• Application background

▶ The Necessity of Patrol Inspection for Cracks and Diseases on Expressway Pavement

Highways are important transportation arteries that bear the traffic pressure of a large number of vehicles. Cracks are a type of road surface disease. If not repaired in a timely manner, it may lead to problems such as road looseness and collapse, affecting traffic safety and smoothness. Therefore, it is necessary to conduct inspections on road surface cracks and diseases.

▶ Advantages of road surface disease (crack) inspection

(1) Improve efficiency

MYUAV DRONE can quickly cover large areas of road surfaces, reducing waste of human resources and time costs. Compared to manual inspections, MYUAV DRONE can complete crack detection tasks more quickly.

(2) Improve accuracy

Real time capture and record of cracks on the road surface, avoiding subjective misjudgment and missed detection by humans. By utilizing techniques such as image processing and machine learning, automated analysis and classification of cracks can be carried out to improve the accuracy and reliability of detection.

(3) Improve security

Cracks are a common condition of road surface diseases, and if not repaired in a timely manner, it can bring potential risks to traffic safety. Using MYUAV DRONE for crack detection can timely detect and record road cracks, providing timely and accurate information for relevant departments, and facilitating maintenance and repair.

(4) Reduce costs

The traditional crack detection method requires a lot of manpower and time, and it is costly to conduct comprehensive inspections on a wide range of road surfaces. Conducting crack inspections on MYUAV DRONE can reduce operational costs, improve detection efficiency and accuracy, and thus reduce maintenance and repair costs for relevant departments.

•  The technical application of crack recognition algorithm

(1) Image acquisition

The MYUAV DRONE is equipped with a high-resolution camera or sensor to take real-time photos or conduct regular inspections of the road surface, and obtain image data of the road surface.

(2) Image preprocessing

Preprocess the collected image data, including image denoising, grayscale, contrast enhancement, and other operations. Denoising and enhancing images are helpful for subsequent crack detection and recognition.

(3) Crack detection

Extract potential crack areas from preprocessed images through algorithms such as edge detection, morphological processing, and filtering.

(4) Feature extraction

Extract features from the detected crack areas, such as color, texture, shape, etc. Feature extraction helps to distinguish cracks from other road texture or shadow interference factors.

(5) Classification and recognition

Use machine learning algorithms such as support vector machines, neural networks, deep learning, etc. to classify and recognize the extracted crack features. When training a model, a large number of annotated crack images can be used as training data, enabling the model to automatically recognize real cracks.

▶ Principle of crack recognition algorithm

The road disease algorithm recognition inspection system is based on deep learning algorithms to train road crack disease images, thereby achieving road crack recognition, segmentation, and statistics. The algorithm has the characteristics of high efficiency, accuracy, and stability, and can continuously recognize real-time road photo data. It can automatically identify cracks above 5mm on the road and ensure a recognition rate of over 85%.

Deep learning adopts transfer training mode for training, which can achieve high adaptability to different scenarios and can be targeted for upgrading according to the needs of specific problems. By retaining the sustainable updating ability of the algorithm, it can be updated in the background to achieve faster detection speed, higher detection accuracy, and more types of disease detection functions, which can continuously improve a set of algorithms within the hardware cycle.

• System design and implementation plan and implementation

▶ system architecture

(1) Sky end: MYUAV DRONE flight platform, high-definition camera

(2) Ground end: remote control, ground station

--DRONE SOLUTION PROVIDER--

MYUAV® TECHNOLOGIES CO.,LTD.

Tax No.: 91320118MA275YW43M  Legal Register No.: 320125000443821

Add:No.89，Pingliang Street，Jianye District, Nanjing,China 210019

M: myuav@myuav.com.cn T:+86 25 6952 1609 W:en.myuav.com.cn

[Caution]MYUAV™ is a manufacturer of defense products and is under security management by a state agency.

Drawings, materials, samples, etc. provided by us shall be used only for the purposes provided.