Research On Instrument Recognition Technology And Development Of Instrument Recognition System Based On Intelligent Robot

With the development of science and technology,more and more fields begin to use intelligent system.As a kind of intelligent processing technology,instrument recognition technology has been widely used in industrial field,and more and more attention has been paid to it.Although there are a lot of researches on fixed instrument recognition,there are few researches on non-fixed instrument recognition.In this paper,a non-fixed instrument recognition system based on intelligent robot is developed.It obtains the instrument image through the camera equipped with the robot,and realizes the recognition of multi-type instruments.The main work of this paper includes the research of instrument recognition algorithm and the development of multi-type instruments recognition system for non-fixed instrument recognition in power field.In order to solve the problem of the randomness of instrument location in non-fixed instrument identification,and to reduce the difficulty of instrument recognition and in the meantime improve the recognition rate,a calculation model based on instrument detection,instrument registration and instrument identification is proposed.The recognition method based on instrument modeling is designed to reduce the complexity of the recognition algorithm and improve the accuracy and robustness of the recognition algorithm,and also to achieve the versatility and compatibility of the system.At the same time,this paper also improved some methods in instrument recognition algorithm.In this paper,a new detection algorithm based on structured support vector machine is proposed.The algorithm uses structured SVM as the classifier,and makes full use of the geometric relationship between the target and other objects in the background to improve the detection accuracy.At the same time,the algorithm uses the method of limiting contrast adaptive histogram equalization to preprocess image,which is helpful to reduce the influence of illumination and other environmental factors on the image.Secondly,this paper proposes an image registration algorithm based on SURF feature.The algorithm performs feature matching using SURF feature and BF algorithm.PROSAC algorithm is used for matching pair selection.At the same time,some improvements are made in the feature point matching algorithm,including the restriction of registration area and the pre-selection of matching pairs.The image registration algorithm is a dependency of the subsequent recognition algorithm,which is helpful to reduce the difficulty of recognition and improve the recognition accuracy.Thirdly,this paper presents a recognition algorithm for two common instrument types.The pointer type instrument recognition is based on a pre-modeling algorithm and an image rotation method.The pre-modeling algorithm makes full use of the prior information of instrument templates,which reduces the difficulty of pointer detection.The image rotation method extracts the matching samples in the horizontal region after rotating the image,which reduces the difficulty and calculation amount compared to rotating the search window directly.Digital instrumentation recognition algorithm identifies numbers by KNN classifiers.The algorithm makes full use of a priori information for the tilt correction of the digital display area and digital rough positioning,which improves the accuracy of digital segmentation.The method proposed in this paper also uses the relationship between contour and convex hull to recognize the decimal point separately,which has better effect than the direct classification.Finally,for the cooperative project with Dali Company,this paper developed the instrument recognition system based on the intelligent robot.The system consists of the instrument modeling software and the instrument recognition algorithm,which completely realized the whole process of detection,registration and recognition with high accuracy and speed when the intelligent robot stop from the preset point.The instrument recognition system has been put into practical.use.