Home Intelligent Cleaning System Based on Machine Vision

[0013] The present invention will be described in detail below in conjunction with the drawings.

[0014] Attached Figure one It is the overall structure diagram of this system. As shown in the figure, the system includes a processor platform, infrared camera, radar, vacuum cleaner, infrared communication module, electric motor, wheels and so on. The DSP chip is the core of the processor, connected to SDRAM and FLASH through the EBIU bus, responsible for data calculation and storage; connected to the video decoder chip through the PPI interface, to receive video image data; the infrared communication module was connected to the PF interface of the DSP chip to transmit the remote control The command. The infrared camera is fixed on the rotating column and can be rotated left and right to scan and capture images, and the back end is connected with the video decoding chip on the processor platform. The wireless radar is placed under the camera and also swings with the camera to measure the distance and specific orientation. The electric motor is placed at the bottom and connected to the wheels to control the direction and speed of the system. The suction port of the vacuum cleaner is placed at the bottom of the front end, connected to the trash bin, and is responsible for cleaning.

[0015] Attached Figure II It is the overall software flow chart of the system except for the remote monitoring function. The system adopts Linux embedded operating system and programming based on hardware drivers. After the infrared camera captures the image, it undergoes image filtering, light compensation, grayscale processing, binarization processing, threshold segmentation, feature extraction and other operations, and then detects and recognizes the topological relationship between the foreign object and the surrounding environment and the characteristics of the foreign object itself. If it is a foreign object, the motor rotates to control the rotation of the wheel and advance to the foreign object. If it is garbage, the vacuum cleaner is turned on to clean up foreign objects. If it is a coin, key, etc., the system will alarm to remind you to pick it up. If the infrared communication part receives the data, the system stops the detection, and completes the corresponding operations of forwarding, turning, cleaning, and stopping according to the command. The core part of the whole system is video processing, foreign body detection and recognition. Its workflow is:

[0016] (1) After the camera acquires the original image, it performs preprocessing such as smoothing and enhancement of the image.

[0017] (2) Using the image segmentation technique of relative entropy threshold to obtain the best global threshold to facilitate subsequent feature extraction and recognition.

[0018] (3) Filter the image obtained after segmentation to remove the outlier noise that may be contained.

[0019] (4) Using machine vision invariance method, for the processed image, extract image feature values ​​such as grayscale, texture, color, and structural features as original features, and compare them with the feature values ​​contained in the foreign objects stored in the memory to determine whether they are foreign objects If yes, the motor brakes.

[0020] (5) Compare the detected foreign objects with the characteristics of various foreign objects stored in the database to determine what kind of foreign objects are. If it is garbage, turn on the vacuum cleaner; if it is a coin, key, pen, or magnetic card, turn on an alarm prompt.