Design and Implementation of an Automated App Advertising Screenshot System

1. Project Background

Mobile internet technology is rapidly developing, with more users preferring mobile devices for information access. Mobile advertising spending is increasing, but instability and compatibility issues can prevent ads from displaying properly. Currently, staff manually capture screenshots of app ads daily, taking about 10 minutes per ad. The author's department developed an automated system to reduce this workload.

2. Business Requirements

The system must automatically capture ads in real-time with high accuracy, support all cities, include timestamps, and distribute screenshots by dealer.

3. System Design and Implementation

The system uses Macaca engine, script recording, and app adaptation engines. It employs WindowManager for floating timestamps, hooks for city simulation, and OpenCV for image matching.

3.1 System Architecture

The system includes device, engine, and application layers. PhoneBus manages devices, Macaca handles automation, and various modules handle tasks like scheduling and distribution.

3.2 Core Workflow

The workflow includes pulling contract data, syncing ad materials, and executing screenshot tasks. Different ad types (text, image, text-image) use tailored strategies.

4. Application Effects

The system has achieved a 99.5% success rate, freeing staff from manual work and identifying issues promptly.

5. Future Plans

The system will be decoupled for better scalability and compatibility with various business needs.