Cucumber Detection Based On Texture And Color In Greenhouse

China is a large agricultural country,the agricultural population is the largest in the world,and agriculture is the foundation in our country.China’s land area accounts for only 7% of the world,but to feed the world’s population accounted for 1/5.The population of agricultural production is declining year by year,in the background of acceleration of aging process.So it is urgent to solve this problem.With the development of robotics and artificial intelligence,it is possible for robots to enter the field of agricultural production.The introduction of agricultural robot in agricultural production can effectively alleviate the conflict and promote the development of the robot industry.Not only liberate the finite labor resources,but also can greatly reduce the production cost of cucumber,increase productivity.Agriculture robot by mechanical harvesting requires automatic detection and counting of fruits in tree canopy.Because of color similarity,shape irregularity,and background complex,fruit identification turns to be a very difficult task and not to mention to execute pick action.Therefore,green cucumber detection within complex background is a challenging task due to all the above-mentioned problems.In this paper,a technique based on texture analysis and color analysis is proposed for detecting cucumber in greenhouse.RGB image was converted to gray-scale image and HSI image to perform algorithm respectively.Color analysis was carried out in the first stage to remove background,such as soil,branches,and sky,while keeping green fruit pixels presented cucumbers and leaves as many as possible.In parallel,MSER and HOG were applied to texture analysis in gray-scale image and green channel image relatively.We can obtain some candidate regions by MSER to obtain the candidate including cucumber.HOG+SVM is the used to eliminate candidate regions that not include cucumber.In order to further remove false positives,key points were detected by a SIFT algorithm based on distribution of the key points.Then,the results of color analysis and texture analysis were merged to get candidate cucumber regions.In the last stage,the mathematical morphology operation was applied to get complete cucumber.After the above operation,we can obtained cucumber contour more accurate compared to the general segmentation method.The recognition rate is higher than the common segmentation method in natural light environment,which ensure accurate picking of agricultural robot follow-up.