Research On Key Technology Of Intelligent Detection Of Conveyor Belt Edge Damage

As one of the important means of modern production and transportation in China,belt conveyor is widely used in many industries,such as coal mine,chemical industry and port.As an important part of belt conveyor,conveyor belt’s common damage forms in the process of use are longitudinal tear and belt edge damage.At present,most of the research on the surface damage of conveyor belt is longitudinal tearing,and the research on the edge damage is insufficient.Therefore,this thesis studies the detection and localization,feature extraction and recognition of the edge damage.It is of great significance to carry out real-time and reliable detection of the damage of the conveyor belt edge and take corresponding effective preventive and remedial measures in time,to extend the service life of the conveyor belt,ensure production safety and improve economic benefits.In this thesis,a new belt edge damage detection system based on photoelectric sensor,intelligent camera and RFID is proposed.The system uses photoelectric sensor to detect the belt edge of conveyor in real time.When the damage is detected,the photoelectric sensor sends out a signal and triggers the smart camera to collect the image,the abnormal damaged area is captured and the damaged image is processed by the image processing unit of the smart camera.Taking the time of conveyor belt rotating for one cycle as the single period of detection,the distance between the damaged area and the fixed position RFID reader can be calculated by the initial response time and final response time recorded by RFID tag in RFID module,combined with the detection time and conveyor running speed,so as to realize the location of the damaged area.In this thesis,the image features of conveyor belt surface damage are analyzed,and a method of belt edge damage feature extraction and recognition based on image processing technology is proposed.Firstly,this thesis compares the common denoising algorithms,and the original denoising algorithm is improved to get the denoising superposition algorithm suitable for this thesis.The algorithm first uses the median filter to denoise the damaged image,and then uses the morphological operation to remove the isolated holes and smooth the edge burr in the image,and obtains the final denoising image.The damage is segmented from the background by automatic threshold segmentation algorithm.Secondly,the gray level co-occurrence matrix is used to calculate the four kinds of characteristic values of the conveyor belt image,and the texture features of the belt edge damage are extracted,and establish the corresponding feature vector.The obtained feature vector is used as the input vector of BP neural network,and the momentum BP method is used to improve the neural network to judge the authenticity of the damage.Finally,the pixel coordinate system is established to calculate the maximum length of the horizontal and vertical directions,and the influence of different shapes and different light intensities on the feature extraction algorithm is analyzed.The experimental results show that the denoising algorithm,the segmentation algorithm and the geometric feature extraction algorithm have good adaptability and high accuracy,which can effectively segment the damaged area and detect the geometric features of the damaged area.In this thesis,the key technology of conveyor belt edge damage detection is studied.With the help of multi-sensor integration technology,the detection and localization,and real-time monitoring of belt edge damage in the process of conveyor belt transportation are realized.Using image processing technology,the staff can diagnose and evaluate the damage degree according to the image information,and make timely judgment and repair for the possible consequences,so that provide reliable and effective decision-making information for the staff,understand the potential safety hazards in time,reduce the accident risk,and improve the safety production efficiency.