As to aided driving by harvesting robot, there is the large amount of image data processing, meanwhile, harvesting robot requires real-time image processing and to calculate the linear parameters in a straight line detection of the walking target process. This paper presents a hardware processing platform to TMS320DM6437 digital signal processor as the core processing chip, and an improved Hough transform algorithm which is based on a determined point is proposed to complete line detection. A camera is to be fitted on the left top of the combine harvester in order to capture images of farmland scenes in the process of harvesting. At first, according to different color characteristics of harvested areas and non-harvested areas, improved methods of Maximum entropy threshold segmentation and morphological approach are employed to determine candidate points of walking goal line. Then the candidate points are selected as the point set. Finally, the improved Hough transform based on a determined point is applied to complete line detection. The algorithm simplifies binary map to unitary map. Comparing with the traditional Hough transform, it saves computing time, and reduces the parameter space greatly. After multiple images processing, tests show that this detection method can detect real-time parameters of harvesting robot’s walking goal line, and the algorithm is well proved with respect to its speed, anti-interference and accuracy.
[1]
Mao Enrong,et al.
Agriculture extra-green image segmentation based on particle swarm optimization and k-means clustering.
,
2009
.
[2]
Mao Enrong,et al.
Recognition and classification for vision navigation application environment of agricultural vehicle.
,
2009
.
[3]
Amir Averbuch,et al.
Color image segmentation based on adaptive local thresholds
,
2005,
Image Vis. Comput..
[4]
He Pei-kun.
High-Speed Image Sampling and Detection System Based on DSPC64X
,
2005
.
[5]
Gao Chi,et al.
Double three-point fix and objective diagram fitting controlling of farming robot.
,
2009
.
[6]
X Hi.
Vehicle flow detection system based on machine vision
,
2008
.
[7]
Wang Ku,et al.
Identifying the navigation route based on TMS320DM642 for agriculture visual robot.
,
2009
.
[8]
N. Otsu.
A threshold selection method from gray level histograms
,
1979
.
[9]
T. W. Ridler,et al.
Picture thresholding using an iterative selection method.
,
1978
.