Research On Instrument Reading Recognition Algorithms Based On LabVIEW

Machine vision is a subject that uses machine to imitate the process of biometric recognition.It collects images by camera and processes them by computer.Then it extracts features and identifies object attributes,States and other information through related algorithms.In industrial production,in order to better monitor various data,a large number of instruments and equipment are needed.In the old factories and some special places,pointer meters still occupy a dominant position.At present,manual reading is mostly used in instrument reading.Manual reading is not only inefficient,but also influenced by subjective factors.Long-term work will inevitably lead to inaccurate reading.Therefore,it is necessary to combine machine vision technology with instrument reading.In this paper,machine vision is applied to the automatic recognition of the reading of pointer instrument,and the image of instrument is captured by camera.Usually due to the complexity of the environment in which the instrument is located and the camera itself,the collected instrument image has low contrast and many noise points.Aiming at the problem of low image contrast,this paper uses histogram equalization to improve image contrast from the perspective of machine vision and digital image processing technology,so as to enhance the contrast of dial area and pointer.Image smoothing technology is used to process the image and filter out the noise of the image.The maximum interclass variance segmentation algorithm is used to segment the image by threshold,and the binary image is obtained.In the phase of pointer line detection,this paper proposes the following two innovations: 1.By establishing a search circle,the pointer position can be determined quickly.2.Based on the traditional Hough transform,an improved Hough transform algorithm based on ROI location is proposed in this paper.By establishing a rectangular ROI to frame the pointer,and then only need to detect in this area,and to constrain the detected line,the equation of the axis of the pointer can be obtained.Finally,the current reading of the instrument is obtained by the mapping relation of direction and value.The system can automatically inspect several instruments in turn.In this paper,a total of 30 state thermometers are selected as test objects,and the traditional algorithm and the algorithm in this paper are used for reading recognition.From the test results,it can be seen that the maximum reference error of this algorithm is 0.4%,which is close to 0.5 level accuracy.It meets the 1.0 level accuracy requirement of the instrument.The detection speed and accuracy are better than traditional Hough linear detection.