CN-121988998-A - Photovoltaic module installation method based on computer vision technology

CN121988998ACN 121988998 ACN121988998 ACN 121988998ACN-121988998-A

Abstract

The invention provides a photovoltaic module installation method based on a computer vision technology, which comprises the steps of calibrating a depth camera, capturing an environment RGB image and depth information, constructing a three-dimensional model, positioning a specific position of a bracket by using a computer vision algorithm, calculating an optimal placement posture of a photovoltaic module according to the position of the bracket, changing the optimal placement posture into a coordinate system which can be executed by a mechanical arm, moving the camera, aligning a module area, collecting RGB image and depth information, positioning the position and posture of the module, converting the identified module posture information into a mechanical arm operation instruction, moving the mechanical arm to a preset position, adsorbing or clamping the module by a grabbing device, moving the mechanical arm carrying module to the position above the calculated bracket and placing the module, and confirming whether the module is installed correctly and finely adjusting by visual feedback or a sensor. The invention not only improves the working efficiency, but also improves the data processing speed, enhances the adaptability of the vision system, and ensures the installation precision and consistency.

Inventors

SHEN TING
Chen haotian
LI YANQING
XIANG YANG
Weng Lewei
Yu Cannan
WU CHUANZHEN
MA HAORUI

Assignees

中国电建集团华东勘测设计研究院有限公司
浙江华东工程建设管理有限公司

Dates

Publication Date: 20260508
Application Date: 20241104

Claims (5)

1. The photovoltaic module installation method based on the computer vision technology is characterized by comprising the following steps of: step 1, precisely installing a depth camera at or near the tail end of a mechanical arm to realize the tight combination of vision and mechanical operation; step 2, starting the depth camera, calibrating various parameters of the depth camera, ensuring that the acquired image is clear and accurate, and laying a foundation for subsequent processing; Step 3, commanding the depth camera to capture RGB images and depth information of the surrounding environment of the bracket, and constructing a three-dimensional visual model of the area; Step 4, identifying and positioning the specific position of the bracket from the acquired data by using a computer vision algorithm, and analyzing the relative layout of the bracket in the environment; Step 5, calculating the optimal placement position and posture of the photovoltaic module according to the bracket position, ensuring stability and facilitating subsequent connection, converting the position and posture from a camera coordinate system to a coordinate system which can be executed by the mechanical arm, and generating an installation position and posture instruction, wherein the posture is the inclination angle and the inclination direction of the photovoltaic module; Step 6, adjusting the view of the depth camera, aiming at the photovoltaic module to be operated, and collecting RGB images and depth information again; step7, recognizing and accurately positioning the position and the gesture of the photovoltaic module in the visual field of the camera by using a visual algorithm, and preparing for accurate grabbing; step 8, converting the identified position and posture information of the photovoltaic module into an instruction under a coordinate system executable by the mechanical arm, so as to ensure that the mechanical arm can accurately understand and execute the grabbing action; step 9, according to the instruction in the step 8, the mechanical arm accurately moves to a preset position above the photovoltaic module, and the grabbing operation is ready to be executed; Step 10, extracting a photovoltaic module through an executing mechanism on a mechanical arm; step 11, according to the installation position and posture instruction in step 5, the mechanical arm carries the photovoltaic module to move above the bracket position corresponding to the instruction, and the mechanical arm is lowered and adjusted to the installation posture; and step 12, after the placement is finished, whether the photovoltaic module is correctly and firmly installed on the bracket is confirmed through visual feedback or a sensor.
2. The method for installing the photovoltaic module based on the computer vision technology according to claim 1 is characterized in that in the computer vision algorithm in the step 4, key feature points of a bracket are firstly identified in an image, the positions of the feature points are determined through a two-dimensional image identification technology and then are converted into coordinates in a three-dimensional space by combining depth information, so that a bracket pose is determined, or a point cloud model is generated by utilizing RGB information and depth information of the image, and then the three-dimensional identification algorithm based on the point cloud is used for directly obtaining the bracket pose, or a hybrid network model is adopted, and the point cloud and RGB depth complement technology are combined so as to improve the accuracy and the robustness of identification.
3. The method for installing the photovoltaic module based on the computer vision technology according to claim 1, wherein in the step 5, the position and the gesture are converted from a camera coordinate system to a coordinate system which can be executed by a mechanical arm, an intelligent evaluation mechanism is integrated, the construction quality of the bracket is analyzed synchronously, and the installation deviation and the firmness of the bracket are checked through an image recognition and structure stability algorithm.
4. The method for installing a photovoltaic module according to claim 1, wherein in step 6, the image and data captured by the depth camera not only includes the photovoltaic module itself but also covers the surrounding environmental features, the RGB image provides color information, which helps the vision algorithm to identify the boundary and detail of the photovoltaic module more accurately, and the depth information supplements the data in terms of distance and space layout.
5. The method for installing the photovoltaic module based on the computer vision technology as set forth in claim 1, wherein a photovoltaic module database is established, the photovoltaic module database contains size and shape information of different photovoltaic modules, and in the process of identification and positioning, the vision algorithm can rapidly identify different types of photovoltaic modules and flexibly adjust the follow-up grabbing strategy according to parameters in the database.

Description

Photovoltaic module installation method based on computer vision technology Technical Field The invention relates to the field of automatic installation of photovoltaic modules, in particular to a method for installing a photovoltaic module based on computer vision. Background Currently, the installation of photovoltaic modules still relies mainly on manual operations, and this traditional manner of labor-intensive operation faces multiple challenges. First, because photovoltaic modules are typically heavy and large in size, workers need to operate at high altitudes or on inclined surfaces during installation, which greatly increases the hazards and difficulty of work. In addition, because photovoltaic power plants are often built in remote areas, the climate conditions in these areas may be very severe, such as high temperature, sand storm, etc., further increasing the difficulty of installation work. The manual operation is not only inefficient, but also prone to installation errors affecting the final performance of the photovoltaic system. Advances in computer vision technology have provided new possibilities for automating the photovoltaic module installation process. By using advanced image processing algorithms and technologies, the computer vision system can realize the functions of real-time tracking, measurement, three-dimensional reconstruction and the like of the photovoltaic module, so as to guide a robot or an automation device to finish the accurate installation of the photovoltaic module. Compared with the traditional binocular camera system, the novel visual technical scheme aims to achieve faster data processing speed and higher precision, and simultaneously adapts to different installation environments and support structures more flexibly. Through combining computer vision and arm technique, can develop a set of intelligent installation system, this system not only can improve the security of installation, reduces the risk of casualties, can also greatly improve work efficiency, reduces the installation error because of human error leads to. More importantly, the automatic system can continuously work under any weather condition without time limitation, so that project period is shortened, operation and maintenance cost is reduced, and economic benefit of the photovoltaic system is improved. In summary, with the development of computer vision technology and its application in the photovoltaic field, future photovoltaic module installation will develop towards more intelligent and efficient, which is not only an effective means for solving the current low installation efficiency, but also one of the key factors for promoting sustainable development of the photovoltaic industry. However, in order to truly achieve this objective, there is a need to overcome technical obstacles such as increasing the speed of data processing, enhancing the adaptability of the vision system, and the like. Disclosure of Invention The invention aims to provide a photovoltaic module installation method based on a computer vision technology. The whole process not only improves the working efficiency, but also improves the data processing speed, enhances the adaptability of a vision system, ensures the installation precision and consistency, and is an important step for intelligent manufacturing application in the photovoltaic field. For this purpose, the invention adopts the following technical scheme: The photovoltaic module installation method based on the computer vision technology is characterized by comprising the following steps of: step 1, precisely installing a depth camera at or near the tail end of a mechanical arm to realize the tight combination of vision and mechanical operation; step 2, starting the depth camera, calibrating various parameters of the depth camera, ensuring that the acquired image is clear and accurate, and laying a foundation for subsequent processing; Step 3, commanding the depth camera to capture RGB images and depth information of the surrounding environment of the bracket and constructing a three-dimensional visual model of the area, wherein the depth information is the distance from an object to the camera in the visual field shot by the camera; Step 4, identifying and positioning the specific position of the bracket from the acquired data by using a computer vision algorithm, and analyzing the relative layout of the bracket in the environment; Step 5, calculating the optimal placement position and posture of the photovoltaic module according to the bracket position, ensuring stability and facilitating subsequent connection, converting the position and posture from a camera coordinate system to a coordinate system which can be executed by the mechanical arm, and generating an installation position and posture instruction, wherein the posture is the inclination angle and the inclination direction of the photovoltaic module; Step 6, adjusting the view of the depth camera, aiming a