Oil Barrel Bunghole Recognition Technique Based On Computer Vision

With the development of computer vision technology, many applied techniques based on computer vision is coming out continuously, which is playing a more and more important role in both production and daily life. Especially in many fields of industrial production auxiliary, computer vision technology has played an important role in situations which can’t be replaced by artificial work,such as high toxicity or high precision requirements and so on. In this paper, on the basis of computer vision technology combined with the problems which came up in the process of production, a method to solve the problem that the distinguishment and position of eccentric bunghole was needed to recongized.The distinguishment and position of eccentric bunghole is a difficult and high-risk and heavy-workload task. At present our country as a whole still adopt artificial detection method, relying on the workers to detect the whole power manually. This method is inevitable, and requires long cycle, heavy task, easy omission, with a high rate of false dection for the detecting worker. In case of dection faults, the detecting time is relatively long, which may lead to a long period without power for production and let down the effectivity of production.To correct the drawbacks of the artificial detection for the overhead power transmission line, in this paper, we put forward a relatively efficient and convenient technique. We design the detection system for overhead power transmission line based on computer vision. In this system, we install cameras for observation, and take photos of the lines, and send them back to the sever and process the pictures. Below is the basic procedure of the system:1. Firstly, the cameras observe the lines and each level’s towers. The real-time video will be send back to the sever.2. Each frame of the received videos will be processed and calculated with the computer vision method and artificial intelligence algorithms.3. The analysis includes mechanical invasion force, detection the illegal buildings, the dump soil test, tree test, impurity detection and high-temperature detection.4. The software then used background modeling object identification, feature extraction and motion detection, which are well-known computer vision methods.5. At last, the sever gives the alarm according to the image damaged degree made by foreign forces.By setting up several video sample points, we implemented the real time observation about the whole power line. Once the line is broken, we can get the feedbacks and take actions at the first time. The reaction time is really cut down, and the efficiency is raised up, compared to former method.The malfunction detection system for overhead power transmission line based on computer vision can solved the problems well. And it has the irreplaceable advantages over the manual method. However, there are still drawbacks remaining to be solved, which will be corrected in the near future.