Method for determining a state of obstruction of at least one camera installed in a stereoscopic system
To improve the performance for determining obstruction of a stereoscopic system using two cameras or more, a hybrid of local and semi-global methods is provided. For each stereoscopic image formed from simultaneous left and right images, a breakdown of each left and right image into corresponding sectors is applied. Obstruction level is determined by a disparity map by sector, based on left or right images, and in which a disparity is assigned to each pixel corresponding to the best matching score. A determination of density by sector of the disparity map is carried out by reference to a fraction of pixels with a disparity considered to be valid. A state of obstruction of at least one camera is determined based on a weighted average of the probabilities of obscuring of the sectors of the disparity map obtained by comparison between the density of the sectors and a predefined density level.
1. A method for determining a state of obstruction of at least one camera in a multi-camera system installed on board a vehicle comprising the following steps:
acquisition of the successive stereoscopic images of a field of view, each stereoscopic image from the multi-camera system being formed from left and right images ( 11 , 12 ) produced simultaneously and digitally stored in the form of pixels,
calculation of a disparity map based on successive multi-camera images, and
calculation of a level of obstruction as being a weighted average determined by the following steps:
breakdown of the disparity map ( 20 ) into sectors ( 10 ),
determination of the density of each sector ( 10 ) by the fraction of pixels with a disparity considered to be valid ( 108 ),
determination of a probability of stereo obstruction by sector ( 109 ) of the disparity map ( 20 ) by comparison between the density of this sector ( 10 ) and a predefined level of obstruction, and
determination ( 119 ) of a weighted average of the probabilities of obscuring of the sectors ( 10 ) as a function of a weighting ( 118 ) of the position of these sectors ( 10 ) within the disparity map ( 20 ),
wherein the method furthermore comprises a complementary mono-stereo combined analysis test including the following steps in order to determine the presence of an object in front of a camera:
selection ( 112 ) of one of the two stereo images ( 11 ; 12 ) to be analyzed, referred to as mono image;
breakdown ( 113 ) of this mono image ( 11 ; 12 ) into sectors ( 10 a ) with the same dimensions as those ( 10 b ) of the disparity map ( 22 ) in order to form a mono map ( 32 );
calculation ( 114 ) of a probability of mono obstruction by sector ( 10 a ) of the mono image ( 11 ; 12 ), carried out by an edge detection analysis and a detection of characteristic points using the detection data, transferred onto the mono map ( 32 );
comparison of the probabilities of stereo and mono obstruction, then merging ( 115 ) between the disparity map ( 22 ), having the probabilities of stereo obstruction ( 109 ) by sector ( 10 b ) and the mono map ( 32 ), having the probabilities of mono obstruction ( 114 ) by sector ( 10 a ), by assigning to each sector ( 10 b ) of the disparity map ( 22 ) the lowest probability of obstruction, corresponding to the highest density, so as to generate a disparity map referred to as merged ( 42 ); and
if the probability of merged obstruction is globally less than the probability of stereo obstruction (test 116 ), the camera corresponding to the image being analyzed ( 11 ; 12 ) is considered as potentially obstructed and a processing is triggered ( 117 );
in the opposite case, an object is considered as having masked the field of view of the camera being analyzed.
2. The method for determining a state of obstruction as claimed in claim 1 , in which a digital filtering of disparity noise ( 105 ) is carried out by sector ( 10 ) on the disparity map ( 20 ) prior to the determination ( 108 ) of density of the sectors.
3. The method for determining a state of obstruction as claimed in claim 2 , in which the digital filtering of the disparity noise ( 105 ) is carried out by the application of mathematical morphology tools to the disparity map ( 20 ).
4. The method for determining a state of obstruction as claimed in claim 2 , in which, each sector ( 10 ) being broken down into subsets called macroblocks ( 30 ), with the same dimensions and regularly distributed within each sector ( 10 ), the digital filtering of the disparity noise ( 105 ) is carried out by sector ( 10 ) by measuring the density of each macroblock ( 30 ) of this sector ( 10 ) by its fraction of pixels whose disparity higher than a threshold is considered to be valid, the proportion of macroblocks ( 30 ) considered to be valid determining ( 108 ) the density of the sector ( 10 ).
5. The method for determining a state of obstruction as claimed in claim 4 , in which the probability of stereo obstruction by sector ( 10 ) is determined by comparison (step 109 ) between the proportion of macroblocks ( 30 ) considered to be valid in this sector ( 10 ) and a predefined level of obstruction.
6. The method for determining a state of obstruction as claimed in claim 1 , in which the number of sectors ( 10 ) is chosen substantially between 10 and 50.
7. The method for determining a state of obstruction as claimed in claim 4 , in which the number of macroblocks ( 30 ) by sector ( 10 ) is chosen substantially between 10 and 100.
8. The method for determining a state of obstruction as claimed in claim 2 , in which the number of sectors ( 10 ) is chosen substantially between 10 and 50.
9. The method for determining a state of obstruction as claimed in claim 3 , in which the number of sectors ( 10 ) is chosen substantially between 10 and 50.
10. The method for determining a state of obstruction as claimed in claim 4 , in which the number of sectors ( 10 ) is chosen substantially between 10 and 50.
11. The method for determining a state of obstruction as claimed in claim 5 , in which the number of sectors ( 10 ) is chosen substantially between 10 and 50.
12. The method for determining a state of obstruction as claimed in claim 5 , in which the number of macroblocks ( 30 ) by sector ( 10 ) is chosen substantially between 10 and 100.