Research On Orange Quality Detection Algorithm Based On DSP

Fruit quality detection based on DSP is an usual topic in post-processing of agricultural products and image processing area. The orange quality detection algorithms as research purpose, with the aid of OpenCV for analysis and comparison of orange image processing algorithms, this research simulated various algorithms by OpenCV and transplanted them to DSP Experimental platform and then optimized them according to characteristics of DSP.The study work mainly reflects in the following aspects:1: analyzing the important part of orange image preprocessing and give a reasonable arrangement of image preprocessing. Paper compared various segmentation methods and finally achieved a good segmentation results with maximization of the measure of reparability on the basis of between-class variance method.2: by eliminating segmentation of interference target using mathematical morphological method, research obtains a binary image of fruit target and then extracted the object boundary according to the difference of the images.3: put forward the method on orange defect detection and distinguished two type defection: disease defect and rot defect. Realized detection algorithm with OpenCV and inspection precision of normal orange is 90%.4: put forward the method on shape detection. This method use average radius and roundness to descript the shape characteristic and inspection precision of normal orange can reach 89%.5: put forward the maturity detection method based on color information. By calculating maturity area proportion, this method can descript maturity quantitatively.6: transplant the validated algorithms to DSP platform and then do some necessary optimization to improve system performance.