into the image feature detection.In addition,the stability of feature detection is improved,and the efficiency of detection is also improved.In the research of cigarette recognition,the texture features of the target image are obviously identified for this thesis.In this thesis,template matching based on feature point detection is used to identify objects.Experiments show that this method has accurate recognition results and short time.Next,for the positioning of the cigarette,because the traditional positioning algorithm takes a long time and the positioning accuracy is low,this thesis proposes an improved positioning method based on SURF algorithm matching.It solves the problem that the traditional binocular matching positioning search feature points need to search in the two-dimensional image plane for poor real-time performance.Given the polar constraint,the search range is reduced from two-dimensional to one-dimensional,which shortens the matching time and improves the matching accuracy.The experiment shows that the method has the advantages of high speed and high accuracy when positioning the cigarette.Finally,a palletizing robot target recognition and positioning system is constructed,and the system hardware and software design is completed.Experiments show that the recognition and localization algorithm used in this thesis not only has good robustness,but also its timeliness meets the requirements of practical engineering.