Application and analysis of lifting monitoring in digital oilfield

Oil fields shoulder the important responsibility of safeguarding national rights and interests and need to make full use of the management function of video surveillance technology. In recent years, the number of monitoring equipment has been increasing and the working environment of the equipment is complex and diverse, which puts forward new requirements for video monitoring. At present, digital video surveillance system has become the mainstream of video surveillance system, especially in oilfield video surveillance system, which plays an irreplaceable role more and more.

Through advanced image digital compression technology, it has changed the traditional way of image simulation transmission, can easily achieve remote storage and system expansion, and at the same time cooperate with intelligent digital video software, able to maximize digital technology

With the advantages of artificial intelligence of video, it saves a lot of human resources and costs, improves the efficiency and reliability of monitoring, and provides an important guarantee for the safe operation and management of oil fields.

The dynamic detection of the digital video surveillance system is to analyze the images obtained by the camera by the controller. If the images of the areas concerned by surveillance kanshisha change (for example, someone passes, brightness value mutation, etc.), the system will automatically alarm in time to remind surveillance kanshisha to pay attention. Dynamic detection of digital video surveillance system can effectively prevent theft and fire. As we all know, in oil field production, if problems can be found tens of seconds or even dozens of seconds in advance and handled in time, the Help to safe production is inestimable. The common dynamic detection function algorithm is mainly time difference method. The time difference (also called adjacent frame difference) method makes full use of the characteristics of video images and extracts the required dynamic target information from the continuously obtained video streams. If the video images collected under normal circumstances are carefully compared with two adjacent frames, it can be found that most of the background pixels remain unchanged, only the pixel difference of some adjacent frames with moving targets in the foreground is relatively large. The time difference method uses the subtraction of adjacent frame images to extract the information of foreground moving targets. If the moving object does not move after it appears in the scene, this method will regard it as a part of the scene and no prompt will be given.

The main problems are as follows:

(1) real-time robust moving target technology for fast detection, tracking, classification and recognition of moving targets.

(2) video surveillance based on mobile cameras.

(3) detect abnormal events, predict and alarm future events.

(4) non-contact identification.

(5) motion analysis and behavior understanding.

3.1 Adjacent frame difference method

The most commonly used algorithm in moving object detection is the adjacent frame difference method. This method uses the foreground and background images to realize the obvious difference image in the two adjacent frames of the image sequence of the moving object.

Basic requirements of adjacent frame difference method: the noise in the target area is small, the speed is not 0, the brightness changes obviously, and the background basically remains unchanged. Relevant methods are adopted to compensate global motion, such as coordinate transformation method, block matching method, regional difference double difference method, double difference method, etc. The double difference method mainly uses the noise to show difficult repetition in the time domain, and the isolated noise is eliminated by multiplication.

3.2 optical flow method

Optical flow is the instantaneous velocity field generated by pixels on the observed surface of space moving objects when moving, which contains important information of object structure and dynamic behavior. Optical flow usually moves through cameras, target Motion or both.

Optical flow method is mainly divided into three categories: Frequency domain-based method, gradient-based method and matching-based method.

(1) method based on frequency domain.

(2) based on the matching method.

(3) gradient-based method.

4.1 remote control of oil exploitation to realize unattended operation

Many oil and gas fields are located in remote locations, working environment is extremely difficult, and many key oil and gas wells have heavy mining tasks, so safe production has not been fully guaranteed. Using advanced intelligent monitoring equipment to realize real-time monitoring of oil and gas exploitation, the monitoring system can locate the abnormal position of the equipment in time and automatically identify and analyze the abnormal situation of the equipment, and notify the corresponding technical personnel to carry out equipment maintenance to realize intelligent monitoring and centralized control. Real-time image acquisition can be performed on the mining site of high-risk projects. For single Wells, camera equipment of fixed focal length and fixed position can also be used to collect production data, and the site, remote monitoring of well site and other models can improve the monitoring quality and further realize unattended operation under the condition of reducing labor force.

4.2 Digital gas gathering station replaces traditional manual gas gathering station

Install panoramic cameras in gas gathering stations, material storage areas, generator rooms and central control rooms, and set automatic alarm systems for unknown personnel to break into key areas such as central control areas, the intrusion detector system is installed on the wall of the gas gathering station, which can comprehensively monitor the personnel access and equipment operation, automatically identify and alarm abnormal situations, and provide video support for the work of the security department.

4.3 Remote access control system

Most oil mining companies in our country have introduced advanced intelligent access control systems, and internal personnel can open the door through face brushing or fingerprints, it basically realizes visual intercom with outsiders and remote control of gate switch. The intelligent access control system adopts the magnetic lock control gate. When there are illegal intrusions by outsiders or internal personnel do not swipe their cards and face according to regulations, the alarm system will be activated immediately for sound prompt, and automatically record the collected video and image information.

4.4 intelligent video alarm

The video camera monitors the key protected areas such as oil pipes and pump channels in the production area in real time. The back-end platform captures and analyzes the video in real time through the front-end monitoring equipment, analyze the target attributes and motion tracks of various alarm information such as oil leakage, flame, smoke and so on that may occur in the production area in real time, and feedback the analyzed alarm information in time, remind the monitoring personnel to handle it in time.

4.5 intelligent temperature alarm

It is not a good way to add some high-precision physical testing equipment to the existing electrification equipment to ensure the safety of the production area. Although physical testing methods can obtain all kinds of effective data information, but at the same time, it also increases the potential hazard sources in the production area. Establishing a video-based intelligent analysis system can obtain effective data in detail without causing too many security risks to the existing production safety. Real-time analysis of video (visible light and infrared) images by computer can identify abnormal temperature rise, smoke, flame and other phenomena in the production area, and automatically alarm.

Intelligent video surveillance system is of great significance in accelerating the informatization construction of domestic oilfield business field and improving the overall management level of enterprises. According to the different characteristics of oil fields and the consideration of related costs, the target detection algorithm with higher precision is adopted, and relevant hardware equipment is applied to carry out system components, and a reasonable networking scheme is proposed. In order to speed up the process of digital construction of domestic oil fields, it is necessary to focus on studying the data obtained by the video detection system in full swing to evaluate the events and ensure the real-time and reliability of the data.