A Trinocular Stereo System for Detection of Thin Horizontal Structures

Abstract

Many vision-based approaches for obstacle detection often state that vertical thin structure is of importance, e.g. poles and trees. However, there are also problem in detecting thin horizontal structures. In an industrial case there are horizontal objects, e.g. cables and fork lifts, and slanting objects, e.g. ladders, that also has to be detected. This paper focuses on the problem to detect thin horizontal structures. We introduce a test apparatus for testing thin objects as a complement for the test pieces for human safety described in the European standard EN 1525 Safety of industrial trucks -Driverless trucks and their systems. The system uses three cameras, situated as a horizontal pair and a vertical pair, which makes it possible to also detect thin horizontal structures. A sparse disparity map based on edges and a dense disparity map is used to identify problems with a trinocular system. Both methods use the Sum of Absolute Difference to compute the disparity maps. Tests show that the proposed trinocular system detects all objects at the test apparatus. If a sparse or dense method is used is not critical. Further work will implement the algorithm in real time and verify it on a final system in many types of scenery.

Enhancement Technique for Improving the Reliability of Disparity Map under Low Light Condition

“…The disparity map for stereo camera obstacle detection is derived using two approaches [11]. First approach involves deriving dense disparity map by correlating every pixel in the left image with right image [5,7] and the second approach involves deriving sparse disparity map by correlating only selected features like edges and corners of the left image with the right image [4,[12][13].…”

“…Sparse disparity map has disparity data only with respect to the features in the stereo image but involves more steps in computation [11][12].…”

“…Dense disparity map calculation is direct and provides good and detailed depth map with lot of information [6,11].…”

“…For smaller window size, the smaller objects (objects at far distances appears smaller) and larger objects can be separated in disparity map but the disparity map obtained will be noisy. For larger window size, the smaller objects are masked and only larger objects can be separated but the disparity map obtained will be smooth and less noisy [11]. A suitable window size usually smaller window with proper combination of smoothing filters can be used for obstacle detection.…”

“…Disparity map gives the difference in the projection of all the points in the scenario obtained by correlating the left and right images captured [5][6][10][11]. Stereo camera based system is dependent on the external lighting conditions and is passive in nature [1].…”