A maize pot experiment with different copper stress gradients was designed in an outdoor greenhouse to explore the sensitive leaf types and spectral ranges of crop pollution response under heavy metal stress. Taking maize leaves as the research object, the hyperspectral reflectance data and heavy metal content data of maize leaves during the heading period were measured using instruments, providing basic data for research. This paper designed the Leaf Spectral Detection Method (LSDM) from the frequency domain perspective, combined with time-frequency analysis, to obtain sensitive leaf shapes and spectral bands under heavy metal copper stress, providing technical support for heavy metal monitoring in crops. Based on the growth process of maize, this study explores the full spectrum and sub-spectrum of the old leaf (O), middle leaf (M), and new leaf (N) spectra from 350 to 1 300 nm. Firstly, the hyperspectral reflectance data of maize leaves under copper stress were subjected to double differentiation (SOD) and envelope removal (CR) and transformed into the frequency domain. The Daubechies wavelet 6-layer decomposition was performed using time-frequency analysis methods. Then, based on the signal anomaly points, wavelet high-value points, and SODCR curve high-value points, the spectral anomaly parameters SAP (Spectral Anomaly Parameters) of maize leaves are defined, namely: Abnormal Changes Reflectivity (ACR), which is the absolute value of the difference between the abnormal reflectance and the reflectance of the next adjacent band; Abnormal Wavelet Coefficients (AWC), which is the absolute value of the difference between the abnormal wavelet coefficients and the wavelet coefficients of the next adjacent band; Abnormal SODCR value (ASR), which is the absolute value of the difference between the abnormal point SODCR value and the SODCR value of the next adjacent band. Finally, by examining the correlation between spectral anomaly parameters and heavy metal content in maize leaves, we aim to explore the leaf types and spectral segments sensitive to copper pollution. The results showed that the leaf spectral detection method LSDM can efficiently enhance weak information in maize leaves and accurately locate the spectral anomaly caused by heavy metal copper stress, with the anomaly range concentrated within 350 to 800 nm; Spectral anomaly parameters can quantitatively measure the spectral anomalies of maize leaves under heavy metal copper stress; Under different copper stress gradients, maize new leaves (N) exhibit sensitive leaf types, with sensitive spectral segments including blue edges, green peaks, yellow edges, and red valleys. This paper can provide technical support for monitoring heavy metals in other cereal crops and their canopy scales.