Indoor corner control of smart car based on binocular stereo vision servo

0 Preface

An important task of the intelligent vehicle environment sensing system is to provide the location information of the vehicle during the exercise in real time. Machine vision has been widely used in smart car environment perception systems because of its wide range of signal detection, complete information acquisition, human cognitive habits, low maintenance costs, and no environmental pollution. The main task in the machine vision vehicle environment sensing system is to complete the identification and tracking of roads and targets, and provide the necessary vehicle position information and surrounding environment distance information for the smart car to exercise.

For the roads in reality, they can generally be divided into two categories, namely: structured roads and unstructured roads. There are obvious road markings on the structured roads, and these markings have strong geometric features. The road pavement is mainly determined by these marking boundaries, such as the yellow continuous marking line or the white intermittent marking line in the middle of the road on the highway. White continuous marking lines on both sides; there are no obvious road markings on unstructured roads. In two-dimensional images, road pavement and non-road pavement are mainly distinguished by texture and color, such as unmarked cement road, wild dirt road or stone road. .

The corner image contains a wealth of road information and environmental information, explaining the scene around the road. The corner detection detects the boundary of the curved lane line from the road image, which is also the basis for understanding the curve. The curve model is established; the lane line pixels are extracted; the lane line model is a commonly used cognitive method, and it shows better detection effect on a specific structured road. Literature [6] introduced the application of curve detection in the fields of lane departure warning, corner speed limit and corner collision prevention warning, and proposed that the curve detection should establish a three-dimensional lane line model to improve applicability.

The Hough transform is used to obtain the straight line equation of the lane line, so as to determine the lowest point and the highest point on the corresponding straight line segment, then judge the bending direction of the curved road according to the corresponding criteria, and finally realize the straight line segment and the curved segment of the lane line. Two-dimensional reconstruction of the lane line.

Curved road detection not only needs to identify the road boundary line, but also needs to judge the road bending direction and determine the radius of curvature of the turning. Commonly used lane detection methods can be divided into two categories: road-based features and road-based models. At present, the road model-based method is mainly used in foreign countries, which converts the curve detection into the solution of mathematical parameters in various curve models. Omit the curve curve model establishment and the complex solution process of mathematical parameters. This paper uses the three-dimensional information of stereo vision perception environment, and uses its parallax principle to judge the position of the corner feature in the surrounding environment image of the indoor road to recover its three-dimensional information. Body cornering and yaw angle. A visual servo control system that directly controls the vehicle body driving off-angle and deviation distance is established. The BP control strategy is adopted. The Simulink simulation environment is used to realize the smart car turning control motion for the unknown curve curvature.

8 Inch LED Downlights is can be detachable in 2 parts. For 8 inch LED Downlights, we can do thw watt from 10W to 40W. 10W, 15W, 18W, 30W, 40W are the conventional product. 8 inch 30W are the popular one. 

The 8 inch downlight use Samsung SMD5630 chips with Philips or Lifud driver. Use high quality PC cover, and adopts precision die-casting aluminum for the module, nice heat dissipation.

Dimmable function can also provide. White, black, silver corlor for option. High CRI >80Ra. High PF >0.95. 100lm/W.


For the 8 inch Led Downlights, it can also up to IP54 protection level. All the LED Downlights is CE RoHS approved. 8 inch LED downlights is easy to transport and install. 185-210mm hole size is suitable for 8 inch LED downlights.

8 inch downlight

8 inch color changeable downlight

8 Inch LED Downlights

8 Inch LED Downlights, 8 Inch Dimmable LED Downlights, Aluminum 8 Inch LED Downlights, LED Recessed Lighting Downlight


This entry was posted in on