Abstract: To solve the problem of multi-mode cotton boll counting in the complicated environment,an in-field cotton boll counting algorithm based on density classification is proposed.Firstly,the algorithm encodes the global context information with a density level classification estimator.Then the input images are converted into high-dimensional feature maps by density map estimator with multi-column structure.Finally,through the feature fusion neural network,the classification information is combined with high-dimensional feature maps to generate high-quality density map,and then the cotton bolls are counted.In addition,a new dataset within 412 in-field cotton boll images is constructed for experiment and comparison,which can be divided by different environment,year and region conditions.Experimental results demonstrate that the proposed algorithm achieves a lower counting error,and better effectiveness and robustness than other comparison algorithms.